Your thing about pre ringing is just wrong, pre-ringing is an artefact caused through linear phase filters, linear phase filters do cause extra latency but this is not reversible, just because there is extra latency doesn’t mean pre-ringing? DAWs have built in delay compensation, they way that works is just by delaying the other tracks to mean they all arrive at the same time. The true peak limiting thing is just a myth, true peak limiting is simply just a more accurate gain reduction since all its doing is interpolating sample point in between the existing sample points, this does not effect the gain reduction circuit other than making it more accurate, the only reason you could ever get a different “sound” is because when you turn true peak limiting on more gain reduction is happening because transients that weren’t being detected before are now being detected, turn the input gain down and it’ll null. Also the way you’ve set your limiter is basically like a clipper, not having oversampling will allow for aliasing (since as previously mentioned your basically clipping the signal and clipping results in an infinite series of harmonics being created which will fold back when they hit nyquist) Whether aliasing is something you need to be concerned about is another topic but by definition is is not “clean” Also you peak matched the results, this is not the way you level match a master, you match it through loudness normalisation to make sure your actually hearing the difference and not just that one is louder. I’ve noticed a lot of just flat out falsehoods on this channel, like the “cut everything under 30hz for extra headroom” thing or “put your low end in mono” etc etc I don’t have anything against the channel, you’ve got a good message but c’mon, you can’t claim to be an expert in mastering and teach others when you don’t understand how delay compensation works? If it worked like that then every single session would have tonnes of pre-ringing over every track and you’d see it in the waveforms when bounced from once session to another, or do you think DAW developers were smart enough to just do it the same way speakers are done in live sound, you delay one signal so that they all arrive in phase with each other.
Also who cares about delay, if the Mixed Stereo-Signal is delayed completely. It only Mathers when diffrent parts of spectrum gets delayed. I Think most thinks in mastering are overhyped. Ther are tons of Tracks that everybody love that could have been done better.
@@moritzschiekel2272 exactly that problem was solved literally decades ago, why would DAW developed choose to use a way that is more complicated and actively damages the fidelity of the audio instead of just delaying everything else so it's in sync, I don't know where they pulled that from, that's a new one for me.
@@dussonthephanter6698 as I said in my original comment, this was things I noticed in the past from this channel, not that they are currently doing it. I agree that the way to teach something like mastering is to teach the fundementals on an advanced level, understanding what EQ's are actually doing to your signal, what compressors are and what they do etc. audio theory is an insanely deep and often confusing thing (I personally love it, it's a whole world of knowledge and the depth seems to never end) theres a bigger problem at play and it's patently false information being perpetuated by even some of the biggest names, when I was starting out you just go "yeah that sounds about right" because you don't know enough to question it properly, when you start learning a bit more about audio and getting more experience you see these things and go "what the hell are these guys on about". again as I mentioned previously, I don't think someone can claim to be an expert in mastering and teach others when they make such fundemental mistakes like not knowing how delay compensation works. to me it just sounds like they found something cool like pre-ringing (which is a facinating subject that you could spend hours on) then just sprinkling it in to sound smart without actually knowing what it is and how it works. be skeptical always when it comes to this stuff, go and double check all the information your getting especially when your starting out becuase you'll be fooled by someones big studio and go "well how could he be wrong"
@@moritzschiekel2272 exactly, it’s fundamentally bizarre why they would think that pre-ringing could occur on a single stereo track, I really think they heard the term “pre ringing” read a couple lines and it mentioned delay then just used the term pre-ringing without actually understanding what it is.
This video is misleading. LATENCY ITSELF DOES NOT CAUSE PRE-RINGING! LINEAR PHASE FILTERING AND FFT PROCESSING CAUSE PRE-RINGING! Oversampling can cause latency and pre-ringing. Pre-ringing is negligible when using linear phase oversampling filters anyway, because it only affects high frequencies that most people can't hear. You can run linear phase crossover filters (for multiband processing) with ZERO pre-ringing, that have sizeable latency. You can run a compressor with lookahead, where latency is necessary for the sound. No pre-ringing. also importantly, pre-ringing does not (in itself) attenuate transients! It simply smears them across time. Any sort of filter you put audio through (whether that be digital EQ, analog EQ, convolution, brickwall lowpasses during oversampling, or linear phase EQ) changes the dynamics by introducing ringing. Minimum-phase filters cause ringing AFTER the signal, and linear phase filters cause ringing before AND after the signal. Filter ringing is easiest to hear with transients, but it happens with slow dynamics just as much. MINIMUM PHASE EQ DOESN'T SOUND "CLEANER" THAN LINEAR PHASE EQ!!! It's a subtle difference, and which you use depends on your goal. Saturation doesn't sound dirty because it decreases the dynamic range. It sounds dirty because it adds harmonics to the signal (which "fills out" the spectrum) and causes the audio to intermodulate (like when a big sub-bass makes everything else grittier). The goal of mastering isn't to preserve dynamic range, it's to enhance the emotional impact of music, while making it louder. Making something louder necessitates decreasing the dynamic range! For this purpose, saturation is extremely useful. Hard clipping can be more transparent than limiting when taming peaks, and certain types of saturation can make music feel more detailed or punchy, even though it's technically dirtier. On oversampling: it's probably not necessary during mastering IF you're using a high enough project sample rate. Oversampling becomes useful if you're using saturation in series (which is more common during mixing), and if that saturation creates higher harmonics (like hard clipping). But if you lowpass your mix after saturation, you'll probably never introduce aliasing. If you're running at a low project samplerate, oversampling becomes crucial for any saturation processes, and some dynamics plugins too. As for exciters, they should be alright to use if they're oversampled properly and aren't pushed too hard. Exciters and high frequency saturators can sound amazing on a full mix! I'd recommend testing your nonlinear plugins to see if they creates aliasing, and to see how well their oversampling works. Running sine waves through it is useful, but it's equally useful to listen to the delta signal and see if it sounds metallic.
Paralell processing: don't be afraid. This video is once again giving bad advice because of its lack of detail. It is simply not true that "all forms of parallel processing will negatively affect the signal in this way." Only filters, really. Keep reading. If a plugin has a dry/wet knob built in, chances are the developers MADE SURE that it doesn't create phase cancellation. They're not stupid enough to include a "sound bad" knob on their interface. So, why might blending in a parallel signal cause phase cancellation? Two reasons come to mind: 1. The plugin is introducing latency and not reporting it to your DAW. This almost never happens with decent plugins, even free ones. I dunno how latency reporting works, but I know that it has never failed for me. Don't worry about it unless you hear something wrong. This rule applies for very short delays as well. But chances are, you're not adding delay effects during mastering. 2. The plugin is shifting the phase in some way. Unlike the previous example, phase shift is extremely common. Any minimum phase EQ or filter creates phase shift. HOWEVER, unlike what the video said, EQ is NOT the worst culprit when it comes to phase cancellation. Regular bell and shelf filters phase shift proportional to the amount of cut/boost. So even crazy 18dB EQ moves would only result in a slight narrowing of the filter (higher Q setting, essentially), when mixed back into the dry signal. The culprit for strange phase effects is high/lowpass filters. Filters 12dB/octave or steeper create notch filters when blended in paralell. So if you're running a filter in paralell, stick to 6dB/octave. Or simply fun linear phase EQ. Oversampling uses veery steep lowpass filters. Sometimes they're linear phase, but often they're minimum phase. Minimum phase oversampling filters cause crazy phase shifts, so avoid using them in paralell. Some analog modeled plugins will use IR filters or something, so always check the phase response when running in paralell. "How do I tell which type of oversampling filter my plugin is using?" Bertom EQ analyzer is a free VST that'll tell you the frequency response and phase response of any plugin or chain. If you click the "show phase" button, and see that the phase is at 180 degrees at any point in the spectrum, it's gonna cause phase issues. Notice what I didn't mention? Dynamics and saturation. These don't shift the phase, so you can feel free to blend any limited, compressed, gated, expanded, saturated, clipped, ring modulated, or otherwise processed signals in parallel.
Quick correction: 8:01 CLIPPING IS SATURATION! Saturation is a term used to describe any waveshaping distortion that has a hard limit of amplitude. This describes digital hard and soft clipping perfectly! I agree that soft clipping is often more transparent than compression or other types of saturation when reducing peaks. Indeed, many saturators "color" the signal without decreasing dynamic range hardly at all (such as oxford inflator). Another quick correction: maximizer vs. limiter They're basically the same thing. A limiter is a compressor that keeps a signal below a certain threshold (it has a ration of infinity to 1). A maximizer is a tool that uses limiting, but can also include other things like clipping, or limiters in series. Pro-L2 does both of these, and is therefore a maximizer. So while this video makes them look like different utilities, they're essentially the same thing. THIRD CORRECTION: 12:24 The P440 EQ had saturation in its algorithm. While this isn't harmful misinformation, it does kind of obscure the truth. " also really man? the shumann frequency? Its FOURTH harmonic is 33.8 hz, which is below what most consumer speakers can reproduce. It's also basically subsonic. Also also the actual schumann resonance changes over time. Watch this video by Benn Jordan (a.k.a. "The Flashbulb") for more info: th-cam.com/video/Q3LVijzZAe4/w-d-xo.htmlsi=_hg9ouuwo1BSoWGJ What's my point? Sage audio, if you're reading this: Please be specific with your terminology. Please explain your reasoning fully. Please stop spreading misinformation. You're clearly a channel aimed at audio beginners. Make sure they come into the profession with a nuanced, SCIENTIFIC understanding of audio processing. There's plenty of good advice here, and your results speak for themselves. But I dislike seeing so many incorrect or misleading claims.
@@sageaudio Ahh, this is great knowledge! So oversampling is like a filter that creates ringing at the nyquist frequency. Only FIR oversampling filters create pre-ringing (like FIR EQ filters). I'll update my comment to be more accurate. Glad to be learning more about audio. The smearing effect in particular is more noticeable than I thought it was. The length of filter ringing, as I suspected, was directly correlated with the order (steepness) of the filter. He also mentioned that the oversampling process uses allpass filters tuned to a high frequency? Still though, the oversampling itself only causes ringing at high frequencies, just like the lowpass filter associated with oversampling. I don't know about you, but my young ears can't hear anything very clearly above 16khz. So is anyone really gonna notice a 22khz ringing? I personally doubt it. In my opinion, transient smearing at 22khz is easily preferable to aliasing on transients. Although both artifacts are subtle, one is clearly preferable. Very notable, too, is that the IIR oversampling created latency without any pre-ringing. Pre ringing is not caused by latency, as your video suggests. In fact, there's no examples of pre-ringing (which can definitely occur in some oversampling algorithms) in the video you linked. Plus, these artifacts are almost guaranteed to be less noticeable than aliasing. That's not relevant to clean mastering, but it's probably useful knowledge for the average mix or mastering engineer. To summarize: oversampling is a useful tool for dealing with digital distortion, and its adverse effects are generally overstated.
Thanks for watching! You're right! I conflated correlation with causation in this instance. That said, latency is a really reliable indicator of behind-the-scenes filtering that will cause pre-ringing, such as oversampling, linear phase filters, or convolution.
Wow. That was an amazing approach to teaching mastering. Best I’ve seen so far. I don’t really want to do my own mastering - just understand it better. Thank you!
I've been using Newfangled Audio's Elevate Mastering Bundle for a while now and I'm really pleased with the results. The dynamic control and EQ control is very handy. Great video!
The best video around clean mastering I've ever seen sir! You really explained the most detailed things that others might not even consider telling! Definately helped understanding the fundamentals of achieving a clean sound in general! Thank you for this!!!
Thank you for watching and it's great to hear you enjoyed it! I hope the video helps or at the very least gives you some things to consider when you're working on your next session
Nice video - Learned about the 192Khz ringing (which people avoid) and about the Soft Knew transients. I also heard some metallic ringing from snare on final master samples, specifically on FF L2. Not that It was bad but it changed the tone and I liked the Pre-FF L2 better. - But it is subjectively
@sageaudio wrote: "Any instance in which the frequency response of a signal is changed (which occurs with all of the processors I'm using in this chain) and latency is present, the reallocation of the delayed signal will shift artifacts or in this case post ringing, back to the source of this change, in turn making it pre-ringing - its really just added signal that occurs after the change which now has to be shifted back in order for the full signal to be in time/compensated." As a simple example, if you had a plugin (call it A) that did nothing but delay a signal, it would work by allocating some samples of memory, and storing the samples it sees for a set number of samples, then outputting after the appropriate delay (i.e. the number of samples that equate to the given delay based on the sample rate). It doesn't "do" anything with the signal to reconstruct it - it literally stores the incoming sample, then outputs it verbatim a bit later. Now, what if you had a plugin (call it B) before the plugin (A) that had no latency, but did some processing, such as the ones you mention in the video. In this case, we agree that (B) would add no pre-ringing to the signal, right? And we've already agreed that plugin (A) doesn't have any pre-ringing as it's literally outputting the samples that come in verbatim. So, the chain of (B) then (A) has no pre-ringing. So, what is we build a new plugin that does exactly what (B) then (A) does, all within the same plugin. That would have no pre-ringing, right? While it's certainly possible that there are some plugins (and DAWs?) that don't handle latency properly, the above proves that the claim that any plugin (or even any plugin in which the frequency response of the signal is changed") that has delay brings artifacts due to that delay is false. If you disagree with this (IMO clear) counter example, I have never heard this argument made anywhere other than this channel. Can you provide some links to other information out there that corroborates it? (I posted this in a thread below, but wanted to promote to top level.)
I understood the premise to be that latency is an indicator that the process is likely using oversampling. The goal is to avoid stacking multiple stages of oversampling to maintain transient integrity, so it's logical to avoid processors with latency.
@@sparella what you say is logical, however, that is not the point that this video makes. I am specifically arguing against the point that latency itself causes pre-ringing, which is clearly false.
The king is back! Can you please make an in-depth video of how you would mix/master a rock sog? I usualy mix/master trap /rap songs and i got a project that is a bit out of my confort zone and i will appreciate if you are willing to make a video on this topic! Thank you in advance!
Thanks for watching! A video like that is included in our new School of Mastering (we go over creating pop, rock, hip-hop, classical, jazz, electronic music, singer-songwriter, and metal chains), so I probably won't replicate it here, but here's a link in case you're interested: SageAudio.com
Best mastering tutorial on TH-cam, best Mixing and Mastering engineer/sound technician 🔥🔥🔥🔥🔥 Thanks Sage Audio I, learned a lot from this, you also use new plugins too, expect the FabFilter Pro-L2 limiter
Ok so I have thought about the latency issue and I have to agree with some comments below, please can you provide evidence. I enjoyed watching your videos butt his one needs clearing up as I have some big doubts about the truth to this statement. Pro Tools did used to have latency issues, and possibly some bugs can exist with this one. I think that needs further proof as I don't trust this - sorry but I am open to evidence if you have some. I checked out your website, but it just wants me to pay to sign up? Think that could do with a little more work to explain the benefits.
Thanks for watching and sharing your thoughts! Any situation in which delay is compensated for, and the frequency response of a signal changes, will cause some form of pre-ringing. So say I process with a saturator that introduces harmonics - these harmonics alter the frequency response. If this plugin was to introduce delay, then when the delay was compensated for by the DAW, this is when we'd notice pre-ringing distortion since what was "post-ringing" is now aligned with the original transient. I'll admit, I may not be the best at explaining this, but here's a plugin designer/coder explaining how oversampling introduces pre-ringing: th-cam.com/video/l0CsAqCtZXo/w-d-xo.html Funny enough, this video was sent to me by someone after I claimed oversampling doesn't affect transients or introduce distortion.
Hey @richertz - I've made a quick video showing the relation between plugin latency and pre-ringing distortion: th-cam.com/video/D-P0u9h9vgA/w-d-xo.html It has what I feel is sufficient evidence to justify the claims I made in this video. Thanks for watching!
Happy you're back, I have some questions. I got the idea of splitting the master's mid and side into separate bus channels and using a transient shaper on the sides for more width and detail and heavy compression on the mids for more thickness then blending it into the original signal the collectively processing all 3. That technique has been working like a charm for me, are you saying I should stop doing that? Also I've been serial limiting with the pro L2 and ozone 10 maximizer (and using it's soft clip feature), is serial limiting something you don't do anymore?
Splitting the master up is okay, so long as you make sure the phase lines up. For example, don't use any minimum phase HP/LP filters. The act of blending a dry signal isn't going to cause any issues unless your DAW is terrible about handling latency compensation (which is rare nowadays). As for serial limiting: go for it! Sage actually used 2 maximizers in this video, so it's definitely something that's normal. Keep in mind that if you use 2 of the same limiter in series, it's usually the same thing as using just 1 witn more gain reduction. But stacking limiters with different behavior can be useful. In my experience, pro-L isn't afraid to handle transients with a little more grit, but ozone maximizer is smoother with handling the body of the sound, so it sounds like it would work well. IDK though, good results are good results. Sound is all that matters in mixing.
If it sounds good and you enjoy it I say keep doing it! In this video I just wanted to make something that sounded as clean as possible, but that doesn't mean it has to be the way you master. I change my approach often depending on what I'm trying to achieve/what sound best suits the genre and project.
Hi, thanks for your Always precious guidelines..You suggest to work at 192 kHz, usually the mix are at 48, so se needed to upsample, processi and the downsample, r8brain free Is enough good for this? Any cheap better alternatives ?
Thanks for watching! Mixing at 48kHz. should be fine - if you upsampled to 192kHz your computer would probably have issues running everything (at least mine would haha). So I'd say mix at 48kHz unless your computer can run higher sampling rates without issue.
The goal of this master was to be clean and maintain transient integrity. I focused on the peaks, which are at the same loudness due to peak normalization, to judge how much their character has changed, and whether it is for the better. It was an okay exercise given the context, imo, but I agree it would have been nice to include loudness normalizing in addition.
I usually avoid loudness normalization during these comparisons since I find it affects the timbre - I find peak normalization as a decent but not perfect alternative.
school of mastering - where can I go to get more info?.. ex. Is The School Accredited By Guitar Center, Sweetwater?.. and What If Enrollees Want To Pay Bi-Annually Instead Of Monthly?
We're making a trailer now to give more info, but for the time being you can learn more about it at our website: SageAudio.com As for accreditation, no, these companies haven't given the course one. Currently, we have a monthly subscription service, but reach out to our team here and maybe they can make an accommodation and also answer any additional questions you may have: admin@sageaudio.com
Hello, just wanted to let you know that with the Weiss MM-1 you aren’t introducing any “amount” of processing until you are reducing gain. I am pretty sure you are only making it louder with level and not maximizing with the plugin. Just my two cents.
@@sageaudio I believe it’s in the user guide, it refers to the function of the amount knob as being processing applied right before the brick-wall limiter once the signal is attenuated
Funny, I use the MM-1 subtly on vocals without hitting the limiter. And while I agree with @jonnyoh4731 (it may just be making it louder) it does so in very different ways depending on what setting you have it on. Almost like it is simply boosting the low level information with different EQ curves, but in a nice way. I also use the Loud setting as it has an obvious mid boost, which fills/fattens up vocals and also full tracks, as demonstrated here.
@@digitalproducer3500 there’s a fair chance the plugin is doing some reduction that doesn’t get put up on the meter. Have you ever noticed this effect when you have the input really low and you can visibly tell the signal isn’t anywhere near the threshold?
@@jonnyoh4731 Yep, completely agree. I think what is going on behind the scenes and what shows up visually aren't really on the same page. That said, I like what it's doing, especially on male vocals when mixing back in at a small (40ish) percent. Never use it for limiting myself.
Sage audio can you please make a video about how to get that stock fruity soft clipper sound i spent 4 hours today just to recreate that in studio one but i couldn’t get that. the plugin has a different vibe and sound please make a video spesific about it because that sound is a standart for trap music ❤
finally someone adresses Preringing as "not so good sounding". i always dislike the dull, soft pillowy sound of pre ringing, be it so subtle.. yeah, nice vide
Thanks for watching! Yeah I noticed whenever I used lower latency plugins I enjoyed the sound more - makes sense that it's affecting the signal is some way with some behind the scenes filtering
To avoid pre-ringing distortion caused by oversampling. Increasing the sampling rate of the session will reduce aliasing (not as much as oversampling but will still help). Thanks for watching!
Cool video. My question is what do you think about the emergent AI mastering assistance tools like the new Ozone 11? And is it worth it to spend the time learning about all this information and buying courses when AI is getting better and better at helping producers and engineers do the job in much less time and more efficiently?
My two bits: AI has less context, since it can't "hear" per se. It can recognize audio patterns and emulate preexisting sounds, but a human can make decisions based on emotional impact. Mastering is about getting new perspective on a track and making adjustments to translate the message best for the listener. Right now, AI is for ideas, but humans are for decisions and big picture. Ozone 11 works great because the processing responds to the type of material it's processing, but the human still needs to be there to carefully decide whether it's improving or worsening the sound.
Thanks for the question - it's a really interesting one I don't get to talk about too often. I agree with what @ampersand64 said - for the time being, it seems like AI can recognize patterns, generate solutions, but whether or not those solutions best fit into an emotional context is hard to say/something AI will have to get better at over time. Weird example, but if you're familiar with the Grand Theft Auto series, when they remade they're original games a year ago or so, they used AI to upscale, re-animate character models, improve textures, etc. The problem was, it altered the art style so much that it no longer fit the tone of the games or story. I think a similar situation is occurring with mastering and music production at the moment - not to say that AI assistance sounds bad, but just that it doesn't always best encapsulate the feeling of something, since it's such a complex and difficult component to fully understand. Also, a lot of the products being marketed as AI right now aren't actual machine-learning, but some form of meticulously coded algorithm (I can't say what Ozone is doing exactly though, so some AI component may actually be at play here). So there isn't something behind the scenes that will actually learn and grow and alter it's behavior over time, but instead a complex set of 'if-then' statements that can intelligently detect various aspects of a mix and then adjust to mimic a pre-designed model of a particular genre. That said, as AI and AI mastering advances (still relatively new, at least commercially speaking), I think we can expect to hear better and better masters made with AI, or an engineer with the assistance of AI. I don't think it'll replace traditional mastering, but will eventually be recognized as something that exists in tandem with it, as an additional option for mastering and production.
@@sageaudio Yes, I know a lot of people are giving AI too much credit these days, particularly the doomsday people... Anyway, I also think AI assistance is here to stay, not to replace everything but to assist those who already know the craft. AI is good to build the frame but humans are to give the touch of uniqueness, or soul. I was watching a video about most people no longer knowing to differentiate singers who use autotune and pitch correction from those who don't, great video on Wings of Pegasus channel titled Nobody knows what music is any more. The conclusion for me is the same, in the end only humans can afford to choose making "wrong" things, to give that uniqueness that makes things sound better.
I don't believe there will ever be a one button AI that decides and completes a mastering by itself, it will always require human input no matter what. But for sure it will remove a lot of steps in the process and it will require less and less technical knowledge from anyone to be able to deliver a solid great mastering. In that sense, I believe a lot of the courses on mastering we have today will have to adapt, specially those who focus on technicalities. Since AI is still in its infancy, the majority of technical issues with mastering will soon just be easily dealt with AI automatically at the first analysis, and the mastering guy will just have to paint the bigger picture, decide what kind of feeling he wants to transmit with the track, like wanting more punch in the kick, more width overall, a more dreamy or a more dark tone. All the technicalities and steps on how to do it manually will be nulled with AI. Guides and courses on mastering will need to be more like teaching how to feel instead of teaching how to reach that feeling.
The project sample rate only matters because of aliasing. Sample rate determines the highest frequency that the digital file can reproduce. So at 192khz, the highest frequency is 96khz. At 48khz, it's 24khz. Aliasing is what happens when you distort a digital signal and the added harmonics exceed the highest frequency that the digital file can reproduce. So at 48khz, if you distort cymbals, the harmonics may be higher than 24khz, and they'll reflect back into the spectrum and sound metallic, harsh, grungy. Typically, the solution is to run the project at 44.1 or 48khz, and use oversampling for any distortion. Oversampling simply make the plugin's sample rate higher, then runs the plugin processes (such as distortion or compression), lowpass filters the output (to remove frequencies higher than the project's maximum), then downsamples it back to the project samplerate. Oversampling can be more CPU efficient, because linear processes (that don't add harmonics) can run at a lower sample rate. Fewer samples means fewer processor cycles, so saving CPU with these processes with no change in quality is desirable. Then you can use oversampling on any dynamics or distortion processes to keep aliasing minimal.
@@ampersand64 sorry to but in but I just have to say that it's nice to have someone who actually understands the fundementals of DSP, it's tough going out here, fight the good fight 😆
@@ampersand64 well you may say that but everything you said was correct, trust me that’s a lot more than most, I’m in the same boat as you, no matter how far you get and how much you think you know there’s always more to learn in audio at some points you’ll wish you picked a less complex subject like rocket science 😂
@@sageaudio love this. 15 years using logic and always learning cool new tips! Thank you for sharing. Loving your channel. Only just discovered so thrilled to see how much great content you have made to watch!
Why would you need a clean master it won't translate very well on all devices, specially if avoiding saturation. However great idea to increase audio res for more dynamics.
It's just an option! If you want to make a super-clean sounding master, maybe one designed primarily for hi-fi devices/playback systems then this is a good route to take. But like with most things in audio, it just depends on what you're trying to achieve.
Thanks for watching - the positive value emulates tube saturation a bit more, and the negative traditional distortion, with the 0 curve value being a blend of the 2. If it sounds good I say use it! I like it on a lot of masters - but for this session I was curious as to how to create the cleanest sound possible.
@Sage Guys...this seems to contradict your tutorial from two years ago,namely How To Master Loud Without Distortion in terms of a)the use of true peak limiting and b)the use of multiple serial limiters.I'm sure the advice imparted here is absolutely valid given you're the professionals,however for those of us totally home schooled with YT,this common phenomenon of seemingly diametrically opposing advice gets a bit confusing.Unless they are designed for different mastering situations perhaps?
Thanks for watching both of these videos! The info both videos are simply things to consider - not a declaration of how you should master (there are so many ways to approach audio production, your personal and subjective thoughts and wants for a project will always play a role). That said, this video was about approaching mastering from a 'how do I alter the original sound as little as possible while still making it loud enough to be enjoyed on other devices' perspective. The How to Master Loud without Distortion video came from a 'How Do I Get this to a super loud level without it sounding bad' perspective. So they're just highlighting different thought processes for different applications.
For next round, please put in why your making those eq changes to this specific track. Like you putting in that 2db bass eq boost because it sounds thin in that area, you said it for the mid about the vocal, but adding any bass in mastering on tracks that have good bass already or bad muddy bass(which for 99% of home studio operators you most likely to have problems with)it can be confusing! As most will not have the ear for it, as they've never worked for 10yrs in a proper room with proper equipment, so have never even heard before how something SHOULD sound, let alone know that you can't even hear those things with it being canceled in there room to begin with, due to the nulls inherent in their room!
This makes sense! I also hope the demo can justify why certain filters were chosen, but I agree some more explanation would be helpful. Thanks for the comment!
Thanks for sharing your thoughts! I've addressed this concern to some other people that watched, but in short, I don't want to normalize the loudness of the original mix, since that would alter its timbre and potentially require peak down attenuation. So, I choose to peak-normalize both to get them a little closer and without altering the timbre of either.
In this session I was attempting to minimize the effect of pre-ringing distortion, which is created by most oversampling filters since most utilize linear phase. To counteract aliasing, I raised the sampling rate, which doesn't work as well as oversampling, but seemed to be a good mid-point between lower levels of aliasing and lower levels of pre-ringing.
Thank you for making good videos, but sometimes it seems to me like your theoretical reasonings are quite out there. 1. Delay compensation causing noise? If the DAW compensates for the delay there should be no phase problems since this would happen before playback. The only way there could be problems like that is if you run a wet/dry mix that isn't compensated at all. If the plugin manual says it's compensated there is no reason to watch out for this. With your reasoning using analog gear would always be super detrimental since the delay in sending and receiving audio is much greater than even some of the heavier plugins. Why then are all the pros doing it? 2. Introducing 7hz in your mix would definitely not be audible, but your compressor and/or saturation post would react to it so it would alter the sound for sure. That is probably more likely due to every other aspect of the plugin you introduced that's audibly colored, not the 7hz.
Thanks for watching! I had a lot of questions/concerns with this claim so I made a video showing the relationship: th-cam.com/video/D-P0u9h9vgA/w-d-xo.html You're right! 7Hz is not audible, but it also introduces harmonics of 7Hz that reach into the perceivable range. When adjusting the parameters of this function during the demo, you can hear the difference.
Thanks for watching! One thing I learned after making the video is that if the plugin introduces lookahead, you'll get latency but won't encounter the type of distortion talked about here. That said, if the plugin introduces latency, it's very likely you'll get some distortion due to some behind-the-scenes filters.
Wow. I’ve only had one other engineer tell me about introducing planetary rhythms in mixing/mastering, super fascinating but thought I should understand a proper workflow before trying unethical ways of working. I’m definitely going to look into becoming a member, I look forward to learning from you
It is super interesting! I'm going to look a little more into it to see if it has any real applications in audio. Thought it was cool enough to bring up nonetheless. Looking forward to having you as a member! Here's the link in case it's helpful: SageAudio.com
When I look at the 0:51 second video, I see a text box showing the number of latency samples caused by the plug-in, which is not the case in my logic. Does it only come out in the latest version of Logic? I use logic 10.6.3 version.
reasoning for latency, even if latency compensation introduced artifacts, there is 0 latency compensation while bouncing and so there would be no introduced artifacts lol
Thanks for watching! The compensation is happening prior to exporting the signal - the DAW does this automatically during playback. If the plugin introduces latency, it likely includes a linear phase filter, linear phase oversampling, or convolution, which will cause pre-ringing distortion. Out of curiosity, what do you find false about the parallel processing aspect?
If all the audio stems are in 48k is it really worth mastering at a higher sample rate? Also, if that is so, aren't the benefits going to be lost when converting the master back down to a commercial sample rate?
It is! If you upsample them from 48kHz to a higher sampling rate, you're still less likely to get aliasing distortion. The benefits won't be lost when converting back down - the aliasing occurs due to processing you add, like high frequency saturators. If you avoid aliasing by giving these plugins more room to work with, then by the time you convert the sampling rate back down, no, or less aliasing will be present.
Yes due to an anti-aliasing filter. Odds are though, if the plugin introduces latency, it's due to linear phase filters, oversampling, or convolution which will cause pre-ringing.
I’ve heard from streaky that you should take the channel linking off on the L2 in masters. Also I always thought I should be using oversampling on that plugin for the master. Am I just totally wrong??
For channel linking it depends - if you want the left and right channels to affect the signal independently, then de-link. This will cause a slightly wider sound and result in less limiting overall, but at the expense of mono compatibility. As for oversampling, I wanted to avoid pre-ringing that comes with it. Since oversampling is used to reduce aliasing distortion, I instead increased the sampling rate of the session, which will have a similar result.
Right now we're seeing what's most feasible for everyone - it currently costs $95 for a biannual membership. This gets you access to our School of Mastering, our mastering services at a 50% discount, 1 free master from our head engineer, lots of additional perks, and any and all educational courses we release in the future.
@@sageaudio great! I have made a account and just waiting on your reply to finish the process. That is a great deal and I'll most definitely be enrolling. And all the courses will be all on line courses, right?
Honestly I think this is a bad, unartistic approach to the mastering, which can be easily replaced by AI mastering available in Ozone etc. You'll loose all the EDM loudness wars by doing this. If your vision is to have sound and loudness contour like Skrillex or Deadmau5, you have to forget about this and focus on how about to squeeze your master as much as possible to target approx -4LUFS while STILL having the clean sound. This is not a clean master, this is a safe master. The only right thing I can relate to in the video is the chain sequence.
No worries! I wouldn't recommend mastering like this in all occasions. Like you said, if you wanted to make something -4 LUFS, this is not the route to take.
I always like your free plugin alternative suggestions, but ... What's a good free modern maximizer plugin? The best one i currently have is very old and only 32bit, so can't use it on my newer machine...
I like MCompressor a lot by Melda Audio - if you use the 'Custom Shape' function you can maximize the signal. If i'm remembering it correctly, you can even change the image that's affected - so you could maximize just the mid image if you wanted to, etc.
I'd be curious to hear why you'd avoid oversampling. Is it purely a pre-ringing concern? Granted, at a 192kHz sample rate, your nyquist frequency is going to be way above human hearing and you're not likely to run into any cramping issues, but still, oversampling will increase the anti-aliasing filter frequency and therefore further reduce aliasing and also quantization distortion (but of course, at 192Khz, we're probably bumping into maximum sample rates of individual plugins, perhaps). Also, something I don't think you mentioned, but is implicit in your assertion to work within a 192kHz session is that you are employing SRC in your DAW for mixes printed at lower sample rates. While many modern DAWs have perfectly acceptable SRC, it still isn't necessarily a given that introducing this DAW SRC is desirable in all cases. For example, in a pitch/catch scenario, it would probably be desirable for the pitching computer to use the native sample rate of the printed mix. Generally, I understand what you are aiming at with your message here, and feel as though this video was probably aimed at those doing their own mastering, in which case, I don't really have any objections to anything you've said, I just feel as though in mastering, there are almost no "always" things. It is always a balance between which types of distortion we're introducing. After all, everything we do in mastering is "distortion," by definition, at least under the broad definition that we are distorting the signal in hopefully desirable ways. 🙂
I hear what you're saying! As for oversampling, yeah it's a latency thing for me - granted, I don't always follow exactly what I've stated here, since sometimes aliasing is more noticeable/more of a problem than the mild affect oversampling has on transients. So it depends! Like you said, there are no "always" when mastering since we're combating some form of distortion at all times just about. Sorry I didn't mention it but yes - sample rate conversion is occurring if the original sampling rate was lower. But that gives me some food for thought - I might look more into how different DAWs perform SRC and see if there's any differences between them, or if they have any impact on the sound.
personally I'm more concered that he mentioned that the delay compensation of your DAW causes pre-ringing? this is just false. also your correct in what you say, oversampling is just an obtimised version of running at that native sample rate, by only running at that sample rate for the processes that could produce aliasing for instance, if your compressor has an oversampling feature, most likely it's only oversampling the gain reduction circuit and not the linear input and output gain sliders.
@@memeswillneverdie Yeah, this stuff gets super nuanced, and even with fairly advanced measurement tools, there's a lot of deduction involved in the specific internals of any given plugin. Plugin makers almost never get as specific as is necessary to really understand what's happening under the covers in their documentation. Although, admittedly, that does stray into intellectual property territory.
@@sageaudio What effect does oversampling have on transients? It's basically the same process that the D/A converter on your output goes through, right? Some transients will be higher because of intersample peaks. But those peaks exist in the analog output, regardless of the digital samples' arrangement.
@@ampersand64 I watched the video Sage linked to in other comments. It illustrates the ringing / smearing of transients that OS introduces. The concern isnt when it happens once or twice. It's the stacking of multiple stages of this that becomes audibly detrimental eventually. Hence the session at 192 and no OS vs 48kHz and several OS stages.
Your thing about pre ringing is just wrong, pre-ringing is an artefact caused through linear phase filters, linear phase filters do cause extra latency but this is not reversible, just because there is extra latency doesn’t mean pre-ringing? DAWs have built in delay compensation, they way that works is just by delaying the other tracks to mean they all arrive at the same time.
The true peak limiting thing is just a myth, true peak limiting is simply just a more accurate gain reduction since all its doing is interpolating sample point in between the existing sample points, this does not effect the gain reduction circuit other than making it more accurate, the only reason you could ever get a different “sound” is because when you turn true peak limiting on more gain reduction is happening because transients that weren’t being detected before are now being detected, turn the input gain down and it’ll null.
Also the way you’ve set your limiter is basically like a clipper, not having oversampling will allow for aliasing (since as previously mentioned your basically clipping the signal and clipping results in an infinite series of harmonics being created which will fold back when they hit nyquist)
Whether aliasing is something you need to be concerned about is another topic but by definition is is not “clean”
Also you peak matched the results, this is not the way you level match a master, you match it through loudness normalisation to make sure your actually hearing the difference and not just that one is louder.
I’ve noticed a lot of just flat out falsehoods on this channel, like the “cut everything under 30hz for extra headroom” thing or “put your low end in mono” etc etc
I don’t have anything against the channel, you’ve got a good message but c’mon, you can’t claim to be an expert in mastering and teach others when you don’t understand how delay compensation works?
If it worked like that then every single session would have tonnes of pre-ringing over every track and you’d see it in the waveforms when bounced from once session to another, or do you think DAW developers were smart enough to just do it the same way speakers are done in live sound, you delay one signal so that they all arrive in phase with each other.
Also who cares about delay, if the Mixed Stereo-Signal is delayed completely. It only Mathers when diffrent parts of spectrum gets delayed. I Think most thinks in mastering are overhyped. Ther are tons of Tracks that everybody love that could have been done better.
@@moritzschiekel2272 exactly that problem was solved literally decades ago, why would DAW developed choose to use a way that is more complicated and actively damages the fidelity of the audio instead of just delaying everything else so it's in sync, I don't know where they pulled that from, that's a new one for me.
@@dussonthephanter6698 as I said in my original comment, this was things I noticed in the past from this channel, not that they are currently doing it.
I agree that the way to teach something like mastering is to teach the fundementals on an advanced level, understanding what EQ's are actually doing to your signal, what compressors are and what they do etc.
audio theory is an insanely deep and often confusing thing (I personally love it, it's a whole world of knowledge and the depth seems to never end)
theres a bigger problem at play and it's patently false information being perpetuated by even some of the biggest names, when I was starting out you just go "yeah that sounds about right" because you don't know enough to question it properly, when you start learning a bit more about audio and getting more experience you see these things and go "what the hell are these guys on about".
again as I mentioned previously, I don't think someone can claim to be an expert in mastering and teach others when they make such fundemental mistakes like not knowing how delay compensation works. to me it just sounds like they found something cool like pre-ringing (which is a facinating subject that you could spend hours on) then just sprinkling it in to sound smart without actually knowing what it is and how it works. be skeptical always when it comes to this stuff, go and double check all the information your getting especially when your starting out becuase you'll be fooled by someones big studio and go "well how could he be wrong"
@@memeswillneverdie Also when u think in therm of Mixing u usually Master a Stereo File, so ther is only one Channel
@@moritzschiekel2272 exactly, it’s fundamentally bizarre why they would think that pre-ringing could occur on a single stereo track, I really think they heard the term “pre ringing” read a couple lines and it mentioned delay then just used the term pre-ringing without actually understanding what it is.
This video is misleading.
LATENCY ITSELF DOES NOT CAUSE PRE-RINGING! LINEAR PHASE FILTERING AND FFT PROCESSING CAUSE PRE-RINGING!
Oversampling can cause latency and pre-ringing. Pre-ringing is negligible when using linear phase oversampling filters anyway, because it only affects high frequencies that most people can't hear.
You can run linear phase crossover filters (for multiband processing) with ZERO pre-ringing, that have sizeable latency.
You can run a compressor with lookahead, where latency is necessary for the sound. No pre-ringing.
also importantly, pre-ringing does not (in itself) attenuate transients! It simply smears them across time.
Any sort of filter you put audio through (whether that be digital EQ, analog EQ, convolution, brickwall lowpasses during oversampling, or linear phase EQ) changes the dynamics by introducing ringing.
Minimum-phase filters cause ringing AFTER the signal, and linear phase filters cause ringing before AND after the signal.
Filter ringing is easiest to hear with transients, but it happens with slow dynamics just as much.
MINIMUM PHASE EQ DOESN'T SOUND "CLEANER" THAN LINEAR PHASE EQ!!! It's a subtle difference, and which you use depends on your goal.
Saturation doesn't sound dirty because it decreases the dynamic range. It sounds dirty because it adds harmonics to the signal (which "fills out" the spectrum) and causes the audio to intermodulate (like when a big sub-bass makes everything else grittier).
The goal of mastering isn't to preserve dynamic range, it's to enhance the emotional impact of music, while making it louder. Making something louder necessitates decreasing the dynamic range!
For this purpose, saturation is extremely useful. Hard clipping can be more transparent than limiting when taming peaks, and certain types of saturation can make music feel more detailed or punchy, even though it's technically dirtier.
On oversampling: it's probably not necessary during mastering IF you're using a high enough project sample rate. Oversampling becomes useful if you're using saturation in series (which is more common during mixing), and if that saturation creates higher harmonics (like hard clipping). But if you lowpass your mix after saturation, you'll probably never introduce aliasing.
If you're running at a low project samplerate, oversampling becomes crucial for any saturation processes, and some dynamics plugins too.
As for exciters, they should be alright to use if they're oversampled properly and aren't pushed too hard. Exciters and high frequency saturators can sound amazing on a full mix!
I'd recommend testing your nonlinear plugins to see if they creates aliasing, and to see how well their oversampling works. Running sine waves through it is useful, but it's equally useful to listen to the delta signal and see if it sounds metallic.
I'm happy to be corrected if I'm wrong.
Thanks for watching! Here's a good video that shows how oversampling causes pre-ringing: th-cam.com/video/l0CsAqCtZXo/w-d-xo.html
Paralell processing: don't be afraid. This video is once again giving bad advice because of its lack of detail.
It is simply not true that "all forms of parallel processing will negatively affect the signal in this way." Only filters, really. Keep reading.
If a plugin has a dry/wet knob built in, chances are the developers MADE SURE that it doesn't create phase cancellation. They're not stupid enough to include a "sound bad" knob on their interface.
So, why might blending in a parallel signal cause phase cancellation? Two reasons come to mind:
1. The plugin is introducing latency and not reporting it to your DAW. This almost never happens with decent plugins, even free ones. I dunno how latency reporting works, but I know that it has never failed for me. Don't worry about it unless you hear something wrong.
This rule applies for very short delays as well. But chances are, you're not adding delay effects during mastering.
2. The plugin is shifting the phase in some way. Unlike the previous example, phase shift is extremely common. Any minimum phase EQ or filter creates phase shift.
HOWEVER, unlike what the video said, EQ is NOT the worst culprit when it comes to phase cancellation. Regular bell and shelf filters phase shift proportional to the amount of cut/boost. So even crazy 18dB EQ moves would only result in a slight narrowing of the filter (higher Q setting, essentially), when mixed back into the dry signal.
The culprit for strange phase effects is high/lowpass filters. Filters 12dB/octave or steeper create notch filters when blended in paralell. So if you're running a filter in paralell, stick to 6dB/octave. Or simply fun linear phase EQ.
Oversampling uses veery steep lowpass filters. Sometimes they're linear phase, but often they're minimum phase. Minimum phase oversampling filters cause crazy phase shifts, so avoid using them in paralell.
Some analog modeled plugins will use IR filters or something, so always check the phase response when running in paralell.
"How do I tell which type of oversampling filter my plugin is using?" Bertom EQ analyzer is a free VST that'll tell you the frequency response and phase response of any plugin or chain. If you click the "show phase" button, and see that the phase is at 180 degrees at any point in the spectrum, it's gonna cause phase issues.
Notice what I didn't mention? Dynamics and saturation. These don't shift the phase, so you can feel free to blend any limited, compressed, gated, expanded, saturated, clipped, ring modulated, or otherwise processed signals in parallel.
Quick correction: 8:01
CLIPPING IS SATURATION!
Saturation is a term used to describe any waveshaping distortion that has a hard limit of amplitude. This describes digital hard and soft clipping perfectly!
I agree that soft clipping is often more transparent than compression or other types of saturation when reducing peaks. Indeed, many saturators "color" the signal without decreasing dynamic range hardly at all (such as oxford inflator).
Another quick correction: maximizer vs. limiter
They're basically the same thing. A limiter is a compressor that keeps a signal below a certain threshold (it has a ration of infinity to 1).
A maximizer is a tool that uses limiting, but can also include other things like clipping, or limiters in series. Pro-L2 does both of these, and is therefore a maximizer.
So while this video makes them look like different utilities, they're essentially the same thing.
THIRD CORRECTION: 12:24
The P440 EQ had saturation in its algorithm. While this isn't harmful misinformation, it does kind of obscure the truth. "
also really man? the shumann frequency? Its FOURTH harmonic is 33.8 hz, which is below what most consumer speakers can reproduce. It's also basically subsonic. Also also the actual schumann resonance changes over time.
Watch this video by Benn Jordan (a.k.a. "The Flashbulb") for more info: th-cam.com/video/Q3LVijzZAe4/w-d-xo.htmlsi=_hg9ouuwo1BSoWGJ
What's my point? Sage audio, if you're reading this: Please be specific with your terminology. Please explain your reasoning fully. Please stop spreading misinformation.
You're clearly a channel aimed at audio beginners. Make sure they come into the profession with a nuanced, SCIENTIFIC understanding of audio processing. There's plenty of good advice here, and your results speak for themselves. But I dislike seeing so many incorrect or misleading claims.
@@sageaudio Ahh, this is great knowledge! So oversampling is like a filter that creates ringing at the nyquist frequency. Only FIR oversampling filters create pre-ringing (like FIR EQ filters).
I'll update my comment to be more accurate. Glad to be learning more about audio.
The smearing effect in particular is more noticeable than I thought it was. The length of filter ringing, as I suspected, was directly correlated with the order (steepness) of the filter. He also mentioned that the oversampling process uses allpass filters tuned to a high frequency?
Still though, the oversampling itself only causes ringing at high frequencies, just like the lowpass filter associated with oversampling.
I don't know about you, but my young ears can't hear anything very clearly above 16khz. So is anyone really gonna notice a 22khz ringing? I personally doubt it.
In my opinion, transient smearing at 22khz is easily preferable to aliasing on transients. Although both artifacts are subtle, one is clearly preferable.
Very notable, too, is that the IIR oversampling created latency without any pre-ringing. Pre ringing is not caused by latency, as your video suggests. In fact, there's no examples of pre-ringing (which can definitely occur in some oversampling algorithms) in the video you linked.
Plus, these artifacts are almost guaranteed to be less noticeable than aliasing. That's not relevant to clean mastering, but it's probably useful knowledge for the average mix or mastering engineer.
To summarize: oversampling is a useful tool for dealing with digital distortion, and its adverse effects are generally overstated.
Latency does not necessarily mean pre-ringing etc... Sometimes it's for looking ahead of the waveform for better peak detection algorithms.
Thanks for watching! You're right! I conflated correlation with causation in this instance. That said, latency is a really reliable indicator of behind-the-scenes filtering that will cause pre-ringing, such as oversampling, linear phase filters, or convolution.
Wow. That was an amazing approach to teaching mastering. Best I’ve seen so far. I don’t really want to do my own mastering - just understand it better. Thank you!
That's awesome thank you! If you ever need a track mastered, please let us know - SageAudio.com
@@sageaudio Will do. Thanks again
I've been using Newfangled Audio's Elevate Mastering Bundle for a while now and I'm really pleased with the results. The dynamic control and EQ control is very handy. Great video!
Thanks for watching! I like they're plugins too!
The best video around clean mastering I've ever seen sir! You really explained the most detailed things that others might not even consider telling! Definately helped understanding the fundamentals of achieving a clean sound in general! Thank you for this!!!
Thank you for watching and it's great to hear you enjoyed it! I hope the video helps or at the very least gives you some things to consider when you're working on your next session
ive noticed you dissapereard on here a little. I think your way of teaching beats all.
That's really kind I appreciate it! Happy to be making videos again
@@sageaudio u make my life make sense thanks for coming back
this channel is a gold mine i love the way you teach
Thank you! Great to hear you like the channel
Nice video - Learned about the 192Khz ringing (which people avoid) and about the Soft Knew transients. I also heard some metallic ringing from snare on final master samples, specifically on FF L2. Not that It was bad but it changed the tone and I liked the Pre-FF L2 better. - But it is subjectively
Thanks for sharing your thoughts - I appreciate it!
Love these guys. Quick and to the point.
Usted es un maestro del master! Master of the Universe! Saludos desde Argentina. Gracias
@sageaudio wrote: "Any instance in which the frequency response of a signal is changed (which occurs with all of the processors I'm using in this chain) and latency is present, the reallocation of the delayed signal will shift artifacts or in this case post ringing, back to the source of this change, in turn making it pre-ringing - its really just added signal that occurs after the change which now has to be shifted back in order for the full signal to be in time/compensated."
As a simple example, if you had a plugin (call it A) that did nothing but delay a signal, it would work by allocating some samples of memory, and storing the samples it sees for a set number of samples, then outputting after the appropriate delay (i.e. the number of samples that equate to the given delay based on the sample rate). It doesn't "do" anything with the signal to reconstruct it - it literally stores the incoming sample, then outputs it verbatim a bit later.
Now, what if you had a plugin (call it B) before the plugin (A) that had no latency, but did some processing, such as the ones you mention in the video. In this case, we agree that (B) would add no pre-ringing to the signal, right? And we've already agreed that plugin (A) doesn't have any pre-ringing as it's literally outputting the samples that come in verbatim. So, the chain of (B) then (A) has no pre-ringing.
So, what is we build a new plugin that does exactly what (B) then (A) does, all within the same plugin. That would have no pre-ringing, right?
While it's certainly possible that there are some plugins (and DAWs?) that don't handle latency properly, the above proves that the claim that any plugin (or even any plugin in which the frequency response of the signal is changed") that has delay brings artifacts due to that delay is false.
If you disagree with this (IMO clear) counter example, I have never heard this argument made anywhere other than this channel. Can you provide some links to other information out there that corroborates it?
(I posted this in a thread below, but wanted to promote to top level.)
I understood the premise to be that latency is an indicator that the process is likely using oversampling. The goal is to avoid stacking multiple stages of oversampling to maintain transient integrity, so it's logical to avoid processors with latency.
@@sparella what you say is logical, however, that is not the point that this video makes. I am specifically arguing against the point that latency itself causes pre-ringing, which is clearly false.
Hey @dyrossmusic - here's a video I just released showing the relation between plugin latency on pre-ringing: th-cam.com/video/D-P0u9h9vgA/w-d-xo.html
@@sageaudio you didn't demonstrate what you think you did!
The king is back! Can you please make an in-depth video of how you would mix/master a rock sog? I usualy mix/master trap /rap songs and i got a project that is a bit out of my confort zone and i will appreciate if you are willing to make a video on this topic!
Thank you in advance!
Thanks for watching! A video like that is included in our new School of Mastering (we go over creating pop, rock, hip-hop, classical, jazz, electronic music, singer-songwriter, and metal chains), so I probably won't replicate it here, but here's a link in case you're interested: SageAudio.com
the only channel I truly trust for audio advice
Thank you so much! Dan Worrall is another great option if you're looking to expand into other channels.
Best mastering tutorial on TH-cam, best Mixing and Mastering engineer/sound technician 🔥🔥🔥🔥🔥
Thanks Sage Audio I, learned a lot from this, you also use new plugins too, expect the FabFilter Pro-L2 limiter
It's a nice limiter! Get your hands on it!
Thanks for watching - great to hear you enjoyed this one!
Ok so I have thought about the latency issue and I have to agree with some comments below, please can you provide evidence. I enjoyed watching your videos butt his one needs clearing up as I have some big doubts about the truth to this statement. Pro Tools did used to have latency issues, and possibly some bugs can exist with this one. I think that needs further proof as I don't trust this - sorry but I am open to evidence if you have some.
I checked out your website, but it just wants me to pay to sign up? Think that could do with a little more work to explain the benefits.
Thanks for watching and sharing your thoughts! Any situation in which delay is compensated for, and the frequency response of a signal changes, will cause some form of pre-ringing. So say I process with a saturator that introduces harmonics - these harmonics alter the frequency response. If this plugin was to introduce delay, then when the delay was compensated for by the DAW, this is when we'd notice pre-ringing distortion since what was "post-ringing" is now aligned with the original transient. I'll admit, I may not be the best at explaining this, but here's a plugin designer/coder explaining how oversampling introduces pre-ringing: th-cam.com/video/l0CsAqCtZXo/w-d-xo.html
Funny enough, this video was sent to me by someone after I claimed oversampling doesn't affect transients or introduce distortion.
th-cam.com/video/l0CsAqCtZXo/w-d-xo.html
That’s applies only if the plugin uses linear phase processing. @@sageaudio
Hey @richertz - I've made a quick video showing the relation between plugin latency and pre-ringing distortion: th-cam.com/video/D-P0u9h9vgA/w-d-xo.html
It has what I feel is sufficient evidence to justify the claims I made in this video. Thanks for watching!
Thank you for your time making this I’ll watch and add some thoughts after as I’m sure it will be quite interesting.
pulsar modular makes incredible plugins!!!
They do! Very cool stuff
I love you my teacher
Thanks for watching! Great to hear you enjoy the channel
AMAZING!!!
Thanks for watching!
Happy you're back, I have some questions. I got the idea of splitting the master's mid and side into separate bus channels and using a transient shaper on the sides for more width and detail and heavy compression on the mids for more thickness then blending it into the original signal the collectively processing all 3. That technique has been working like a charm for me, are you saying I should stop doing that?
Also I've been serial limiting with the pro L2 and ozone 10 maximizer (and using it's soft clip feature), is serial limiting something you don't do anymore?
Splitting the master up is okay, so long as you make sure the phase lines up. For example, don't use any minimum phase HP/LP filters. The act of blending a dry signal isn't going to cause any issues unless your DAW is terrible about handling latency compensation (which is rare nowadays).
As for serial limiting: go for it! Sage actually used 2 maximizers in this video, so it's definitely something that's normal. Keep in mind that if you use 2 of the same limiter in series, it's usually the same thing as using just 1 witn more gain reduction.
But stacking limiters with different behavior can be useful. In my experience, pro-L isn't afraid to handle transients with a little more grit, but ozone maximizer is smoother with handling the body of the sound, so it sounds like it would work well. IDK though, good results are good results. Sound is all that matters in mixing.
If it sounds good and you enjoy it I say keep doing it! In this video I just wanted to make something that sounded as clean as possible, but that doesn't mean it has to be the way you master. I change my approach often depending on what I'm trying to achieve/what sound best suits the genre and project.
Hi, thanks for your Always precious guidelines..You suggest to work at 192 kHz, usually the mix are at 48, so se needed to upsample, processi and the downsample, r8brain free Is enough good for this? Any cheap better alternatives ?
Thanks for watching! Mixing at 48kHz. should be fine - if you upsampled to 192kHz your computer would probably have issues running everything (at least mine would haha). So I'd say mix at 48kHz unless your computer can run higher sampling rates without issue.
Please make a series mixing different instruments with the artiria fx plugins. Thanks as always
Thanks for watching! I'll use Arturia more in the future :] they make great stuff
6:58 - the esoteric plugin! i like it lol
Haha same here!
I’ve used Sage on many of my commercial releases 🤩💯
That's awesome, thank you for watching!
You're back, good!
Thanks for watching! Glad to be back
confused about the peak normalization via comparison at the end; one is way louder still so of course it sounds better.
The goal of this master was to be clean and maintain transient integrity. I focused on the peaks, which are at the same loudness due to peak normalization, to judge how much their character has changed, and whether it is for the better. It was an okay exercise given the context, imo, but I agree it would have been nice to include loudness normalizing in addition.
I usually avoid loudness normalization during these comparisons since I find it affects the timbre - I find peak normalization as a decent but not perfect alternative.
Bro just one more question, I have the weiss EQ-MP is there any difference in me using it instead of the EQ1?
hi sage, in my case my audio interface is only 48khz, can i still change the sample rate to 192 finish the master and export to wav at 192?
school of mastering - where can I go to get more info?.. ex. Is The School Accredited By Guitar Center, Sweetwater?.. and What If Enrollees Want To Pay Bi-Annually Instead Of Monthly?
We're making a trailer now to give more info, but for the time being you can learn more about it at our website: SageAudio.com
As for accreditation, no, these companies haven't given the course one.
Currently, we have a monthly subscription service, but reach out to our team here and maybe they can make an accommodation and also answer any additional questions you may have: admin@sageaudio.com
Who is the singer/artist?
Thanks for sharing!
Of course, thanks for watching!
what is the reason to not use oversampling on the limiter ?
To avoid pre-ringing distortion. Thanks for watching!
what is pre-ringing distortion ? I've not come across that one before@@sageaudio
Hello, just wanted to let you know that with the Weiss MM-1 you aren’t introducing any “amount” of processing until you are reducing gain. I am pretty sure you are only making it louder with level and not maximizing with the plugin. Just my two cents.
Thanks! I tried testing it with plugin doctor but it's crashing on my end. If you get a chance to look into this some more please let me know!
@@sageaudio I believe it’s in the user guide, it refers to the function of the amount knob as being processing applied right before the brick-wall limiter once the signal is attenuated
Funny, I use the MM-1 subtly on vocals without hitting the limiter. And while I agree with @jonnyoh4731 (it may just be making it louder) it does so in very different ways depending on what setting you have it on. Almost like it is simply boosting the low level information with different EQ curves, but in a nice way. I also use the Loud setting as it has an obvious mid boost, which fills/fattens up vocals and also full tracks, as demonstrated here.
@@digitalproducer3500 there’s a fair chance the plugin is doing some reduction that doesn’t get put up on the meter. Have you ever noticed this effect when you have the input really low and you can visibly tell the signal isn’t anywhere near the threshold?
@@jonnyoh4731 Yep, completely agree. I think what is going on behind the scenes and what shows up visually aren't really on the same page. That said, I like what it's doing, especially on male vocals when mixing back in at a small (40ish) percent. Never use it for limiting myself.
Glad you are back!
Thanks! Me too!
Sage audio can you please make a video about how to get that stock fruity soft clipper sound i spent 4 hours today just to recreate that in studio one but i couldn’t get that. the plugin has a different vibe and sound please make a video spesific about it because that sound is a standart for trap music ❤
I'll take a listen to it and see if I can figure out what it's doing behind the scenes! I could add something about it to a future video
@@sageaudio tysm because i did a lot of testing but couldn’t figure it out
finally someone adresses Preringing as "not so good sounding".
i always dislike the dull, soft pillowy sound of pre ringing, be it so subtle..
yeah, nice vide
Thanks for watching! Yeah I noticed whenever I used lower latency plugins I enjoyed the sound more - makes sense that it's affecting the signal is some way with some behind the scenes filtering
@@sageaudio i dont know whether or not its filtering involved? What else could it be, hmm.
so if i produce at 48khz i should export on 192khz for master, after mix at 48khz?
Upsampling to 192kHz during your mix's bounce/export should help alleviate some issues during mastering, so yes!
@@sageaudio so produce al 48khz, export stems at 48khz, and on pro tools upscaling to 192 and then export the final master at 192 also? or 44100 24bit
What's the reason for avoiding oversampling on the Pro L2?
To avoid pre-ringing distortion caused by oversampling. Increasing the sampling rate of the session will reduce aliasing (not as much as oversampling but will still help). Thanks for watching!
Cool video. My question is what do you think about the emergent AI mastering assistance tools like the new Ozone 11? And is it worth it to spend the time learning about all this information and buying courses when AI is getting better and better at helping producers and engineers do the job in much less time and more efficiently?
My two bits:
AI has less context, since it can't "hear" per se. It can recognize audio patterns and emulate preexisting sounds, but a human can make decisions based on emotional impact.
Mastering is about getting new perspective on a track and making adjustments to translate the message best for the listener.
Right now, AI is for ideas, but humans are for decisions and big picture.
Ozone 11 works great because the processing responds to the type of material it's processing, but the human still needs to be there to carefully decide whether it's improving or worsening the sound.
@@ampersand64 Humm. I like that answer and agree, haven't thought about that perspective. Thank you!
Thanks for the question - it's a really interesting one I don't get to talk about too often. I agree with what @ampersand64 said - for the time being, it seems like AI can recognize patterns, generate solutions, but whether or not those solutions best fit into an emotional context is hard to say/something AI will have to get better at over time.
Weird example, but if you're familiar with the Grand Theft Auto series, when they remade they're original games a year ago or so, they used AI to upscale, re-animate character models, improve textures, etc. The problem was, it altered the art style so much that it no longer fit the tone of the games or story.
I think a similar situation is occurring with mastering and music production at the moment - not to say that AI assistance sounds bad, but just that it doesn't always best encapsulate the feeling of something, since it's such a complex and difficult component to fully understand.
Also, a lot of the products being marketed as AI right now aren't actual machine-learning, but some form of meticulously coded algorithm (I can't say what Ozone is doing exactly though, so some AI component may actually be at play here). So there isn't something behind the scenes that will actually learn and grow and alter it's behavior over time, but instead a complex set of 'if-then' statements that can intelligently detect various aspects of a mix and then adjust to mimic a pre-designed model of a particular genre.
That said, as AI and AI mastering advances (still relatively new, at least commercially speaking), I think we can expect to hear better and better masters made with AI, or an engineer with the assistance of AI. I don't think it'll replace traditional mastering, but will eventually be recognized as something that exists in tandem with it, as an additional option for mastering and production.
@@sageaudio Yes, I know a lot of people are giving AI too much credit these days, particularly the doomsday people... Anyway, I also think AI assistance is here to stay, not to replace everything but to assist those who already know the craft. AI is good to build the frame but humans are to give the touch of uniqueness, or soul. I was watching a video about most people no longer knowing to differentiate singers who use autotune and pitch correction from those who don't, great video on Wings of Pegasus channel titled Nobody knows what music is any more. The conclusion for me is the same, in the end only humans can afford to choose making "wrong" things, to give that uniqueness that makes things sound better.
I don't believe there will ever be a one button AI that decides and completes a mastering by itself, it will always require human input no matter what. But for sure it will remove a lot of steps in the process and it will require less and less technical knowledge from anyone to be able to deliver a solid great mastering. In that sense, I believe a lot of the courses on mastering we have today will have to adapt, specially those who focus on technicalities. Since AI is still in its infancy, the majority of technical issues with mastering will soon just be easily dealt with AI automatically at the first analysis, and the mastering guy will just have to paint the bigger picture, decide what kind of feeling he wants to transmit with the track, like wanting more punch in the kick, more width overall, a more dreamy or a more dark tone. All the technicalities and steps on how to do it manually will be nulled with AI. Guides and courses on mastering will need to be more like teaching how to feel instead of teaching how to reach that feeling.
then should I set sample rate at 192khz even when I'm recording or making beats?
It would definitely help! That said, it's really taxing on a computer, so it may be hard to do a full mix session at 192kHz.
The project sample rate only matters because of aliasing. Sample rate determines the highest frequency that the digital file can reproduce. So at 192khz, the highest frequency is 96khz. At 48khz, it's 24khz.
Aliasing is what happens when you distort a digital signal and the added harmonics exceed the highest frequency that the digital file can reproduce.
So at 48khz, if you distort cymbals, the harmonics may be higher than 24khz, and they'll reflect back into the spectrum and sound metallic, harsh, grungy.
Typically, the solution is to run the project at 44.1 or 48khz, and use oversampling for any distortion. Oversampling simply make the plugin's sample rate higher, then runs the plugin processes (such as distortion or compression), lowpass filters the output (to remove frequencies higher than the project's maximum), then downsamples it back to the project samplerate.
Oversampling can be more CPU efficient, because linear processes (that don't add harmonics) can run at a lower sample rate. Fewer samples means fewer processor cycles, so saving CPU with these processes with no change in quality is desirable.
Then you can use oversampling on any dynamics or distortion processes to keep aliasing minimal.
@@ampersand64 sorry to but in but I just have to say that it's nice to have someone who actually understands the fundementals of DSP, it's tough going out here, fight the good fight 😆
@@memeswillneverdie i don't even understand that much 😭 I just like to KNOW what's going on with my audio
@@ampersand64 well you may say that but everything you said was correct, trust me that’s a lot more than most, I’m in the same boat as you, no matter how far you get and how much you think you know there’s always more to learn in audio at some points you’ll wish you picked a less complex subject like rocket science 😂
This video was as helpful as helpful can be! Great job!
Awesome! Thanks for watching!
amazing. thanks a lot !! amazing
Thanks for watching!
how do you get that hovering over the plugin to check latency in Logic? i cant find the tick box!? Great video!
Here's a helpful article about it! discussions.apple.com/thread/250429130
@@sageaudio love this. 15 years using logic and always learning cool new tips! Thank you for sharing. Loving your channel. Only just discovered so thrilled to see how much great content you have made to watch!
Why would you need a clean master it won't translate very well on all devices, specially if avoiding saturation. However great idea to increase audio res for more dynamics.
It's just an option! If you want to make a super-clean sounding master, maybe one designed primarily for hi-fi devices/playback systems then this is a good route to take. But like with most things in audio, it just depends on what you're trying to achieve.
The Inflator shouldn’t be changing the tone of song if you’re leaving the Tone at 0, right?
Inflator is a waveshaper, so it introduces harmonic distortion. The goal of this master is to be as clean as possible.
@@sparella right! I forgot it was a waveshaper lol I usually only use a tiny bit of it
@@Diego-rn6cy and maybe you don’t actually want a “clean” master
Thanks for watching - the positive value emulates tube saturation a bit more, and the negative traditional distortion, with the 0 curve value being a blend of the 2. If it sounds good I say use it! I like it on a lot of masters - but for this session I was curious as to how to create the cleanest sound possible.
Thank you so much you are back
Thanks for watching! Happy to be making videos on TH-cam again
VERY useful information! Thank you.
Thank you! Hope it was helpful!
Wow 🔥
Thanks for watching!
@Sage Guys...this seems to contradict your tutorial from two years ago,namely How To Master Loud Without Distortion in terms of a)the use of true peak limiting and b)the use of multiple serial limiters.I'm sure the advice imparted here is absolutely valid given you're the professionals,however for those of us totally home schooled with YT,this common phenomenon of seemingly diametrically opposing advice gets a bit confusing.Unless they are designed for different mastering situations perhaps?
Yes, different situations: Loud priority vs Clean priority.
Thanks for watching both of these videos! The info both videos are simply things to consider - not a declaration of how you should master (there are so many ways to approach audio production, your personal and subjective thoughts and wants for a project will always play a role).
That said, this video was about approaching mastering from a 'how do I alter the original sound as little as possible while still making it loud enough to be enjoyed on other devices' perspective. The How to Master Loud without Distortion video came from a 'How Do I Get this to a super loud level without it sounding bad' perspective.
So they're just highlighting different thought processes for different applications.
For next round, please put in why your making those eq changes to this specific track. Like you putting in that 2db bass eq boost because it sounds thin in that area, you said it for the mid about the vocal, but adding any bass in mastering on tracks that have good bass already or bad muddy bass(which for 99% of home studio operators you most likely to have problems with)it can be confusing! As most will not have the ear for it, as they've never worked for 10yrs in a proper room with proper equipment, so have never even heard before how something SHOULD sound, let alone know that you can't even hear those things with it being canceled in there room to begin with, due to the nulls inherent in their room!
This makes sense! I also hope the demo can justify why certain filters were chosen, but I agree some more explanation would be helpful. Thanks for the comment!
Woah
Thanks for watching!
Thank You AUDIO GODS!!!!
Hahaha thanks for watching!
I think you missed stereo imaging
I liked the stereo image in this one, but I see what you're saying - maybe it could have used some more width
Nice to see you back. Thanks for this.
Thank you! Hope it was helpful
thank you for this
but to me personally the final master sounds "dead"
Thanks for watching and no worries! I don't expect everyone to enjoy the final sound :]
✌🏼
Thanks for watching!
I hope you’re doing well? Been a while since I get notifications about your videos .. 🙏🏾
Doing great! Thanks for asking - hope you've doing doing well too
Remember when this channel had less than 5k, everyone found the pot of gold now
Good times :] Thanks for sticking around!
@@sageaudio cant help it, thanks for being the most consistent source for vocal mixing and mastering tutorials I have found.
I've posted my comments on the follow-up vid but suffice to say, the statements at 0:30 are simply untrue & are misconstruing a correlation.
Thanks for your comment!
def not seeing the mouse over detail of latency per plug in..
Here's how to enable it: discussions.apple.com/thread/250429130
❤
Thanks for watching!
7:14 WTF
It's interesting for sure!
Your A/B comparison is totally unfair. Only a pair of ears is needed to notice the difference in loudness between them.
Thanks for sharing your thoughts! I've addressed this concern to some other people that watched, but in short, I don't want to normalize the loudness of the original mix, since that would alter its timbre and potentially require peak down attenuation. So, I choose to peak-normalize both to get them a little closer and without altering the timbre of either.
9:05 havent you updated saturate yet?
Haha I haven't, need to get on that
What i just watched omg.. is it a prank?) So much bs in one vid.
It's not but thanks for sharing your thoughts!
exactely
Why avoid oversampling on the limiter and clipper?
I expect it's to reduce latency. Recall that the project sample rate is already at 192kHz.
In this session I was attempting to minimize the effect of pre-ringing distortion, which is created by most oversampling filters since most utilize linear phase. To counteract aliasing, I raised the sampling rate, which doesn't work as well as oversampling, but seemed to be a good mid-point between lower levels of aliasing and lower levels of pre-ringing.
If your mastering videos are like this video in terms of speaking pace and visuals..I'd get it
They are! I hope you check them out
Thank you for making good videos, but sometimes it seems to me like your theoretical reasonings are quite out there.
1. Delay compensation causing noise? If the DAW compensates for the delay there should be no phase problems since this would happen before playback.
The only way there could be problems like that is if you run a wet/dry mix that isn't compensated at all. If the plugin manual says it's compensated there is no reason to watch out for this.
With your reasoning using analog gear would always be super detrimental since the delay in sending and receiving audio is much greater than even some of the heavier plugins. Why then are all the pros doing it?
2. Introducing 7hz in your mix would definitely not be audible, but your compressor and/or saturation post would react to it so it would alter the sound for sure. That is probably more likely due to every other aspect of the plugin you introduced that's audibly colored, not the 7hz.
Thanks for watching! I had a lot of questions/concerns with this claim so I made a video showing the relationship: th-cam.com/video/D-P0u9h9vgA/w-d-xo.html
You're right! 7Hz is not audible, but it also introduces harmonics of 7Hz that reach into the perceivable range. When adjusting the parameters of this function during the demo, you can hear the difference.
Awesome useful information never have heard so much insight into latency on plugins makes sense though -
Thanks for watching! One thing I learned after making the video is that if the plugin introduces lookahead, you'll get latency but won't encounter the type of distortion talked about here. That said, if the plugin introduces latency, it's very likely you'll get some distortion due to some behind-the-scenes filters.
Wow. I’ve only had one other engineer tell me about introducing planetary rhythms in mixing/mastering, super fascinating but thought I should understand a proper workflow before trying unethical ways of working. I’m definitely going to look into becoming a member, I look forward to learning from you
It is super interesting! I'm going to look a little more into it to see if it has any real applications in audio. Thought it was cool enough to bring up nonetheless. Looking forward to having you as a member! Here's the link in case it's helpful: SageAudio.com
How about using The God Particle plugin?
I'll have to check out that plugin some more! Only used it a couple of times.
When I look at the 0:51 second video, I see a text box showing the number of latency samples caused by the plug-in, which is not the case in my logic. Does it only come out in the latest version of Logic? I use logic 10.6.3 version.
Here's how you turn that option on: Preferences > Display > General > select 'Show Help Tags'
Nice video, but what if i can choose max 48kHz?
Thanks for watching! That's fine - but you're more likely to run into aliasing distortion. It may be noticeable but might not, depends on the session
That’s brilliant! Shall we connect?
Thanks for watching! You can reach out to us here: admin@sageaudio.com
the thing about latency introducing distortion is false. so is the thing about parallel processing.
reasoning for latency, even if latency compensation introduced artifacts, there is 0 latency compensation while bouncing and so there would be no introduced artifacts lol
Thanks for watching! The compensation is happening prior to exporting the signal - the DAW does this automatically during playback. If the plugin introduces latency, it likely includes a linear phase filter, linear phase oversampling, or convolution, which will cause pre-ringing distortion. Out of curiosity, what do you find false about the parallel processing aspect?
No compression?
Not on this one - but it can be useful when needed!
Auto Control D
What do you mean?
@@sageaudio I'm bookmarking your video?
If all the audio stems are in 48k is it really worth mastering at a higher sample rate? Also, if that is so, aren't the benefits going to be lost when converting the master back down to a commercial sample rate?
It is! If you upsample them from 48kHz to a higher sampling rate, you're still less likely to get aliasing distortion. The benefits won't be lost when converting back down - the aliasing occurs due to processing you add, like high frequency saturators. If you avoid aliasing by giving these plugins more room to work with, then by the time you convert the sampling rate back down, no, or less aliasing will be present.
Dosen't newfangled audio saturate have high latency?
Yes due to an anti-aliasing filter. Odds are though, if the plugin introduces latency, it's due to linear phase filters, oversampling, or convolution which will cause pre-ringing.
I’ve heard from streaky that you should take the channel linking off on the L2 in masters. Also I always thought I should be using oversampling on that plugin for the master. Am I just totally wrong??
For channel linking it depends - if you want the left and right channels to affect the signal independently, then de-link. This will cause a slightly wider sound and result in less limiting overall, but at the expense of mono compatibility.
As for oversampling, I wanted to avoid pre-ringing that comes with it. Since oversampling is used to reduce aliasing distortion, I instead increased the sampling rate of the session, which will have a similar result.
thanks a bunch! great tips as always@@sageaudio
How much does it cost to take your on line course? I'm interested and I want to be a student to learn the ins and outs of music production.
Right now we're seeing what's most feasible for everyone - it currently costs $95 for a biannual membership. This gets you access to our School of Mastering, our mastering services at a 50% discount, 1 free master from our head engineer, lots of additional perks, and any and all educational courses we release in the future.
@@sageaudio great! I have made a account and just waiting on your reply to finish the process. That is a great deal and I'll most definitely be enrolling. And all the courses will be all on line courses, right?
Learn more about how Plugin Latency indicates Pre-Ringing Distortion: th-cam.com/video/DHYoT3RxFww/w-d-xo.html
Honestly I think this is a bad, unartistic approach to the mastering, which can be easily replaced by AI mastering available in Ozone etc. You'll loose all the EDM loudness wars by doing this. If your vision is to have sound and loudness contour like Skrillex or Deadmau5, you have to forget about this and focus on how about to squeeze your master as much as possible to target approx -4LUFS while STILL having the clean sound. This is not a clean master, this is a safe master. The only right thing I can relate to in the video is the chain sequence.
No worries! I wouldn't recommend mastering like this in all occasions. Like you said, if you wanted to make something -4 LUFS, this is not the route to take.
you Lowkey taught me everything, wish you'd go more into mixing over MP3/WAV beats but it's cool :(
what do you mean "mixing over mp3/WAV beats"?
Thanks for watching! So you want more videos about mixing beats?
Mixing vocals and a two track beat
@@sageaudio mixing over a two track, understanding volume level , masked frequency
I always like your free plugin alternative suggestions, but ... What's a good free modern maximizer plugin? The best one i currently have is very old and only 32bit, so can't use it on my newer machine...
I like MCompressor a lot by Melda Audio - if you use the 'Custom Shape' function you can maximize the signal. If i'm remembering it correctly, you can even change the image that's affected - so you could maximize just the mid image if you wanted to, etc.
@@sageaudio hey thanks a lot 👍
I'd be curious to hear why you'd avoid oversampling. Is it purely a pre-ringing concern? Granted, at a 192kHz sample rate, your nyquist frequency is going to be way above human hearing and you're not likely to run into any cramping issues, but still, oversampling will increase the anti-aliasing filter frequency and therefore further reduce aliasing and also quantization distortion (but of course, at 192Khz, we're probably bumping into maximum sample rates of individual plugins, perhaps). Also, something I don't think you mentioned, but is implicit in your assertion to work within a 192kHz session is that you are employing SRC in your DAW for mixes printed at lower sample rates. While many modern DAWs have perfectly acceptable SRC, it still isn't necessarily a given that introducing this DAW SRC is desirable in all cases. For example, in a pitch/catch scenario, it would probably be desirable for the pitching computer to use the native sample rate of the printed mix. Generally, I understand what you are aiming at with your message here, and feel as though this video was probably aimed at those doing their own mastering, in which case, I don't really have any objections to anything you've said, I just feel as though in mastering, there are almost no "always" things. It is always a balance between which types of distortion we're introducing. After all, everything we do in mastering is "distortion," by definition, at least under the broad definition that we are distorting the signal in hopefully desirable ways. 🙂
I hear what you're saying! As for oversampling, yeah it's a latency thing for me - granted, I don't always follow exactly what I've stated here, since sometimes aliasing is more noticeable/more of a problem than the mild affect oversampling has on transients. So it depends! Like you said, there are no "always" when mastering since we're combating some form of distortion at all times just about.
Sorry I didn't mention it but yes - sample rate conversion is occurring if the original sampling rate was lower. But that gives me some food for thought - I might look more into how different DAWs perform SRC and see if there's any differences between them, or if they have any impact on the sound.
personally I'm more concered that he mentioned that the delay compensation of your DAW causes pre-ringing? this is just false.
also your correct in what you say, oversampling is just an obtimised version of running at that native sample rate, by only running at that sample rate for the processes that could produce aliasing for instance, if your compressor has an oversampling feature, most likely it's only oversampling the gain reduction circuit and not the linear input and output gain sliders.
@@memeswillneverdie Yeah, this stuff gets super nuanced, and even with fairly advanced measurement tools, there's a lot of deduction involved in the specific internals of any given plugin. Plugin makers almost never get as specific as is necessary to really understand what's happening under the covers in their documentation. Although, admittedly, that does stray into intellectual property territory.
@@sageaudio What effect does oversampling have on transients? It's basically the same process that the D/A converter on your output goes through, right? Some transients will be higher because of intersample peaks. But those peaks exist in the analog output, regardless of the digital samples' arrangement.
@@ampersand64 I watched the video Sage linked to in other comments. It illustrates the ringing / smearing of transients that OS introduces. The concern isnt when it happens once or twice. It's the stacking of multiple stages of this that becomes audibly detrimental eventually. Hence the session at 192 and no OS vs 48kHz and several OS stages.
Great tutorial, gracias, saludos!
Thanks for watching!