Haha! Thanks James. That's to pre-empt the people who leave comments along the lines of, "This was a waste of my time, you could have just said set it to 12...."
4 people thought this video was too loud for their fragile ears so they downvoted. ;) Thanks for making this in Davinci Resolve Curtis! I hope you'll be doing more Resolve/Fairlight content.
Interesting - because 99% of the music on TH-cam is 3 LUFS louder than this video. Hahaha! I'm still dabbling with Fairlight. Version 16 is better than ever but I'm still getting a lot of crashes when mixing a short film with 25+ audio tracks. My computer is plenty powerful to handle that and more so I'm assuming that's a Resolve issue. But yes, we'll probably have more Fairlight content in the future as it stabilizes and matures.
These are excellent videos. So much detailed information delivered in very clear and concise method. The text summary at the start of the video and the detailed text during the video is really appreciated
Thanks Curtis, again you make it look so straight forward, but the more I watch the more I appreciate what the 'sound guy' does and the more I appreciate your channel . I think I will go and make another coffee then re-watch the previous video and this one . stay well
Good info on the part where you have to consider how much room you have left for loudness and using compression as well. I don't have the full iZotope tools because $$$ but I purchased RX Elements for like $10 when B&H had a sale. Unfortunately that still doesn't have the loudness part in it (or breath and mouth clicks). I just got it for the denoise. What does have a nice tool for loudness is iZotope Ozone 9 Elements. You can tell it to target -16 LUFS and set a maximum peak. So that's what I do! Best of all Sweetwater had a campaing where they were giving it out for free after signing up with them.
Thank you, that is very helpful!! FYI: I just discovered you can get the audio analysis in Resolve without listening through the whole thing as well: right-click the audio track in the sound editor and there you can choose 'analyse audio levels'.
@@curtisjudd I don't know. But I did right-click on the clip in order to get this option, so I guess that was clip-specific.I didn't yet try to get a reading for an entire mix.
Accessing RX from resolve has been an EPIC discovery. I have several recordings that are at the extreme ends of the range as far as db I need to find a few more RX lessons that address the workflow for leveling out the difference. Audience interviews and different mics can create a variety of levels. :) Fairlight is getting better and I like working in RX a bit more. Thanks for the great content and I look forward to learning more as I improve my skills with audio and video. Gratitude for the support.
Thanks Russ. You'll find several RX videos and demos on this channel and my other channel: Curtis Judd Audio. We also have a course dedicated to RX over at school.learnlightandsound.com
Hi Curtis, Great video but could you please explain what happens if you normalise a clip or track using the built in normaliser in Resolve. I see it has various setting such as ITU-R BS. 1770-, ITU-R BS. 1770-4, ATSC A/85, EBU R128, OP-59, TR-B32, AGCOM 219/09/CSP and Netfilx. Would using these do the same thing as you are describing in your video. I have searched to try and find what all these setting in fact do but drawn a blank so hoping you can help.
Hi Dave, those are just different loudness targets with slightly different specifications such as momentary loudness in additional to integrated (total program) loudness in LUFS or LKFS. Yes, they would do the same thing, but you will probably need to do a bit of compression on most dialogue clips first so that the transients aren't too heavily limited. For videos going to web, I just choose ITU-R BS.1770-4 and set the target to -16 LUFS.
@@curtisjudd Thank you so much for getting back. Are these then going to a different job than choosing Sample Peak Program or Sample Peak Program with a Target Level in dBFS chosen by me and would it be best to use this normalising track by track before mixing especially throes with dialogue? Also if you were making a video for a client to be given to them on a file which one would you advise?
@@DaveKnowlesFilmmaker I would just mix with your ears and then focus on loudness normalization at the end once you have the mix sounding like you want and balanced. Then bounce out the mix and loudness normalize the mix. If the video is going to be online, I would choose ITU-R BS.1770-4 and set the target to -16 LKFS/LUFS.
if you have to reduce the noise (preamp & mic noise), when do you do that step? first, before the compressor, after the compressor, or after loudness normalize? thanks
Huge thanks for your work, Curtis! This topic is very important for me. Now I’m trying to master my first short film. I already have an experience working with videos like one in this tutorial, but films looks like much more complicated. I’ve seen your review of your short film audio editing. There was many post processing applied on the tracks, but final loudness optimization hadn’t covered there. Can steps from this video be applied on the short too? (Mixing everything in the one track and making it louder with compression and gain) Or workflow with films is different? I listened to my audio tracks and it seems like most of them requires their specific noise reduction, EQ and compression. Also it’s hard for me to master them without separate volume changing for each clip. That’s why I’m a bit confused.
Hi Galach, generally, the workflow for narrative film is to edit and mix first. Clean and adjust each clip until it all sounds good. Then bounce the entire mix out to a stereo track and loudness normalize that like we showed here.
@@curtisjudd thanks for answer🙏 Also I’d like to ask once more thing you don’t mind. In your video about short film editing, you worked in Audition with project file from FCPX. Now you’re using Davinci, I’m using it too. Does it allow to link workflow with audition? For editing each individual clips and tracks and have an opportunity to make changes in both programs from time to time. Or it’s better to use Fairlight for this purpose. And the last question about separate clips editing in davinci. I used to apply all my effects while editing audio to have an opportunity to see changed waveform (for checking loudness, noice ration, etc) like you showed in this tutorial (I used audition before). But in fairlight i found only methods to add effects on the track, not to apply it. Also it hasn’t as nice wide waveform interface as audition Izotope do. (There even loudness on the mixer is strange, -10 is set to default for unknown reason) This makes my edit process kind a “blind”. But I don’t know. Maybe it’s the only way o should to work
@@galachiev DaVincia Resolve's Fairlight is a very different beast to Audition's waveform editor as you have discovered. I imagine there is a way to round trip between Resolve and Audition, but I haven't worked out how to do it. It is straightforward to set up round-trips to Izotope RX using the external editor option. I don't know if that is possible with Audition or not.
@@curtisjudd understood. I guess I'll try to deal with Fairlight) Read in the comments that you're planning to make more vids about it in future. I've already subscribed and turned on notifications)
1. Did not understand much of this. I promise i paid attention. 2. I am interested in this kind of staff (I actually think it is fascinating) 3. I want to learn more about it. Which course from you should I buy? I think you are great professor. I am sure this is all about me not having enough basic knowledge to follow. But i will fix that.
Thanks. Did you watch the previous video I talked about and linked on loudness vs. volume? That is foundational to understanding this video. If you are using Resolve, we have a Fairlight Fundamentals course. We do not have a course specifically on loudness normalization yet. This is the closest we have. 😀
I think after watching this you should watch the videos of how he applied it in IZotope RX and try to do similar things with your software. Just watching isn't enough.
Thanks, Curtis! This was really helpful, knowing the concepts is much more useful than knowing what buttons to click in some tool. (Especially if your laptop is not powerful enough for DaVinci and your hobby is not at a level yet where you're ready to spend that kind of money on iZotope RX 😄)
Thanks very much for taking the time to upload these great videos. This video was educational - and timely - as it outlines a complex process in simple steps. One note though: I got told that it'd be better to apply target loudness normalization for the dialogoue track(s) not the entire mix. That approach would make sense to me because the music, effects and ... are built around the dialogue anyway. Your thoughts?
You’ll still generally want to do a final loudness check. Often music is substantially louder than dialogue and especially in cases wher e you have not music and dialogue you need to measure the overall loudness if the goal is to produce consistency.
This was incredibly helpful, as are the other videos you've done on the subject. I've been trying to figure out the best way to handle this. Now I need to go look at how to do this in Audacity.
@@curtisjudd Thanks! I went and looked at izotope, and it's a bit out of my price range and even though Audacity has a plugin, it seems pretty basic. This looks perfect for a small channel like mine.
Thanks Curtis. This is super helpful and has already helped me fix some audio I was having a tough time with. I particularly appreciate seeing the work flow from different apps. I didn't know Resolve had a round trip option like that. ~Chris
Excellent video Curtis.. but I have a question and sorry if it’s a dumb one (I’m learning).. When mixing my audio do I have to focus on what db I mix at? I have music and dialogue/vocals. I have read that I should aim for -12db for vocals (which I have done) and mixed music too however my overall export is way too quiet.. When you adjust the LUFS then does that not make my vocal tracks exceed the recommended -12db??
I'm not sure where the -12dB FS recommendation comes from, unless you are producing video for TV? If you use LUFS and aim for a target loudness for TV, then you do NOT need to worry about a -12dB recommendation. Instead, you'll want to hit the loudness target (e.g., -23 LUFS for European broadcast TV) and the true peak level no higher than -1.5dB, then you should be fine. But first you need to decide on a target level and that has to do with where your video will be screened.
Really interesting content, thank you for your knowledge! Currently I am looking for a audio equipment, I shot some corporate, wedding doc and skate videos. I would like to record interviews, voice overs, and sound effects. I was thinking about Tascam or Zoom H5, Audio Technica AT875R (shotgun, voice exterior) and Lewitt LCT 040 (sound effects and small places). Do you recomend me something different? my budget is around 500 ThankS!!!
Hi Miguel, The Tascam DR-60DmkII is decent and should work decently with the AT875R. I haven't used the LCT 040 but seems like a good option for close miking sound effects. Happy recording!
This was a great help to me. I tend to spend too much time on individual tracks instead of just going forward with the mix and then normalizing. I also need to figure out a way to export just the audio mix track and then bring it back in without producing the video file plus audio twice. Thanks.
Thank you so much for your great videos. I recently started producing eLearning videos with Davinci and thanks to you I slowly understand what I am doing :-). I have got two questions. 1. In a much older video you recommend to target normalize to 19 lufs in case of a mono dialogue track. Do you still hold to this recommendation? 2. In Davinci there are among others two normalization modes BS.1770-1 and BS.1770-4. Is there a big differerene and if so which one do you recommend for dialogue?
When I reach a True-Peak of -1,5db with Loudness Control and look on the right side where the Y-Axis shows the dB I can see that my audio is nowhere over -3db. Why does the loudness control show me a True Peak of -1,5db instead??
True peak takes into account that when digital audio is converted back to analogue to play out via speakers or headphones, the waveform can actually go to a higher amplitude between samples.
If all you have is RX Elements, you can do exactly the same workflow. Instead of setting the slider and hitting Render in Loudness Control you can just open up Waveform Stats, calculate the difference between integrated loudness and the target yourself and then click Gain and enter it there.
Great video as always, watched a couple of times. Would it be possible to cover how you could achieve this using Fairlight/Resolve alone? Whilst it might take more time to do, it would be useful to see the steps you'd go through? Also Resolve 17 now seems to have a capability in the Fairlight module to right click and "Analyze Audio Levels", this seems to fix some of the issues with playing through the entire track to get a reading? What are your thoughts on this feature?
GReat video, thanks. I've just started mixing audio for film and video so I'm on a low budget and using Steinberg Waves LE. This doesn't feature LUFS metering. I just have db metering so, would keeping the peaks in my audio at -16db be the correct thing to do?
Thanks Giancarlo. After all the crashes I've experienced over the last couple of weeks, I think I'll be heading back to FCPX and learning how to round-trip to Logic.
@@curtisjudd Davinci is great for color correction / grading but I find FCPX unbeatable for editing (I don't like Premiere). According to some professionals, Logic is exceptional for working with music but not so great for editing the audio section of a video where Pro tools reigns supreme: in my modest experience I am happy with Logic even if you have to adopt different precautions in the workflow (I will do a video about it .. but in Italian though); I hope Apple makes this workflow (XML from FCPX to Logic and vice versa) better.
I've found the Youlean Loudness Meter is pretty good. It's "freemium" model, so you can check out an already very useful version gratuit. Or, pay a pretty reasonable $40 to unlock the Pro features.
At the moment i'm only comfortable with the Pro Tools stock and CLA73 (non post) compressor. I've seen you confidently reduce the attack time to zero a couple of times. When I do that I get distortion unless I add a knee. Just wondered if you/anyone ever had this problem? Obviously if you reduce the attack time, some loud spikes could still sneak through. Thanks for this great content.
I'm sure you've answered these questions a million times, but if you have one more in you: Link to sound proofing Link to sound treatment My office has greater than -60db noise floor for 250hz on down Above 250hz, the noise floor is -60db to -100db with the microphone, XLR cable, and audio interface in the supply line. Not sure when it's time to replace the microphone or anything else. Hoping treating the room is where i need to start. High pass filter at 80hz, compression, noise reduction at a minimum.
Sound blankets: th-cam.com/video/uzyEaVYCk3s/w-d-xo.html they won’t eliminate everything below 150Hz but they help a bit. In my mix room, I use GIK Acoustics broadband traps.
Hey Curtis, I just started editing in Davinci Resolve full time but I am still not sure how good Fairlight is. How would rate Fairlight against pro tools or adobe audition? Also, would you consider doing some kind of extensive Audio tutorial in Fairlight? I am beginning to shoot scripted work and the audio comes out well, but every now and then I record something that needs quite a bit of work to correct, and I am just wondering if Fairlight stands up in that regard, or should I go about editing my audio in Pro Tools or something along those lines. I've seen Fairlight tutorials online for other folks some of them are useful but I honestly trust you more since you really do this for a living. And I'm interested in editing sound for scripted and narrative work versus a talking head on youtube. Thanks and I love the channel.
Hi and thanks! I just used Resolve 16 on a scripted short project (6 and 1/2 minute piece) which had a decent amount of sound design elements in the mix. I used about 25 tracks and started to experience some crashes. But aside from the crashes, I quite like Fairlight. In fact, I have a Fairlight Fundamentals course at school.learnlightandsound.com. You can see a walkthrough of the mix before we finished it over at th-cam.com/video/IR60F5FJZ5s/w-d-xo.html
Hello Curtis, Thanks so much for this great video. There is this very strange thing in davinci fairlight that happens when I try to follow your steps. I put a short music clip on a stereo track to normalize it (nothing else). After normalisation it has an integ. Lufs of -20 and true peak of -4.2 (this is also confirmed by iZotopeRX8). But when I look at Bus1 of fairlight the music will clip at one point. The gain level of Bus1 is higher as in track1, but the gain knob of Bus1 is not elevated, no sound effects or pan or eq is activated. Do you have any idea what could be the reason? I think if you dont nobody will :-). By the way when I change the track type to mono the problem is gone. But than of course the nice stereo effect is gone, too. The only way to stop clipping I found is to place a limiter in Bus1.
Very helpful video showng how to approach audio laudness normalization. Do you maybe know any plugins just for normalization? I'm sure iZotope RX is great, but it's quite pricey for my current needs.
Hey Curtis! Thank you for another awesome guide! I noticed you didn't touch "short-term limit" option. What's your general opinion on it? May be just disable it to avoid unnecessary additional limiting?
Thank you for this.. I want to try it on a podcast project with multiple people speaking on different tracks.. do you have any suggestion for this kind of workflow? i.e. roughly ballance the tracks to sound approximately equally loud and then bouncing them and use the process? is there any way to use fairlight tools only other than showing the loudness history crappy delayed graph?
Yes, exactly. Though if I am mixing a short film and I want to create the best possible sound, I will go into the two tracks and cut out the parts where that person is not speaking which cleans up the overall mix by a LOT. Then bounce it and use this process.
@@curtisjudd Hi, just followed it the way we discussed.. but I am kinda confused.. in davinci my loudness history meter says the bounced track is sitting at -16 LUFS but when I open it in RX it says it is at -23. when I follow your procedure and compress the audio so the peaks won't clip and bump it up to get the loudness in RX to -16, when I move back to davinci it is much much louder and loudness history meter reads +5 LUFS 🤷♂️ do you have any idea what is going on?
@@curtisjudd it was set correctly in project settings. Only thing I was doing wrong was that I had EBU R128 set in the loudness meter on fairlight page instead of bs1770. Didn’t know it could have such effect
I use it *sometimes*. I just mixed a short film in Resolve 16 but had a lot of crashes once we had more than 20 tracks in Fairlight. So I'm hopeful for its future, but not convinced it is ready for bigger jobs like that just yet.
@@curtisjudd I just moved from the Zoom F4 to the SD MixPre 6II. Am I going overboard if I will record a three person 1-2 hour long podcasts all in 32 bit? And what about the sampling rate? I was also wondering, if you use the mix assist plug in is is just recorded to the L+R mix? Which means that you cannot tweak all the individual tracks in post with the mix assist plug in? Thanks.
@@OliGudbjartsson12 Hi Oli, I don't usually use 32-bit for recording spoken word audio, but you can if you want. SD cards are cheap so it isn't much of a problem. I'd use a 48kHz sample rate for spoken word recordings. MixAssist only affects the mix channels, but you can still record isolated channels for each of the mics so that you can do a more manual/in-depth mix in post if you want. The MixPre gives you total freedom that way. (so did the F4, for that matter).
Hello Curtis, Thank you very much for your video. With respect to your example, have you tried in the past using the loudness control module of RX for normalisation without using compression? I was surprised that you first compressed the audio in order to get more headroom before using izotopeRX. Using iZotopeRX8 myself I found out that the loudness control module of RX takes care of that automatically when loudness normalizing. I am not a pro, but if I set eg. integrated lkfs to -16 and true peak to - 1.5 then RX will deliver this even though the loudness increase was higher then my available headroom. RX will make sure that the peaks of your waveform will not pass the true peak limit while increasing the integrated loudness lkfs to the desired value. It would be very interesting to hear whether it also works for you.
Hi Vali, yes, it certainly works, but it tends to limit too heavily and that can be heard if it is too heavy. That's why I compress first so I have more control over how the final audio sounds.
@@curtisjudd Thanks, that was very helpful. Since I do mainly dialogue with very few wafeform peaks that need attention I never noticed heavy compressed audio by RX but very good to know that compression can be necessary in other situations.
I'm using a Canon G7X Mark II with no external mic. (they're ordered) My question is where do you generally have your computer speaker volume set and then your computer sound volume set to begin editing? In DaVinci...
Volume and loudness are not the same thing. For your speaker volume, you should set it to a comfortable listening level. If you want to get technical, it should be at or less than 78 dB SPL.
Thank you so much for the video. but I getting error message after hitting Izotop ' ~ MixL_IZOTOP_n898B.wav' could not be opened. Product Portal cannot open files in the 'Waveform audio' format. Any advice? Thank you~!!
Hey Curtis, I've been studying your videos all week and been learning a ton. Thank you. For Loudness Normalization, would an easier way be to just put in the loudness normalization settings in the Adobe Premiere Pro export menu? I see you can also specify the true peak value (like at -1 dB), so nothing clips.
You can do that, but often if you're normalizing for web to -16 LUFS, you need to do some compression first so that the audio doesn't sound distorted. Sometimes if you apply a limiter, like the export functionality does, and the audio hits up hard against that limiter, it will still sound distorted. Best to use a compressor first to manage the waveform peaks and get it sounding smooth while leaving enough headroom to get the loudness target you need.
Hey Curtis, thanks for the video! One question, is there a specific reason you apply the compression to the entire mix instead of just the dialogue track beforehand?
As part of the mix, you can apply compression to the dialogue before bouncing and doing the loudness normalization process. Then I'd do a final pass of compression on the loudness mix just to ensure we're not limiting any transients too much, IF NEEDED. Often when you have a more complex mix with lots of overlapping tracks, their sum ends up producing larger amplitude than expected so this still needs to be managed at the end with a compression or limiter pass.
@@curtisjudd Great, thanks for the clarification! I purchased a new mic recently, and as part of my research I scrolled through comments you replied to on old videos. I found the extra follow-ups you leave in the comments very useful.
Thanks for the video. Just what I've been dabbling with the past couple of days while finalising an online course. Btw great to see some Resolve examples. I do it a little differently (apart from the fact that I don't own the big version of RX). Namely I try to avoid compression on the full mix. The reason being that I am using stock music which was already heavily compressed by the composer, and I don't want to compress that even further. Instead I add Ozone Limiter (set to -1.1dB True Peak) and Youlean Loudness meter (with TH-cam preset) to the mixbus. Then I increase the level in Resolve (using BX Masterdesk as a glorified master fader) until Youlean shows me I am between -15 and -14dB (recommended for TH-cam). Finally I check how the aforementioned music track is doing LUFS-wise when it is playing at full level. Usually I have to pull it's level back until Youelean gets below -14dB again. And lastly I drag the final export on to the standalone Youlean meter app which calculates a LUFS curve chart for the entire video plus indicating any Peaks above -1dB (YT setting) in case there are any.
Thanks for the insights Heiner. I find that most popular music is already HEAVILY compressed as well. What that usually means is that I have to attenuate it quite a lot to get it sounding right with the dialogue and then it basically never gets any further compression when I work on the final bounce/mix because its transients already leave plenty of headroom.
I was super excited when you announced this video, but even though I have a copy of iZotope RX8 Advanced, and despite decades of reading everything about audio production, I feel helpless to grasp how to use these advanced tools successfully. It seems impossible to be self-taught on this topic. I'm desperate to learn these tools, but just cannot figure out how to approach the task.
Hi Danny, I'm hoping to make a course to cover this in more detail. But the main idea here is just using compression to make enough headroom so that you can boost everything up to your target level (LUFS).
Thanks for another amazing video, Curtis! When you do your mix, do you first compress the vocals, then once everything is mixed, compress the whole mix again (as shown in this video) for loudness normalization? If so, when you first compress the vocals, how do you know how much compression is enough? I thought we only had to compress the vocals. Ahh, this compression thing is so complex...grateful for your teaching in demystifying it.
@@curtisjudd Gotcha, thank you! One question I have remaining is that of where to place the loudness levels of dialogue, music, sfx in your mix. I'm confused because I learned dialogue should be bouncing in between -12 to -15 dB, but then I saw you normalized your dialogue to -0.1 dB. If that's the case, where should music and sfx generally be sitting at?
iZotope rx is too expensive. How to perform the same operation using Aсon Acoustica? I didn't find a way to set the headroom target value and the Integrated LKFS target value when trying to normalize.
Good news, there are several options covered in these previous videos: th-cam.com/video/RrRn8stkEBs/w-d-xo.html th-cam.com/video/OKSWPrT5upo/w-d-xo.html
Hi, Curtis! Looking for plugin with pre-setted loudness target lvl, for automatic loudness correction on master in DAW, before render. My vst chain is close to target lufs, but every 3-rd mixdown I fain to re-level and re-render just becouse mix appear louder by 1-ich dB (. can U reccomend some?
Izotope RX Advanced costs way too much comparing to Resolve Studio. Could you show how to do loudness normalization with the latest Resolve Studio Fairlight only?
I have not regrets making the purchase for advanced for my corporate and short narrative work. You can see the differences here: www.izotope.com/en/products/rx/features.html#comp and if you go for the standard version, you can always upgrade later.
So, correct me if I’m wrong, but you’re saying the parts of the waveform which are lower db (lower peaks) we’re perceiving as the same loudness as those parts which are higher db (higher peaks)? I don’t understand why you wouldn’t want a waveform that looks more uniform across the timeline. Or are you saying that you don’t want the listener to hear the same loudness throughout and you’d rather them hear the louder and quieter parts of the voice? Or is it that having the same peak throughout the entire waveform is bad because we’ll be perceiving higher pitched parts differently than lower pitched parts?
Clarification: Waveforms with large amplitude (tall) only sound loud to us when they are sustained for more than a few milliseconds. Smaller amplitude portions are always quieter to us. But we only perceive something as loud if the amplitude is tall AND the tall amplitude lasts for more than a few milliseconds. So in that context, when you're finishing/mixing dialogue audio, you first need to decide where the audio will be heard (online, TV, theater, etc.) and make choices about how "loud" you need it to be. If the audio will be consumed online, that means people will often be listening with less than ideal circumstances - in cars, trains, planes, and listening with phone speakers or ear buds - you'll probably need to normalize it to a louder level. And to do that you generally need to reduce the dynamic range of the audio. There are several tools for that including levelers and compressors. For TV or theater, the listening environment is usually better optimized. So you can retain more dynamic range and do NOT have to normalize so loud. I hope that helps.
@@curtisjudd Yes that helps. Thank you. Just to make sure though, going through and doing it manually on your timeline with the pen tool and adjusting the levels is the same thing right? These tools are only designed to expedite the process?
@@carlsayers7370 You could use clip gain, yes, but that can often end up sounding rather odd if you've got any substantial noise floor because it will obviously jump to. louder levels. So I usually try to avoid any sort of extreme automation or clip gain like that (with the pen tool, as you said). Yes, compression automates this. And the critical tool is a loudness meter, regardless of which way you choose to optimize your loudness.
Running into a few issues and I'd love some help! I'm editing podcast audio and consistently run out of headroom to hit -17lufs even after compression. Anything helps! Thanks!
Really good explanation but there's something I don't understand. If you keep it at -16 LUFS (with -1.9 headroom) and youtube normalises to -14 isn't it going to clip once the audio is turned up? I understand the reasoning from the other video to conserve dialogue audio quality but I think this method could be a false friend. If you don't compress the audio enough and don't leave enough headroom you could inadvertently make youtube clip your audio once it turns it up. Maybe I'm missing something here. Either way this was a really good video!
Hi Bruno. TH-cam does NOT change the loudness of your video unless it is LOUDER than -14 LUFS (i.e., your video would have to be -13, or -12, or even louder). If your video is louder than -14 LUFS, it reduces the loudness to -14 LUFS. So if you normalize to -16, TH-cam will leave it at -16 LUFS. So no false friends here. Just normalize to -16 LUFS and you are totally safe.
@@curtisjudd That makes so much sense!! Thanks for your reply. My background is more in music and since Spotify does normalise it's risky to go under because you don't know what they will do. Once again thank you for your response and the very informative video. I'm really enjoying your RX tutorials.
Another way to do this in rx and DaVinci is to just do the round trip with just dialogue/VO track and normalize it to - 18db. And then bring it back I to DaVinci and use the eq and faders in the music and effects tracks to target - 16. This works well because I often find myself using RX for other things in the voice so I only have one round trip.
Yes, and I would suggest a final loudness check to make sure the final mix sits where it should. But yes, that should get you very close in most cases.
I do the final loudness check in resolve using the loudness line graph below the master track. That way I have a visual of loudness over time. I just run it while doing my final check before rendering. If there is an issue I can see where and then fix that part and spot check it.
Still working on that. For live, most people still just use a compressor or limiter and push the levels up until it sounds loud enough or meets an RMS level. The trick is to put a loudness meter close to the end of the signal chain. I haven't found a good way to do that yet since I stream directly from an ATEM Mini.
@@curtisjudd Kudos for the lightning fast response! I see how that makes it more difficult. Perhaps it's worth a shot to recreate the workflow in software as a showcase? Either way I'm curious what you'll come up with. Good luck!
“You are dismissed, we don’t cover that here.” Love it, that’s why I come here.
A polite way of saying Buh Bye now.
Haha! Thanks James. That's to pre-empt the people who leave comments along the lines of, "This was a waste of my time, you could have just said set it to 12...."
Good one.
@@curtisjudd When I advise your videos to people, I always tell them, Curtis Judd does not waste your time.
@@kirkdarling4120 👍
Can't wait for your video on how to do this with Final Cut and iZotope RX !
I don’t think there’s a smooth transition from fcp to rx. Hope I’m wrong
@@DavidKfilmmaker Yea, you need to export your final mix and apply iZotope to that.
We did a sort of wonky approach here using Youlean Loudness meter a few years back: th-cam.com/video/xgDKF5Rjj7s/w-d-xo.html
4 people thought this video was too loud for their fragile ears so they downvoted. ;) Thanks for making this in Davinci Resolve Curtis! I hope you'll be doing more Resolve/Fairlight content.
Interesting - because 99% of the music on TH-cam is 3 LUFS louder than this video. Hahaha! I'm still dabbling with Fairlight. Version 16 is better than ever but I'm still getting a lot of crashes when mixing a short film with 25+ audio tracks. My computer is plenty powerful to handle that and more so I'm assuming that's a Resolve issue. But yes, we'll probably have more Fairlight content in the future as it stabilizes and matures.
@@curtisjudd are you on a Windows box? I get issues with Resolve too. 15 was so stable...16 has been a nightmare for me.
@@joenicklo I'm running it on a Mac Pro.
These are excellent videos. So much detailed information delivered in very clear and concise method. The text summary at the start of the video and the detailed text during the video is really appreciated
Thanks Chris!
Thanks Curtis, again you make it look so straight forward, but the more I watch the more I appreciate what the 'sound guy' does and the more I appreciate your channel . I think I will go and make another coffee then re-watch the previous video and this one .
stay well
Thanks John!
Good info on the part where you have to consider how much room you have left for loudness and using compression as well.
I don't have the full iZotope tools because $$$ but I purchased RX Elements for like $10 when B&H had a sale. Unfortunately that still doesn't have the loudness part in it (or breath and mouth clicks). I just got it for the denoise. What does have a nice tool for loudness is iZotope Ozone 9 Elements. You can tell it to target -16 LUFS and set a maximum peak. So that's what I do! Best of all Sweetwater had a campaing where they were giving it out for free after signing up with them.
Thanks - good to know that Ozone Elements can get the job done!
Thank you, that is very helpful!!
FYI: I just discovered you can get the audio analysis in Resolve without listening through the whole thing as well: right-click the audio track in the sound editor and there you can choose 'analyse audio levels'.
Thanks Thomas. Is it possible to analyze the levels for the entire mix or is that just for the selected clip?
@@curtisjudd I don't know. But I did right-click on the clip in order to get this option, so I guess that was clip-specific.I didn't yet try to get a reading for an entire mix.
There’s the video I’ve been waiting for!
I hope it helps and thanks for coming by!
Thanks for the video, Curtis!
👍
Accessing RX from resolve has been an EPIC discovery. I have several recordings that are at the extreme ends of the range as far as db I need to find a few more RX lessons that address the workflow for leveling out the difference.
Audience interviews and different mics can create a variety of levels. :)
Fairlight is getting better and I like working in RX a bit more. Thanks for the great content and I look forward to learning more as I improve my skills with audio and video.
Gratitude for the support.
Thanks Russ. You'll find several RX videos and demos on this channel and my other channel: Curtis Judd Audio. We also have a course dedicated to RX over at school.learnlightandsound.com
Hi Curtis, Great video but could you please explain what happens if you normalise a clip or track using the built in normaliser in Resolve. I see it has various setting such as ITU-R BS. 1770-, ITU-R BS. 1770-4, ATSC A/85, EBU R128, OP-59, TR-B32, AGCOM 219/09/CSP and Netfilx. Would using these do the same thing as you are describing in your video. I have searched to try and find what all these setting in fact do but drawn a blank so hoping you can help.
Hi Dave, those are just different loudness targets with slightly different specifications such as momentary loudness in additional to integrated (total program) loudness in LUFS or LKFS. Yes, they would do the same thing, but you will probably need to do a bit of compression on most dialogue clips first so that the transients aren't too heavily limited. For videos going to web, I just choose ITU-R BS.1770-4 and set the target to -16 LUFS.
@@curtisjudd Thank you so much for getting back. Are these then going to a different job than choosing Sample Peak Program or Sample Peak Program with a Target Level in dBFS chosen by me and would it be best to use this normalising track by track before mixing especially throes with dialogue?
Also if you were making a video for a client to be given to them on a file which one would you advise?
@@DaveKnowlesFilmmaker I would just mix with your ears and then focus on loudness normalization at the end once you have the mix sounding like you want and balanced. Then bounce out the mix and loudness normalize the mix. If the video is going to be online, I would choose ITU-R BS.1770-4 and set the target to -16 LKFS/LUFS.
@@curtisjudd Thanks for your advice.
if you have to reduce the noise (preamp & mic noise), when do you do that step? first, before the compressor, after the compressor, or after loudness normalize? thanks
Before loudness normalization (which includes compression).
Thank you for this video! It's one of the most useful tutorial I've watched!
You're most welcome, thanks for coming by!
Huge thanks for your work, Curtis! This topic is very important for me. Now I’m trying to master my first short film. I already have an experience working with videos like one in this tutorial, but films looks like much more complicated. I’ve seen your review of your short film audio editing. There was many post processing applied on the tracks, but final loudness optimization hadn’t covered there. Can steps from this video be applied on the short too? (Mixing everything in the one track and making it louder with compression and gain) Or workflow with films is different? I listened to my audio tracks and it seems like most of them requires their specific noise reduction, EQ and compression. Also it’s hard for me to master them without separate volume changing for each clip. That’s why I’m a bit confused.
Hi Galach, generally, the workflow for narrative film is to edit and mix first. Clean and adjust each clip until it all sounds good. Then bounce the entire mix out to a stereo track and loudness normalize that like we showed here.
@@curtisjudd thanks for answer🙏 Also I’d like to ask once more thing you don’t mind. In your video about short film editing, you worked in Audition with project file from FCPX. Now you’re using Davinci, I’m using it too. Does it allow to link workflow with audition? For editing each individual clips and tracks and have an opportunity to make changes in both programs from time to time. Or it’s better to use Fairlight for this purpose.
And the last question about separate clips editing in davinci. I used to apply all my effects while editing audio to have an opportunity to see changed waveform (for checking loudness, noice ration, etc) like you showed in this tutorial (I used audition before). But in fairlight i found only methods to add effects on the track, not to apply it. Also it hasn’t as nice wide waveform interface as audition Izotope do. (There even loudness on the mixer is strange, -10 is set to default for unknown reason)
This makes my edit process kind a “blind”. But I don’t know. Maybe it’s the only way o should to work
@@galachiev DaVincia Resolve's Fairlight is a very different beast to Audition's waveform editor as you have discovered. I imagine there is a way to round trip between Resolve and Audition, but I haven't worked out how to do it. It is straightforward to set up round-trips to Izotope RX using the external editor option. I don't know if that is possible with Audition or not.
@@curtisjudd understood. I guess I'll try to deal with Fairlight) Read in the comments that you're planning to make more vids about it in future. I've already subscribed and turned on notifications)
1. Did not understand much of this. I promise i paid attention.
2. I am interested in this kind of staff (I actually think it is fascinating)
3. I want to learn more about it. Which course from you should I buy?
I think you are great professor. I am sure this is all about me not having enough basic knowledge to follow. But i will fix that.
Thanks. Did you watch the previous video I talked about and linked on loudness vs. volume? That is foundational to understanding this video. If you are using Resolve, we have a Fairlight Fundamentals course. We do not have a course specifically on loudness normalization yet. This is the closest we have. 😀
Same. Went too fast for me
I think after watching this you should watch the videos of how he applied it in IZotope RX and try to do similar things with your software. Just watching isn't enough.
@@rsmith02 which video(s) are you referring to?
Great info. I’d buy a full course on RX.
Thanks Danny - good idea on an RX course. If only I could shake this pesky day job to free up some time.... 😉
Great explanations again ..... Now I got it 8 months later, it means that I am learning 😉
Thanks Magnus.
Thanks for the tutorial - very informative and concise!
Thanks.
Thanks, Curtis! This was really helpful, knowing the concepts is much more useful than knowing what buttons to click in some tool. (Especially if your laptop is not powerful enough for DaVinci and your hobby is not at a level yet where you're ready to spend that kind of money on iZotope RX 😄)
Thanks Arthur!
This is so good , I came back. 2x in less than 24 minutes :). Thank you for sharing !
👍
Each time I think I reach some decent level of audio editing You show me the whole new dimension to learn:)))) Thank you very much!!!
You bet! Keep making great sound!
Thanks very much for taking the time to upload these great videos. This video was educational - and timely - as it outlines a complex process in simple steps. One note though: I got told that it'd be better to apply target loudness normalization for the dialogoue track(s) not the entire mix. That approach would make sense to me because the music, effects and ... are built around the dialogue anyway. Your thoughts?
I was wondering the same. I only apply loudness to dialogue because music is often times quite loud and compressed already.
You’ll still generally want to do a final loudness check. Often music is substantially louder than dialogue and especially in cases wher e you have not music and dialogue you need to measure the overall loudness if the goal is to produce consistency.
I checked out both part I and this part II, both very helpful... thank you Curtis!
👍
It was very helpful in understanding the concept! 🙂👍
👍
Always awesome and I really love the RX and Resolve combination for audio. Gratitude Curtis.
Thanks Russ.
This was incredibly helpful, as are the other videos you've done on the subject. I've been trying to figure out the best way to handle this. Now I need to go look at how to do this in Audacity.
Thanks! You might look at the free version of the Youlean Loudness meter over at youlean.co/youlean-loudness-meter/
@@curtisjudd Thanks! I went and looked at izotope, and it's a bit out of my price range and even though Audacity has a plugin, it seems pretty basic. This looks perfect for a small channel like mine.
@@TheGameBench 👍
Excellent, once again. Made my Sunday morning!
Thanks Marshall.
23 seconds into the video and I love it already
👍
Super interesting. Love how you lay stuff out.
Thanks Jacob.
Thank for information
Could you please to make a video how to do that in Resolve native effects ?
Unfortunately, it is a fair bit more complex in Resolve/Fairlight.
Thanks Curtis. This is super helpful and has already helped me fix some audio I was having a tough time with. I particularly appreciate seeing the work flow from different apps. I didn't know Resolve had a round trip option like that. ~Chris
Thanks Chris!
Excellent video Curtis.. but I have a question and sorry if it’s a dumb one (I’m learning)..
When mixing my audio do I have to focus on what db I mix at? I have music and dialogue/vocals. I have read that I should aim for -12db for vocals (which I have done) and mixed music too however my overall export is way too quiet..
When you adjust the LUFS then does that not make my vocal tracks exceed the recommended -12db??
I'm not sure where the -12dB FS recommendation comes from, unless you are producing video for TV? If you use LUFS and aim for a target loudness for TV, then you do NOT need to worry about a -12dB recommendation. Instead, you'll want to hit the loudness target (e.g., -23 LUFS for European broadcast TV) and the true peak level no higher than -1.5dB, then you should be fine. But first you need to decide on a target level and that has to do with where your video will be screened.
@@curtisjudd Thank you Curtis for taking the time to respond, I really do appreciate it. That clears up my confusion..
Thanks for the great video Curtis. What version of RX is this? Considering buying.
This was 9 I believe but it hasn’t changed for version 10 - all of this still works exactly the same.
@@curtisjudd I guess “level” is maybe what I meant. Basic, elements, etc. Thank you for responding.
@@FlyingJackalope oh, sorry about that - you’ll need the standard or advanced version for the loudness control module.
@@curtisjudd Thank you Curtis
Great channel, learnt so many great "hands on" production tips from you. Keep it up!
Thanks Mikael!
Great video. Would really appreciate a short follow up of the equivalent steps in Premiere Pro/Audition 👍
We got you covered with some previous videos: th-cam.com/video/OKSWPrT5upo/w-d-xo.html th-cam.com/video/RrRn8stkEBs/w-d-xo.html
Thank you, Curtis. Very helpful.
Thanks Ethan.
Really interesting content, thank you for your knowledge! Currently I am looking for a audio equipment, I shot some corporate, wedding doc and skate videos.
I would like to record interviews, voice overs, and sound effects. I was thinking about Tascam or Zoom H5, Audio Technica AT875R (shotgun, voice exterior) and Lewitt LCT 040 (sound effects and small places). Do you recomend me something different? my budget is around 500 ThankS!!!
Hi Miguel, The Tascam DR-60DmkII is decent and should work decently with the AT875R. I haven't used the LCT 040 but seems like a good option for close miking sound effects. Happy recording!
Best in the biz Curtis. Thanks for all your thoughtful content. I would love to see this workflow in Audition, and if you need plugins to optimize it.
Brett, this might help: th-cam.com/video/sUAfW7lZNPg/w-d-xo.html
This was a great help to me. I tend to spend too much time on individual tracks instead of just going forward with the mix and then normalizing. I also need to figure out a way to export just the audio mix track and then bring it back in without producing the video file plus audio twice. Thanks.
Hi Joe - in this case, I just muted all of the original audio tracks.
Thank you Curtis, your videos are the best! Why not using a compressor in fairlight after loudness normalization?
No need to compress after the loudness normalization is done - that’s the very last step.
Thanks a lot for the extensive guide! It’s going to help me refine some of my audio techniques when it comes to talking head videos!
👍 Happy mixing!
Thank you so much for your great videos. I recently started producing eLearning videos with Davinci and thanks to you I slowly understand what I am doing :-).
I have got two questions.
1. In a much older video you recommend to target normalize to 19 lufs in case of a mono dialogue track. Do you still hold to this recommendation?
2. In Davinci there are among others two normalization modes BS.1770-1 and BS.1770-4. Is there a big differerene and if so which one do you recommend for dialogue?
1) yes 2) use the latest
Pure gold - thanks Curtis!
Thanks Derek!
which version of iZotope do you use? standard? full suite? thanks
Is the Advanced
Yes, this was the Advanced version here.
When I reach a True-Peak of -1,5db with Loudness Control and look on the right side where the Y-Axis shows the dB I can see that my audio is nowhere over -3db. Why does the loudness control show me a True Peak of -1,5db instead??
True peak takes into account that when digital audio is converted back to analogue to play out via speakers or headphones, the waveform can actually go to a higher amplitude between samples.
If all you have is RX Elements, you can do exactly the same workflow. Instead of setting the slider and hitting Render in Loudness Control you can just open up Waveform Stats, calculate the difference between integrated loudness and the target yourself and then click Gain and enter it there.
Yes, being sure to do your maths and make sure you won’t clip or heavily limit the audio.
I need this. Thank you, Curtis!
You're most welcome, Homer!
Thank you! Does the breath and de clicker affect background audio? Would it be better to clean the vocal track first?
Yes, do any cleanup before loudness normalization.
Great video as always, watched a couple of times. Would it be possible to cover how you could achieve this using Fairlight/Resolve alone? Whilst it might take more time to do, it would be useful to see the steps you'd go through? Also Resolve 17 now seems to have a capability in the Fairlight module to right click and "Analyze Audio Levels", this seems to fix some of the issues with playing through the entire track to get a reading? What are your thoughts on this feature?
Hi Neil, once I finish mixing a film I started a couple of months ago in v16, I’ll take a look at 17.
Yes I’ve been waiting for this!
👍
Very good information as always!
Thanks ahriik.
Will be waiting for the premiere/audition video
Those are linked in the upper right corner of the video.
GReat video, thanks. I've just started mixing audio for film and video so I'm on a low budget and using Steinberg Waves LE. This doesn't feature LUFS metering. I just have db metering so, would keeping the peaks in my audio at -16db be the correct thing to do?
Hi Nicholas, no, that's not the same thing. You might try installed the free Youlean Loudness meter and use that. Best wishes!
@@curtisjudd Great, thank you Curtis.
Great video Curtis as usual! I work with Final Cut Pro and manage the audio with Logic and Izotope RX
Thanks Giancarlo. After all the crashes I've experienced over the last couple of weeks, I think I'll be heading back to FCPX and learning how to round-trip to Logic.
@@curtisjudd Davinci is great for color correction / grading but I find FCPX unbeatable for editing (I don't like Premiere). According to some professionals, Logic is exceptional for working with music but not so great for editing the audio section of a video where Pro tools reigns supreme: in my modest experience I am happy with Logic even if you have to adopt different precautions in the workflow (I will do a video about it .. but in Italian though); I hope Apple makes this workflow (XML from FCPX to Logic and vice versa) better.
I've found the Youlean Loudness Meter is pretty good. It's "freemium" model, so you can check out an already very useful version gratuit. Or, pay a pretty reasonable $40 to unlock the Pro features.
Yes, great point. Thanks Jason!
HOFA has a free one but it is just a meter.
At the moment i'm only comfortable with the Pro Tools stock and CLA73 (non post) compressor. I've seen you confidently reduce the attack time to zero a couple of times. When I do that I get distortion unless I add a knee. Just wondered if you/anyone ever had this problem? Obviously if you reduce the attack time, some loud spikes could still sneak through.
Thanks for this great content.
Soft knee is not a bad idea. If it works, stick with it 👍
I'm sure you've answered these questions a million times, but if you have one more in you:
Link to sound proofing
Link to sound treatment
My office has greater than -60db noise floor for 250hz on down
Above 250hz, the noise floor is -60db to -100db with the microphone, XLR cable, and audio interface in the supply line.
Not sure when it's time to replace the microphone or anything else. Hoping treating the room is where i need to start. High pass filter at 80hz, compression, noise reduction at a minimum.
Sound blankets: th-cam.com/video/uzyEaVYCk3s/w-d-xo.html they won’t eliminate everything below 150Hz but they help a bit. In my mix room, I use GIK Acoustics broadband traps.
Thank You! Can you do same thing using Davinci Resolve plugins?
Yes
I loved this video, very interesting would love to work out how to do all this with open source software ?
Me too! But I don't know of any options in the Open Source space, sadly.
@@curtisjudd not open source but I am playing with wavepad and it's much better than I remember.
I love you, you save my life.
❤️
Thanks Curtis. What would we do without you? xx
Thanks!
It has to be -14 LUFS for TH-cam videos na ?
No. It has to be no louder than -14 LUFS for TH-cam and for spoken word, I’d stay closer to -16.
Hey Curtis, I just started editing in Davinci Resolve full time but I am still not sure how good Fairlight is. How would rate Fairlight against pro tools or adobe audition? Also, would you consider doing some kind of extensive Audio tutorial in Fairlight? I am beginning to shoot scripted work and the audio comes out well, but every now and then I record something that needs quite a bit of work to correct, and I am just wondering if Fairlight stands up in that regard, or should I go about editing my audio in Pro Tools or something along those lines. I've seen Fairlight tutorials online for other folks some of them are useful but I honestly trust you more since you really do this for a living. And I'm interested in editing sound for scripted and narrative work versus a talking head on youtube.
Thanks and I love the channel.
Hi and thanks! I just used Resolve 16 on a scripted short project (6 and 1/2 minute piece) which had a decent amount of sound design elements in the mix. I used about 25 tracks and started to experience some crashes. But aside from the crashes, I quite like Fairlight. In fact, I have a Fairlight Fundamentals course at school.learnlightandsound.com. You can see a walkthrough of the mix before we finished it over at th-cam.com/video/IR60F5FJZ5s/w-d-xo.html
Thanks! This helps me to know what I’m doing when adjusting loudness. How would 32 bit float files effect this workflow?
Workflow stays the same for 32-bit float, might just have to reduce the overall levels to hit your target.
Good. Now what about Film applications? You wouldn't typically compress them as much. Would you do loudness control in the same way?
Yes, and probably target -24 LUFS.
Hello Curtis, Thanks so much for this great video. There is this very strange thing in davinci fairlight that happens when I try to follow your steps. I put a short music clip on a stereo track to normalize it (nothing else). After normalisation it has an integ. Lufs of -20 and true peak of -4.2 (this is also confirmed by iZotopeRX8). But when I look at Bus1 of fairlight the music will clip at one point. The gain level of Bus1 is higher as in track1, but the gain knob of Bus1 is not elevated, no sound effects or pan or eq is activated. Do you have any idea what could be the reason? I think if you dont nobody will :-). By the way when I change the track type to mono the problem is gone. But than of course the nice stereo effect is gone, too. The only way to stop clipping I found is to place a limiter in Bus1.
It sounds like maybe there is some processing happening on Bus 1. You might also contact BMD support to see what they say. Best wishes!
Very helpful video showng how to approach audio laudness normalization. Do you maybe know any plugins just for normalization? I'm sure iZotope RX is great, but it's quite pricey for my current needs.
Youlean Loudness Meter. Also, Adobe Audition has some good included loudness tools.
I use TBProAudio's free dpMeter plugin in Audacity and it works pretty okay if you want to play around with some things for free
@@arthurrump Thanks Arthur, hadn't heard of that one.
Thanks again for the tutorial! You mentioned Izotope Rx, which costs $400. Do you know any other cheaper software? Thanks!
DaVinci Resolve with it's Fairlight tab is free and quite good.
This was awesome!
Thanks!
Hey Curtis! Thank you for another awesome guide! I noticed you didn't touch "short-term limit" option. What's your general opinion on it? May be just disable it to avoid unnecessary additional limiting?
For online video, yes. For broadcast video, you'd need to follow the spec for the region and the station.
Good stuff. Thank you, Sir...
You're most welcome!
Thank you for this.. I want to try it on a podcast project with multiple people speaking on different tracks.. do you have any suggestion for this kind of workflow? i.e. roughly ballance the tracks to sound approximately equally loud and then bouncing them and use the process? is there any way to use fairlight tools only other than showing the loudness history crappy delayed graph?
Yes, exactly. Though if I am mixing a short film and I want to create the best possible sound, I will go into the two tracks and cut out the parts where that person is not speaking which cleans up the overall mix by a LOT. Then bounce it and use this process.
@@curtisjudd thank you
@@curtisjudd Hi, just followed it the way we discussed.. but I am kinda confused.. in davinci my loudness history meter says the bounced track is sitting at -16 LUFS but when I open it in RX it says it is at -23. when I follow your procedure and compress the audio so the peaks won't clip and bump it up to get the loudness in RX to -16, when I move back to davinci it is much much louder and loudness history meter reads +5 LUFS 🤷♂️ do you have any idea what is going on?
@@creativevisualssk most likely you need to change the loudness target in the project settings to -16 LUFS. It is probably set at -23 presently.
@@curtisjudd it was set correctly in project settings. Only thing I was doing wrong was that I had EBU R128 set in the loudness meter on fairlight page instead of bs1770. Didn’t know it could have such effect
Great video! Curtis, you've been putting up some quality content lately. You're my first source for audio equipment advice - lighting too. Thank you.
Thanks Robert!
Which version of izotope RX are you using? Is it basic, Advanced or one of the studio bundles? Thanx.
This is advanced, but this is all applicable to the standard version as well.
@@curtisjudd Thank you, though my RX6 doesn’t have loudness!
I think that it’s awesome that you use Davinci. I am in the process of learning it.
I use it *sometimes*. I just mixed a short film in Resolve 16 but had a lot of crashes once we had more than 20 tracks in Fairlight. So I'm hopeful for its future, but not convinced it is ready for bigger jobs like that just yet.
@@curtisjudd I just moved from the Zoom F4 to the SD MixPre 6II. Am I going overboard if I will record a three person 1-2 hour long podcasts all in 32 bit? And what about the sampling rate? I was also wondering, if you use the mix assist plug in is is just recorded to the L+R mix? Which means that you cannot tweak all the individual tracks in post with the mix assist plug in? Thanks.
@@OliGudbjartsson12 Hi Oli, I don't usually use 32-bit for recording spoken word audio, but you can if you want. SD cards are cheap so it isn't much of a problem. I'd use a 48kHz sample rate for spoken word recordings. MixAssist only affects the mix channels, but you can still record isolated channels for each of the mics so that you can do a more manual/in-depth mix in post if you want. The MixPre gives you total freedom that way. (so did the F4, for that matter).
Did you the built-in compressor
I use Izotope Nectar, but pretty much any compressor would work the same.
Hello Curtis, Thank you very much for your video. With respect to your example, have you tried in the past using the loudness control module of RX for normalisation without using compression? I was surprised that you first compressed the audio in order to get more headroom before using izotopeRX. Using iZotopeRX8 myself I found out that the loudness control module of RX takes care of that automatically when loudness normalizing. I am not a pro, but if I set eg. integrated lkfs to -16 and true peak to - 1.5 then RX will deliver this even though the loudness increase was higher then my available headroom. RX will make sure that the peaks of your waveform will not pass the true peak limit while increasing the integrated loudness lkfs to the desired value. It would be very interesting to hear whether it also works for you.
Hi Vali, yes, it certainly works, but it tends to limit too heavily and that can be heard if it is too heavy. That's why I compress first so I have more control over how the final audio sounds.
@@curtisjudd Thanks, that was very helpful. Since I do mainly dialogue with very few wafeform peaks that need attention I never noticed heavy compressed audio by RX but very good to know that compression can be necessary in other situations.
@@valis1680 It is often better to use compression with dialogue, too. Just depends on how dynamic the voice is that you've recorded.
I'm using a Canon G7X Mark II with no external mic. (they're ordered) My question is where do you generally have your computer speaker volume set and then your computer sound volume set to begin editing? In DaVinci...
Volume and loudness are not the same thing. For your speaker volume, you should set it to a comfortable listening level. If you want to get technical, it should be at or less than 78 dB SPL.
@@curtisjudd Thank you.
thank you 8 years later chris !!
👍
Thank you so much for the video. but I getting error message after hitting Izotop ' ~ MixL_IZOTOP_n898B.wav' could not be opened. Product Portal cannot open files in the 'Waveform audio' format. Any advice? Thank you~!!
I'm sorry, I don't know how to solve that. Probably best to contact iZotope support. Best wishes!
Hey Curtis, I've been studying your videos all week and been learning a ton. Thank you. For Loudness Normalization, would an easier way be to just put in the loudness normalization settings in the Adobe Premiere Pro export menu? I see you can also specify the true peak value (like at -1 dB), so nothing clips.
You can do that, but often if you're normalizing for web to -16 LUFS, you need to do some compression first so that the audio doesn't sound distorted. Sometimes if you apply a limiter, like the export functionality does, and the audio hits up hard against that limiter, it will still sound distorted. Best to use a compressor first to manage the waveform peaks and get it sounding smooth while leaving enough headroom to get the loudness target you need.
Hey Curtis, thanks for the video! One question, is there a specific reason you apply the compression to the entire mix instead of just the dialogue track beforehand?
As part of the mix, you can apply compression to the dialogue before bouncing and doing the loudness normalization process. Then I'd do a final pass of compression on the loudness mix just to ensure we're not limiting any transients too much, IF NEEDED. Often when you have a more complex mix with lots of overlapping tracks, their sum ends up producing larger amplitude than expected so this still needs to be managed at the end with a compression or limiter pass.
@@curtisjudd Great, thanks for the clarification! I purchased a new mic recently, and as part of my research I scrolled through comments you replied to on old videos. I found the extra follow-ups you leave in the comments very useful.
@@SpokenFrameMedia Thanks for the feedback! Happy recording with your new microphone! 🎙
Great Job as Always 👍 Love these type videos 😀 A little surprised you did not run the iZotope Phase tool?
That was outside the scope of this topic. Typically I run it on dialogue tracks if the waveform is asymmetric.
You're a resource!
👍
Thanks for the video. Just what I've been dabbling with the past couple of days while finalising an online course. Btw great to see some Resolve examples.
I do it a little differently (apart from the fact that I don't own the big version of RX). Namely I try to avoid compression on the full mix. The reason being that I am using stock music which was already heavily compressed by the composer, and I don't want to compress that even further.
Instead I add Ozone Limiter (set to -1.1dB True Peak) and Youlean Loudness meter (with TH-cam preset) to the mixbus. Then I increase the level in Resolve (using BX Masterdesk as a glorified master fader) until Youlean shows me I am between -15 and -14dB (recommended for TH-cam).
Finally I check how the aforementioned music track is doing LUFS-wise when it is playing at full level. Usually I have to pull it's level back until Youelean gets below -14dB again.
And lastly I drag the final export on to the standalone Youlean meter app which calculates a LUFS curve chart for the entire video plus indicating any Peaks above -1dB (YT setting) in case there are any.
Thanks for the insights Heiner. I find that most popular music is already HEAVILY compressed as well. What that usually means is that I have to attenuate it quite a lot to get it sounding right with the dialogue and then it basically never gets any further compression when I work on the final bounce/mix because its transients already leave plenty of headroom.
I was super excited when you announced this video, but even though I have a copy of iZotope RX8 Advanced, and despite decades of reading everything about audio production, I feel helpless to grasp how to use these advanced tools successfully. It seems impossible to be self-taught on this topic. I'm desperate to learn these tools, but just cannot figure out how to approach the task.
Hi Danny, I'm hoping to make a course to cover this in more detail. But the main idea here is just using compression to make enough headroom so that you can boost everything up to your target level (LUFS).
Thanks for another amazing video, Curtis! When you do your mix, do you first compress the vocals, then once everything is mixed, compress the whole mix again (as shown in this video) for loudness normalization? If so, when you first compress the vocals, how do you know how much compression is enough? I thought we only had to compress the vocals. Ahh, this compression thing is so complex...grateful for your teaching in demystifying it.
Generally I just compress at the end, after everything is mixed, to prepare for loudness normalization. But you can do it in stages if you prefer.
@@curtisjudd Gotcha, thank you! One question I have remaining is that of where to place the loudness levels of dialogue, music, sfx in your mix. I'm confused because I learned dialogue should be bouncing in between -12 to -15 dB, but then I saw you normalized your dialogue to -0.1 dB. If that's the case, where should music and sfx generally be sitting at?
@@curtisjudd Thank you!
iZotope rx is too expensive. How to perform the same operation using Aсon Acoustica? I didn't find a way to set the headroom target value and the Integrated LKFS target value when trying to normalize.
Yes Please Curtis, Adobe Premier & Audition.
Good news, there are several options covered in these previous videos: th-cam.com/video/RrRn8stkEBs/w-d-xo.html th-cam.com/video/OKSWPrT5upo/w-d-xo.html
Hi, Curtis! Looking for plugin with pre-setted loudness target lvl, for automatic loudness correction on master in DAW, before render. My vst chain is close to target lufs, but every 3-rd mixdown I fain to re-level and re-render just becouse mix appear louder by 1-ich dB (.
can U reccomend some?
I use the loudness module in Izotope RX but it isn’t a standalone plugin.
Vegas pro is excellent audio editor.
👍
Izotope RX Advanced costs way too much comparing to Resolve Studio. Could you show how to do loudness normalization with the latest Resolve Studio Fairlight only?
That's the thing - it is a fair bit MORE work in Resolve. I'm going to cover that in my upcoming Fairlight course.
Debating acquiring RX8 - if I do mostly dialogue editing, would the Standard edition have the tools I need? I already have Nectar 3 - wicked tool.
I have not regrets making the purchase for advanced for my corporate and short narrative work. You can see the differences here: www.izotope.com/en/products/rx/features.html#comp and if you go for the standard version, you can always upgrade later.
So, correct me if I’m wrong, but you’re saying the parts of the waveform which are lower db (lower peaks) we’re perceiving as the same loudness as those parts which are higher db (higher peaks)? I don’t understand why you wouldn’t want a waveform that looks more uniform across the timeline. Or are you saying that you don’t want the listener to hear the same loudness throughout and you’d rather them hear the louder and quieter parts of the voice? Or is it that having the same peak throughout the entire waveform is bad because we’ll be perceiving higher pitched parts differently than lower pitched parts?
Clarification: Waveforms with large amplitude (tall) only sound loud to us when they are sustained for more than a few milliseconds. Smaller amplitude portions are always quieter to us. But we only perceive something as loud if the amplitude is tall AND the tall amplitude lasts for more than a few milliseconds.
So in that context, when you're finishing/mixing dialogue audio, you first need to decide where the audio will be heard (online, TV, theater, etc.) and make choices about how "loud" you need it to be. If the audio will be consumed online, that means people will often be listening with less than ideal circumstances - in cars, trains, planes, and listening with phone speakers or ear buds - you'll probably need to normalize it to a louder level. And to do that you generally need to reduce the dynamic range of the audio. There are several tools for that including levelers and compressors.
For TV or theater, the listening environment is usually better optimized. So you can retain more dynamic range and do NOT have to normalize so loud.
I hope that helps.
@@curtisjudd Yes that helps. Thank you. Just to make sure though, going through and doing it manually on your timeline with the pen tool and adjusting the levels is the same thing right? These tools are only designed to expedite the process?
@@carlsayers7370 You could use clip gain, yes, but that can often end up sounding rather odd if you've got any substantial noise floor because it will obviously jump to. louder levels. So I usually try to avoid any sort of extreme automation or clip gain like that (with the pen tool, as you said). Yes, compression automates this. And the critical tool is a loudness meter, regardless of which way you choose to optimize your loudness.
Appreciate the video!
👍
Running into a few issues and I'd love some help! I'm editing podcast audio and consistently run out of headroom to hit -17lufs even after compression. Anything helps! Thanks!
You'll need a touch more compression.
@@curtisjudd Thank you for getting back to me! I'll keep massaging it!
Really good explanation but there's something I don't understand. If you keep it at -16 LUFS (with -1.9 headroom) and youtube normalises to -14 isn't it going to clip once the audio is turned up? I understand the reasoning from the other video to conserve dialogue audio quality but I think this method could be a false friend. If you don't compress the audio enough and don't leave enough headroom you could inadvertently make youtube clip your audio once it turns it up.
Maybe I'm missing something here. Either way this was a really good video!
Hi Bruno. TH-cam does NOT change the loudness of your video unless it is LOUDER than -14 LUFS (i.e., your video would have to be -13, or -12, or even louder). If your video is louder than -14 LUFS, it reduces the loudness to -14 LUFS. So if you normalize to -16, TH-cam will leave it at -16 LUFS.
So no false friends here. Just normalize to -16 LUFS and you are totally safe.
@@curtisjudd That makes so much sense!! Thanks for your reply. My background is more in music and since Spotify does normalise it's risky to go under because you don't know what they will do.
Once again thank you for your response and the very informative video. I'm really enjoying your RX tutorials.
Great video thank you
Thanks IVE595
Another way to do this in rx and DaVinci is to just do the round trip with just dialogue/VO track and normalize it to - 18db. And then bring it back I to DaVinci and use the eq and faders in the music and effects tracks to target - 16. This works well because I often find myself using RX for other things in the voice so I only have one round trip.
Yes, and I would suggest a final loudness check to make sure the final mix sits where it should. But yes, that should get you very close in most cases.
I do the final loudness check in resolve using the loudness line graph below the master track. That way I have a visual of loudness over time. I just run it while doing my final check before rendering. If there is an issue I can see where and then fix that part and spot check it.
Could you show this process for a more challenging situation: Live streams / live podcasts?
Still working on that. For live, most people still just use a compressor or limiter and push the levels up until it sounds loud enough or meets an RMS level. The trick is to put a loudness meter close to the end of the signal chain. I haven't found a good way to do that yet since I stream directly from an ATEM Mini.
@@curtisjudd Kudos for the lightning fast response! I see how that makes it more difficult. Perhaps it's worth a shot to recreate the workflow in software as a showcase? Either way I'm curious what you'll come up with. Good luck!