Man… what a blessing this channel is. I learn something really valuable with every new upload. Thanks Dan, you’re genuinely making a really meaningful contribution by sharing with us.
"Honey I shrunk the guitars" LMAO that's why I love these videos of yours, not only educational but also filled with this kind of jokes. Will never get tired of watching your content.
01:25 *Does it work?* - 04:11 _Phase_ - 05:11 _Null test_ *Tests:* 06:58 _ReaComp_ - 08:16 _Decapitator_ · 10:08 - 10:59 _Ultra 530_ - 12:34 _Twin 2_ · 14:03 15:26 *Chain OS on all tracks?* - 20:12 - 23:41 *Summary & Outro* - 00:00 Introduction 01:25 Does it work? 04:11 Linear phase 05:11 Null test 06:58 ReaComp for waveshaping distortion 08:16 Soundtoys Decapitator 10:08 The plugin has to support stupidly high samplerates. 10:59 Mercuriall Tube Amp Ultra 530 CME 12:34 FabFilter Twin 2 14:03 Twin 2: oscillator sync 15:26 *Chain OS on all tracks?* 16:43 plugins that don't benefit 17:52 internal OS 18:35 individual > chain OS 20:12 Where chain OS makes sense 23:41 *Summary & Outro*
This is amazing. I was already starting to make a slow transition from Protools to Reaper, but this seals the deal, I can’t wait for my primary DAW to be made by a company that actually wants their product to be “good,” and puts recourses into making it so, rather than just putting them towards cornering the market so people feel like they don’t have choices
A few days ago I asked one youtuber sound engineer why he is still on ProTools and he said that most studios and production houses use PT so he is forced to use PT if he wants to cooperate with them. I hope one day finally professionals will abandon that overpriced platform to work with devs like Reaper and let them grow...just because Reaper is a hell better than any DAW and is an open platform, any dev can write their script and add new functionality to their DAW. Welcome to Rea World :)
Start your own plugin company, hire a couple of DSP/JUCE guys, you explain, they code. With the amount of vision you accumulated I'm pretty sure end products will be amazing.
I know there was some criticism regarding your own music in the comments recently. While I’ve always liked your rather ambient/psychedelic music tracks that often were excellently crafted to demonstrate the topics you were discussing, this one is just straight up an absolute banger!
The real star of many of your videos, for me at least, is your music. I often pay no attention to your thesis because I’m too busy finding my way through your music. Beautiful stuff.
Please never stop making videos, no matter how in depth and intricate the information that you're sharing is, someone, somewhere at the edge of the world just gets it, as good as you do and they sleep in relief, knowing that. And because he understands what's up, everyone before him who discovered and denied you, now has evident reason to believe in this level of passion and love you put into sonics.
REAPER Devs straight killing it! Illuminating point about making a script to just trigger oversampling on everything. Never thought of it like that. That said, would be cool to see them add an oversampling action to the FX bay. You could go and search through your project by plugin name and batch trigger oversampling on those during mixdown :)
If you're going to oversample everything, you may as well run your project at a higher rate and then you only need to downsample once to your target rate..
Dan Worrall: putting the engineering part back into audio engineering. BTW that was the exact feature I asked in the forum a couple months ago, and I was satisfied to hear that it was already in beta testing.
Hi dan worrall im new to your channel... Didnt realise til i found your channel how nice it is to have videos on mixing etc that actually helps me to learn something new and i genuinely cant thank you enough for your content. Its weird to me that helpful honest content can seem to be such a novelty on youtube, seems to be trending gear/plugin porn mostly. Thanks for demonstrations that are practical and help me learn more interesting things in this challenging arena. 👍
I got Reaper-pilled by Dan. Left Logic behind for mixing and tracking duties. Barely open Ableton any more. Try it, kings! Hard to set up properly but worth the effort.
I don’t think any other DAW’s would dare implement this. For a less tweaky “power user” DAW they’d get inundated with support requests about plugins that don’t work right with it (as demonstrated in this video). I can see third party developers not liking it too because now their plugins that don’t support a 768kHz sample rate seem broken to users that don’t know better. They don’t support a big feature of the host DAW. All of which is to say: for those that do know what’s up, Reaper is a winner ;)
@@mrnelsonius5631 i'm pretty sure bitwig will have this at some point. bitwig has a lot of container tracks, which are basically plugin-wrappers with special functionality like multiband-splitting, mid/side encoder, parallel-processing etc. even the simplest container, the chain, is already so much more useful than any other daw's plugin chain, because you can apply modulators to it that then can act on every parameter of every plugin in the chain. you often end up with a single automation that controls so many things that the evolution of the sound is just amazing. ok not saying it's definitely "better" than reaper. reaper has a lot of nerdy details that make this daw unique in its own way. but if there is one daw where i expect oversampling to become a thing as well it is bitwig.
@@Beatsbasteln Ableton also has a kind of wrapper setup with Effect Racks. I use them for complex routing and parallel processing; essentially making my own unified plugin window with multiple processors. It’d be really easy to implement OS there in a way that only knowledgeable users would tinker with, but I’m not holding my breath ;)
@@gabrieldoudna6570 yeah. i think reaper even has the most flexible routing options of all daws. it's just that they are hidden in weird menus sometimes
I might've made the feature request you referred to (or a similar one). The intention isn't to toggle oversampling on/off for every track, just where it was manually turned on to save CPU while mixing similar to how many plugins with OS have separate options for real-time and rendering. Ideally Reaper would just include real-time/rendering OS options, but this was a possible workaround.
I don't think you mentioned this, but it looks like you're using the r8brain resampling algorithm based on your results. I think the default is something else, so for people to get the same quality results, they need to go to project settings and change the playback & rendering resampling mode to r8brain. You can reduce latency at the cost of less-good antialiasing by using some of the lower quality modes for the playback resampling mode. Personally I still think it's better to just track without oversampling and then add it for mixing, where necessary.
honestly in most of the demos I prefered the sound without oversampling. the aliased frequencies to me are like an extra layer of crust and crunch right in the frequencies where I want that out of a distorted sound.
I wonder if they could add a feature to specify a certain set of plugins within the chain to be oversampled as a chain so you could use that where appropriate without messing up the rest of the chain. Subchains, basically, or folders for specific groups of plugins within a chain. It's a pretty niche use case, so I can see why they wouldn't bother, but it could still be handy.
You're honestly such a legend, so much technical knowledge and still always focused on making sure it actually sounds good instead of just preferring something that's "objectively" better
I really love what you, David Vignola & Michael White do, although you are all slightly different you all have a great way of teaching which translates to all levels of understanding and I'm currently working my way through all your videos. Thank you!
I don't claim to understand _half_ of this. But, as always, you left my jaw on the floor. Maybe one day, I will learn something from this. In the meantime, I just enjoy being absolutely flabbergasted.
This was such a great lesson on oversampling. Thank you. I noticed it in Reaper, but I didn't really understand whether I'd need to use it or not. It was great to hear some of these examples.... and I mean hear. Sometimes someone will demonstrate some very subtle mixing technique, and my ears are too untrained to hear the difference. Twin 2 was a cool example... the difference was fairly dramatic. I guess the lesson here is just try it out on various plugins and listen for enough of a difference. And I think I'll steer clear of chain oversampling. It was also very helpful to get the explanation about the plugin's max sample rates, and things like filters not operating well above a certain rate. I had tried Reaper's oversampling and got weird results from a couple of plugins, and I'd have to say that was due to them not supporting high sample rates.. at least not 768khz.
For anyone without a DAW that can do this (so basically everyone not using Reaper), I recommend Blue Cat patchwork. I recommend it anyway for adding parallel capabilities to any plugin(s) as well as using formats in Daws that don't' support it, like VST3 and AU in Pro Tools. It has 3 quality modes of oversampling up to 16x, and it's *per* plugin fully configurable, not a chain. A must have plugin host/chainer. BTW there's also a standalone mode so it can turn your PC into a hardware synth.
I loved using it on every track for convenience, but it took to much cpu that way. They should have made a mini DAW with Patchwork as the base for live setups.
@@christianwn Neat idea. Patchwork is amazing but he won't implement a plugin browser and relies on the OS which is the one big flaw with it for me. I use Plugin buddy or Studio Verse now unless it's a feature of Patchwork I need. They are all great in their own right, Metaplugin is fantastic too. I have them all. BTW Studio verse and PI Buddy are 100% free.
When chain-oversampling, it could be very beneficial to put nice low-pass filters in between the layers. This way we stop most of the frequencies producing the aliasing, but we'll safe ourselves the upsampling processes between the layers, giving us the best of both worlds. Sam Fisherman suggested this in the interview on "White Sea Studios" channel
Great test! I'm glad you spend time making these! One thing I'd like to know though, is the CPU usage the same if you oversample each plugin in a chain vs the entire chain?
Individual oversampling will require more CPU cycles, all else being equal, as you have more AA filters running. But probably not significantly so, and this is easily offset by optimising the oversampling factor for each stage. If you stick to 2x or 4x for compressors and gentle saturators and only go 8x or 16x for the harder distortions, you'll probably end up using less CPU than oversampling a whole chain to accommodate its worst offender.
Thanks Dan !!! I have one question: can oversampling affect the quality of reverb processing, and is it worth to do it ? How linear phase type of oversampling will affect the operation of the reverb ? I'm very interested in your opinion.
That's two questions. Answer to both: maybe. Unless it's convolution reverb, in which case it's just a waste of CPU. If the reverb is 100% linear (like convolution) also probably no benefit. If there's a saturation stage, that will sound different. Possibly modulation inside the tail could be affected. But don't assume it will sound better. Some of the magic of early digital reverbs comes from the limited bandwidth and naive interpolation adding extra richness and depth to the tail. Valhalla Vintage Verb provides bandlimited 70s and 80s modes that aim to recreate this sound: oversampling this would be kind of pointless as the plug-in downsamples to the same rate internally anyway.
This is great for favorite older vst synths like Waldorf Largo that sound noticeably different (like you said, not necessarily better) when oversampled, especially when abusing the FM capabilities. EDIT: LOL just got to the Twin2 part
Yep! I always A/B oversampling when I am thinking about using it. There are many many times I actually prefer that aliasing. Totally program dependant!
Great video! Interestingly, in most situations where I actually want saturation or heavy clipping, it tends to be on sounds where the fundamental frequency fairly low, so the lack of AA on the Soundtoys plugins never really bothered me. Nice to have this option, though.
*laughs nervously in FL Studio’s “max available oversampling during rendering” option* As a side note, Mercuriall’s guitar sims are worth more exploration, if this was really the first time you’ve looked at their stuff. It’s not all super high gain and their interfaces have come a long way since the demo version you used. Can recommend having a look.
The Reacomp and Decapitator examples…. I vastly preferred the OS audio- which surprises me because I don’t typically worry with aliasing too much. They immediately sounded more analog to me as well, which is a description I don’t typically think means much!! 🙃🤷♂️
Dan, that's a great video and I'm sure different people will get different things out of it. I especially liked the section where you demonstrated what an over-sampled Decapitator would sound like as the difference was reasonably clear and if you had that on a number of tracks then the cumulative effect would be quite obvious.
16:38 It's not as though one renders a mix and then sends it off for distribution before listening to the results. Whether the effects of the oversampling are evaluated pre- or post-render is of little consequence. Rendering first and listening second strikes me as a perfectly valid option if the session is unplayable in realtime. Moreover, it's far easier and smoother a process to compare multiple stereo mixes than it is to toggle effects in and out of bypass in a session. There's no question that oversampling convolution reverbs or other linear effects is waste of CPU cycles. But some folks have cycles to spare, and if you're rendering offline anyway then the extra time will be insignificant, especially with a late-gen CPU. My point with all this is that experimentation is its own reward, so if curiousity compels you to oversample every effect, track and bus in your session then do so. The worst that can happen is you'll come to your senses and turn it all off because Dan is usually right. ;)
Very nice, I started using Reaper because of your vids, I also thought I could rewire audio to use vsts in pro tools (rip). I've been tracking and mixing in pro tools then mastering with Reaper, it's been great thanks again!
I was hyped about this feature more than I'd like to admit, however - I /think/ that the filter used to cut the hypersonic content has slight ring to it (just my ear for now), and I'm rather surprised that they didn't include nerdy stuff like resampling modes and filter selection for online/offline processing. I hope they will include more features in the future //edit Of course you were testing the filter ring...
I was about to mention that the aliasing can sound cool on crunchy drums and then you said it. Also if you just eq some sizzle out... You'll have lessened some high sizzle and kept anything else you wanted from it.... If you wanted it. With synths it's often very horrible but sometimes it's what you need as long as you can tame those nasty Hz. You've showed that too. This is a great video, it's great for people to sometimes recognise there's aren't so much as rules, but guides in music production 👍👍
great new feature. what's a bit surprising with the drum loop demo is that it really improved stereo separation to me. it's clearly got some less distortion, but it also sounds much wider and the elements can be clearly picked out in the stereo image, rather than being smeared into a more mono image.
Man, do rewatch this for Dan's music alone! It's just fantastic (e.g. near the end, but much better from around 14(?) min, also for the thrilling process of developing it).
@@briefcasemanx definitely not the same. More bandwidth is more IMD. Oversampling is less bandwidth because you're filtering out those higher frequencies. Pretty basic stuff and exactly the reason why working in 48khz is best.
@@RutgerS. they are effectively the same. I don't care about intermodulation distortion that's below the hearing threshold. Intermodulation distortion from high sample rate is not a big deal on real world signals being exported at 384khz (again, real world signals, not talking sine sweeps up to 192khz or whatever). Here's a rather contrived example that shows distortion with THREE separate saturation stages and it's still too low for me to care about at 192k no oversampling. I export at double that, so it would be even lower for me. Check at 20:50 for the example: th-cam.com/video/-jCwIsT0X8M/w-d-xo.html I still oversample distortion plugins if they have oversampling. If they don't have oversampling, I don't worry.
It appears that it works with UAD but does eat DSP and it won't let you run different oversampling. Example: On track 7 you oversample the UAD 1176 and then on track 27 you want to run a non oversampled Lexicon 480, it won't let you. They have to match...ugh...back to 96k
Great work as always Dan. I really appreciate the time and effort you put into these videos. Incidentally, can you make a new version of the backing track for this video where you replace the "Bum Bum Buuum" with "Dan Dan Daaan"?
Hello Dan, first off, great vidoe as always. Since Reaper does linear phase, from my limited understanding, that's not always needed. Does it take advantage of minimal phase oversampling which in turns takes a lot less cpu power. Something to think about. That's one of the reasons I love SIR's Standard Clip.
thanks for another amazing video! I have been thinking about trying reaper for some time, and i wanted to ask does reaper have the option to use post fader effects without having to send it to another channel?
@@DanWorrall i use airwindows console when mixing, and i love the effect it has on my tracks but unfortunately i cannot use my faders because they require unity gain to work properly. the only daws i’ve found that allow post fader effect slots are cubase, ardour, and harrison mixbus. i’ve tried ardour and mixbus and they don’t seem to be stable enough to migrate over to, and i’ve been avoiding cubase just because of the price but i think that’s going to end up being my best choice.
Hey, good video and good new feature :) So, right now I am working on a project in 96khz with Reaper, all audio was recorded at 96. There are amp simulations, distortion plugins and just basically a big load of effects. Now I was wondering if there is any need for me to oversample? I watched your other video about sample rates, but I am still not 100% sure what the best solution is for me. Also, because the project is in 96kHz, it's very CPU demanding, so it reacts pretty slow, which ist still tolerable but not ideal. When I choose 48kHz in project/sound driver settings it will resample all media, right? I tried this and when I just change the sample rate without any further actions the sound is very different and definitely worse, much less subtle of a difference then in your examples here, and I think it was aliasing. I then tried to stick to 48 but oversample every plugin that has the ability to do oversampling for a test. Which were a small proportion of all plugins (the fabfilter and some izotope stuff), and it didn't make a difference, or not big enough of a difference. (maybe something like the not oversampled Decap is the problem in that case, which would make sense) So two questions, would it be less CPU demanding to have the session downsample the media material to 48kHz by Reaper and then oversample most FXs and are there any pitfalls regarding sound quality? (for instance do I need to worry about resampling quality in Reaper or do any Low Passing manually) And secondly, how can I achieve the best quality for playback and rendering, would I need to care about oversampling plugins when I would choose to run/render the session at 96? Would be extremely helpful if you or anyone else can give me a hint about this. Best Regards
Seems like a great addition. Would love to see the over sampling compared to FL studios too. Which has had 2 4 8 and 16x in many plugins for years. Not it's compressors though unfortunately.
@@anissbenthami It was requested and discussed on the forums a little while back. And while I do not expect it for FL21, we can only hope to get it at some point. :)
Sorry I missed this one. Yes, the answer above is correct. Once aliasing has been added there's currently no way to separate it from the wanted audio again, which is why you need to oversample and catch it before it actually reflects down into the audio range.
Thanks! An extremely useful series of videos, we are waiting for the continuation. Big about tests, oversampling, aliasing, analog summing, transcates, dittering and other horror of digital.I would really like to understand which is better 24/96 or 64(2.8 MHz / 1 Bit)..
Man… what a blessing this channel is. I learn something really valuable with every new upload. Thanks Dan, you’re genuinely making a really meaningful contribution by sharing with us.
The stuff on his Bandcamp is impeccably engineered.
I wonder who did that ? ... 😜
"Honey I shrunk the guitars" LMAO that's why I love these videos of yours, not only educational but also filled with this kind of jokes. Will never get tired of watching your content.
01:25 *Does it work?* - 04:11 _Phase_ - 05:11 _Null test_
*Tests:* 06:58 _ReaComp_ - 08:16 _Decapitator_ · 10:08 - 10:59 _Ultra 530_ - 12:34 _Twin 2_ · 14:03
15:26 *Chain OS on all tracks?* - 20:12 - 23:41 *Summary & Outro*
-
00:00 Introduction
01:25 Does it work?
04:11 Linear phase
05:11 Null test
06:58 ReaComp for waveshaping distortion
08:16 Soundtoys Decapitator
10:08 The plugin has to support stupidly high samplerates.
10:59 Mercuriall Tube Amp Ultra 530 CME
12:34 FabFilter Twin 2
14:03 Twin 2: oscillator sync
15:26 *Chain OS on all tracks?*
16:43 plugins that don't benefit
17:52 internal OS
18:35 individual > chain OS
20:12 Where chain OS makes sense
23:41 *Summary & Outro*
yeah sure, thanks! but I was gonna watch it all, anyways :D
Thanks wanted to get a look at the tests again 🤘😎
This is amazing. I was already starting to make a slow transition from Protools to Reaper, but this seals the deal, I can’t wait for my primary DAW to be made by a company that actually wants their product to be “good,” and puts recourses into making it so, rather than just putting them towards cornering the market so people feel like they don’t have choices
WELCOME to the Reaper community. Buy this great DAW. IT IS INCREDIBILE.
A few days ago I asked one youtuber sound engineer why he is still on ProTools and he said that most studios and production houses use PT so he is forced to use PT if he wants to cooperate with them.
I hope one day finally professionals will abandon that overpriced platform to work with devs like Reaper and let them grow...just because Reaper is a hell better than any DAW and is an open platform, any dev can write their script and add new functionality to their DAW.
Welcome to Rea World :)
@@mariodario9033 Which youtuber was it?
Welcome to the greatness that is REAPER. A DAW for functionality and creativity vs overpriced fluff.
That was the exact feature I asked in the forum a couple months ago, and I was satisfied to hear that it was already in beta testing.
Start your own plugin company, hire a couple of DSP/JUCE guys, you explain, they code. With the amount of vision you accumulated I'm pretty sure end products will be amazing.
I think he works close enough with Tokyo Dawn and Fabfilter already :)
he already made some plugins under the name of platinumears but they're old now
@@e_cetii bilmiyordum ismail, tesekkur ederim.
I know there was some criticism regarding your own music in the comments recently. While I’ve always liked your rather ambient/psychedelic music tracks that often were excellently crafted to demonstrate the topics you were discussing, this one is just straight up an absolute banger!
I agree. bought it before the video dropped ;)
Oh how rad, Dan's on spotify as well. I had no idea he released his music until I saw this comment. Thanks.
Well stated...
The real star of many of your videos, for me at least, is your music. I often pay no attention to your thesis because I’m too busy finding my way through your music. Beautiful stuff.
the ending with that music really made me feel like we just gained some new superpower, this shit's exciting man!
Please never stop making videos, no matter how in depth and intricate the information that you're sharing is, someone, somewhere at the edge of the world just gets it, as good as you do and they sleep in relief, knowing that. And because he understands what's up, everyone before him who discovered and denied you, now has evident reason to believe in this level of passion and love you put into sonics.
REAPER Devs straight killing it!
Illuminating point about making a script to just trigger oversampling on everything. Never thought of it like that.
That said, would be cool to see them add an oversampling action to the FX bay. You could go and search through your project by plugin name and batch trigger oversampling on those during mixdown :)
If you're going to oversample everything, you may as well run your project at a higher rate and then you only need to downsample once to your target rate..
@@lambd01d for a specific plugin I said.
Dan Worrall: putting the engineering part back into audio engineering.
BTW that was the exact feature I asked in the forum a couple months ago, and I was satisfied to hear that it was already in beta testing.
The fact that I kind of understand all of this is as amazing as the knowledge you provided. So Thank you.
Thanks!
Hi dan worrall im new to your channel... Didnt realise til i found your channel how nice it is to have videos on mixing etc that actually helps me to learn something new and i genuinely cant thank you enough for your content. Its weird to me that helpful honest content can seem to be such a novelty on youtube, seems to be trending gear/plugin porn mostly. Thanks for demonstrations that are practical and help me learn more interesting things in this challenging arena. 👍
For each example, the distortion was more fun, even more clear distortion, when oversampled. Thanks for this, Dan.
22:53
from reaper release details:
if both chain and instance are set, the higher of the two is used
This needs to be in every daw
I have learnt so much since watching this channel! Thanks for all the information
Always happy and grateful when another Dan Worrall video lecture is bestowed upon us.
That Bum Bum song is AMAZING!
Aliasing is what makes the classic mpcs sound so good, and other samplers like emu sp1200, emax etc
The audio geek stuff you deep-dive is so inspiring, I'm always excited to mix again after your videos
I’ve been waiting for months for this feature to get in the official release branch! I love it!
I don't use Reaper. I'm just here to listen to Dans voice. What a voice, indeed. Makes my little Cubase soul green with envy. Makes no sense? I know.
informative as always but that TRACK WHOOOOOOOOOOOPPPP this is SO heavy... i love it
4:21 thanks for playing the spike for demonstration purposes 😂😂😂
I got Reaper-pilled by Dan. Left Logic behind for mixing and tracking duties. Barely open Ableton any more. Try it, kings! Hard to set up properly but worth the effort.
Comment because the algorithm loves comments. Love your work Dan!
this feature is sick as hell. Reaper proves itself the best DAW yet again
I don’t think any other DAW’s would dare implement this. For a less tweaky “power user” DAW they’d get inundated with support requests about plugins that don’t work right with it (as demonstrated in this video). I can see third party developers not liking it too because now their plugins that don’t support a 768kHz sample rate seem broken to users that don’t know better. They don’t support a big feature of the host DAW. All of which is to say: for those that do know what’s up, Reaper is a winner ;)
@@mrnelsonius5631 i'm pretty sure bitwig will have this at some point. bitwig has a lot of container tracks, which are basically plugin-wrappers with special functionality like multiband-splitting, mid/side encoder, parallel-processing etc. even the simplest container, the chain, is already so much more useful than any other daw's plugin chain, because you can apply modulators to it that then can act on every parameter of every plugin in the chain. you often end up with a single automation that controls so many things that the evolution of the sound is just amazing. ok not saying it's definitely "better" than reaper. reaper has a lot of nerdy details that make this daw unique in its own way. but if there is one daw where i expect oversampling to become a thing as well it is bitwig.
@@Beatsbasteln Ableton also has a kind of wrapper setup with Effect Racks. I use them for complex routing and parallel processing; essentially making my own unified plugin window with multiple processors. It’d be really easy to implement OS there in a way that only knowledgeable users would tinker with, but I’m not holding my breath ;)
@@Beatsbasteln I wish Reaper had this. It's got such flexible plugin routing but the workflow is as if it was in beta
@@gabrieldoudna6570 yeah. i think reaper even has the most flexible routing options of all daws. it's just that they are hidden in weird menus sometimes
THANK YOU!!! I`m a distorted- drumloop- junkie and this video opened my eyes !🤘
I might've made the feature request you referred to (or a similar one).
The intention isn't to toggle oversampling on/off for every track, just where it was manually turned on to save CPU while mixing similar to how many plugins with OS have separate options for real-time and rendering. Ideally Reaper would just include real-time/rendering OS options, but this was a possible workaround.
Nothing better than an informative video paired with a badass track. This guy shows deep knowledge in both his teaching and his music work. Props
16x oversampled drum loop sounds a tiny tiny tiny bit less harsh when the cymbal is playing alone. Tiny difference, but it's there!
Those little differences add up
@@UncleBenjs oh for sure, I'm actually impressed at how audible the difference is.
“With great power comes great responsibility.” Words to live by
Great video I love when you answer questions I didn't know I had
I don't think you mentioned this, but it looks like you're using the r8brain resampling algorithm based on your results. I think the default is something else, so for people to get the same quality results, they need to go to project settings and change the playback & rendering resampling mode to r8brain. You can reduce latency at the cost of less-good antialiasing by using some of the lower quality modes for the playback resampling mode. Personally I still think it's better to just track without oversampling and then add it for mixing, where necessary.
ya, ive changed the playback resample to IIR or IIR x2 setting to get oversampling in realtime
Dan Worral - you are the baddest geek on the circuit.
Excellent exploration with excellent explanations along the way! Keep up the good work, prof. Worrall!
honestly in most of the demos I prefered the sound without oversampling. the aliased frequencies to me are like an extra layer of crust and crunch right in the frequencies where I want that out of a distorted sound.
Thank you for your honesty.
Yeah and to me I preferred the opposite. All a good reminder of how subjective all this stuff is! This is a tool, not a prescription
Dan - You added vocals to your background music - Sweet! It works nicely. And the video is epic as always - especially since I use Reaper ;-)
BUM BUM (bum) BUM BUM (bum)
I wonder if they could add a feature to specify a certain set of plugins within the chain to be oversampled as a chain so you could use that where appropriate without messing up the rest of the chain. Subchains, basically, or folders for specific groups of plugins within a chain. It's a pretty niche use case, so I can see why they wouldn't bother, but it could still be handy.
This was the coolest video I've watched in a long time. And I watch a lot of them. Happy Friday!
You're honestly such a legend, so much technical knowledge and still always focused on making sure it actually sounds good instead of just preferring something that's "objectively" better
I really love what you, David Vignola & Michael White do, although you are all slightly different you all have a great way of teaching which translates to all levels of understanding and I'm currently working my way through all your videos. Thank you!
Your vids are simply top notch. A great coverage of an excellent new REAPER feature.
I don't claim to understand _half_ of this. But, as always, you left my jaw on the floor. Maybe one day, I will learn something from this. In the meantime, I just enjoy being absolutely flabbergasted.
This was such a great lesson on oversampling. Thank you. I noticed it in Reaper, but I didn't really understand whether I'd need to use it or not. It was great to hear some of these examples.... and I mean hear. Sometimes someone will demonstrate some very subtle mixing technique, and my ears are too untrained to hear the difference. Twin 2 was a cool example... the difference was fairly dramatic. I guess the lesson here is just try it out on various plugins and listen for enough of a difference. And I think I'll steer clear of chain oversampling.
It was also very helpful to get the explanation about the plugin's max sample rates, and things like filters not operating well above a certain rate. I had tried Reaper's oversampling and got weird results from a couple of plugins, and I'd have to say that was due to them not supporting high sample rates.. at least not 768khz.
Speaking of Synths with a Great UI, have you tried Vital yet Dan? Its a triumph of UI design and user feedback.
That was a wild jam on this one, that beginning was great. Good stuff!
For anyone without a DAW that can do this (so basically everyone not using Reaper), I recommend Blue Cat patchwork. I recommend it anyway for adding parallel capabilities to any plugin(s) as well as using formats in Daws that don't' support it, like VST3 and AU in Pro Tools. It has 3 quality modes of oversampling up to 16x, and it's *per* plugin fully configurable, not a chain. A must have plugin host/chainer. BTW there's also a standalone mode so it can turn your PC into a hardware synth.
I loved using it on every track for convenience, but it took to much cpu that way. They should have made a mini DAW with Patchwork as the base for live setups.
@@christianwn Neat idea. Patchwork is amazing but he won't implement a plugin browser and relies on the OS which is the one big flaw with it for me. I use Plugin buddy or Studio Verse now unless it's a feature of Patchwork I need. They are all great in their own right, Metaplugin is fantastic too. I have them all. BTW Studio verse and PI Buddy are 100% free.
When chain-oversampling, it could be very beneficial to put nice low-pass filters in between the layers. This way we stop most of the frequencies producing the aliasing, but we'll safe ourselves the upsampling processes between the layers, giving us the best of both worlds. Sam Fisherman suggested this in the interview on "White Sea Studios" channel
TDR have released a plugin for that, since I made this video. Ultrasonic from the Special Filters bundle.
Mind blown. This video is gold
Great test! I'm glad you spend time making these!
One thing I'd like to know though, is the CPU usage the same if you oversample each plugin in a chain vs the entire chain?
Individual oversampling will require more CPU cycles, all else being equal, as you have more AA filters running. But probably not significantly so, and this is easily offset by optimising the oversampling factor for each stage. If you stick to 2x or 4x for compressors and gentle saturators and only go 8x or 16x for the harder distortions, you'll probably end up using less CPU than oversampling a whole chain to accommodate its worst offender.
@@DanWorrall good insight, thanks! 🙌
Say what you want, but the image of the analyzer without oversampling looks pretty awesome!
LASERS!1!!! * pew pew *
wow the gate - kick kick- trick is what I do not expect in this video but really head me up to this comment, haha great content again Mr.Worrall
Thanks Dan !!! I have one question: can oversampling affect the quality of reverb processing, and is it worth to do it ?
How linear phase type of oversampling will affect the operation of the reverb ? I'm very interested in your opinion.
That's two questions. Answer to both: maybe. Unless it's convolution reverb, in which case it's just a waste of CPU. If the reverb is 100% linear (like convolution) also probably no benefit. If there's a saturation stage, that will sound different. Possibly modulation inside the tail could be affected. But don't assume it will sound better. Some of the magic of early digital reverbs comes from the limited bandwidth and naive interpolation adding extra richness and depth to the tail. Valhalla Vintage Verb provides bandlimited 70s and 80s modes that aim to recreate this sound: oversampling this would be kind of pointless as the plug-in downsamples to the same rate internally anyway.
@@DanWorrall Got it, thanks for answering the questions.
Your videos are very informative and helpful !!!
This is great for favorite older vst synths like Waldorf Largo that sound noticeably different (like you said, not necessarily better) when oversampled, especially when abusing the FM capabilities. EDIT: LOL just got to the Twin2 part
Yep! I always A/B oversampling when I am thinking about using it. There are many many times I actually prefer that aliasing. Totally program dependant!
I want the audio! Intriguing sounds.
I always enjoy the songs on these videos, and this one is my new favorite. Not sure what genre that would be, but it's hella groovy.
Great video! Interestingly, in most situations where I actually want saturation or heavy clipping, it tends to be on sounds where the fundamental frequency fairly low, so the lack of AA on the Soundtoys plugins never really bothered me. Nice to have this option, though.
ditto
Cakewalk way ahead of everyone. Had this feature for 10 years.
*laughs nervously in FL Studio’s “max available oversampling during rendering” option*
As a side note, Mercuriall’s guitar sims are worth more exploration, if this was really the first time you’ve looked at their stuff. It’s not all super high gain and their interfaces have come a long way since the demo version you used. Can recommend having a look.
The Reacomp and Decapitator examples…. I vastly preferred the OS audio- which surprises me because I don’t typically worry with aliasing too much. They immediately sounded more analog to me as well, which is a description I don’t typically think means much!! 🙃🤷♂️
These match my own conclusions about oversampling using Metaplugin in PT
I was waiting for this video!!!!
…and your track in the background is DOPE! 😎👍
Dan, that's a great video and I'm sure different people will get different things out of it. I especially liked the section where you demonstrated what an over-sampled Decapitator would sound like as the difference was reasonably clear and if you had that on a number of tracks then the cumulative effect would be quite obvious.
Shut up mom, there's Dan Worral on TV.
16:38 It's not as though one renders a mix and then sends it off for distribution before listening to the results. Whether the effects of the oversampling are evaluated pre- or post-render is of little consequence. Rendering first and listening second strikes me as a perfectly valid option if the session is unplayable in realtime.
Moreover, it's far easier and smoother a process to compare multiple stereo mixes than it is to toggle effects in and out of bypass in a session.
There's no question that oversampling convolution reverbs or other linear effects is waste of CPU cycles. But some folks have cycles to spare, and if you're rendering offline anyway then the extra time will be insignificant, especially with a late-gen CPU.
My point with all this is that experimentation is its own reward, so if curiousity compels you to oversample every effect, track and bus in your session then do so. The worst that can happen is you'll come to your senses and turn it all off because Dan is usually right. ;)
You just started up my tinnitus again with thst sweep
I love the vocal parts in the song.
Hey Dan, is there any chance to test Tukan JS plugins for Reaper and give us your opinion? 🙂👍
Very nice, I started using Reaper because of your vids, I also thought I could rewire audio to use vsts in pro tools (rip). I've been tracking and mixing in pro tools then mastering with Reaper, it's been great thanks again!
I was hyped about this feature more than I'd like to admit, however - I /think/ that the filter used to cut the hypersonic content has slight ring to it (just my ear for now), and I'm rather surprised that they didn't include nerdy stuff like resampling modes and filter selection for online/offline processing. I hope they will include more features in the future
//edit
Of course you were testing the filter ring...
they did, it uses the playback resample mode (it is under project settings)
@@kytdkut I wasn't aware that this is the case. Thanks
Great stuff Dan!
I was about to mention that the aliasing can sound cool on crunchy drums and then you said it. Also if you just eq some sizzle out... You'll have lessened some high sizzle and kept anything else you wanted from it.... If you wanted it. With synths it's often very horrible but sometimes it's what you need as long as you can tame those nasty Hz. You've showed that too. This is a great video, it's great for people to sometimes recognise there's aren't so much as rules, but guides in music production 👍👍
I heard this was in the works a few months ago and it's the first time I've ever got excited about an upcoming feature in a DAW.
great new feature. what's a bit surprising with the drum loop demo is that it really improved stereo separation to me. it's clearly got some less distortion, but it also sounds much wider and the elements can be clearly picked out in the stereo image, rather than being smeared into a more mono image.
I'm watching with no sound/captions on and I'm still learning so much. Dan is a master teacher. Don't do anything by default still holds true with OS.
Man, do rewatch this for Dan's music alone! It's just fantastic (e.g. near the end, but much better from around 14(?) min, also for the thrilling process of developing it).
Love the music on this one, Dan.
I hope you release the music for ACE and the Repros some day.
Fantastic job as usual of filling in missing gaps in the usual audio tutorial space! Where do I listen to your wonky music??
All the usual outlets, or here on Bandcamp: dan-worrall.bandcamp.com
Would love to see a similar review of the DDMF Metaplugin oversampling.
Great video (again). Every time I hear about these issues in the digital domain a little voice says to me "do saturation in the analog domain".
Can’t thank you enough for all the things I’ve learned from you! ❤️
Thanks Dan Worrall for being Dan Worrall
I just export at a super high sample rate. It's never messed up a mix, just makes things sound slightly better than while mixing.
But that introduces a lot of intermodulation distortion... might even sound worse than aliasing!
@@RutgerS. Not sure what you're talking about. It's the same as oversampling. The DAW just goes to a higher sample rate during export.
@@briefcasemanx definitely not the same. More bandwidth is more IMD. Oversampling is less bandwidth because you're filtering out those higher frequencies. Pretty basic stuff and exactly the reason why working in 48khz is best.
@@RutgerS. they are effectively the same. I don't care about intermodulation distortion that's below the hearing threshold. Intermodulation distortion from high sample rate is not a big deal on real world signals being exported at 384khz (again, real world signals, not talking sine sweeps up to 192khz or whatever).
Here's a rather contrived example that shows distortion with THREE separate saturation stages and it's still too low for me to care about at 192k no oversampling. I export at double that, so it would be even lower for me. Check at 20:50 for the example: th-cam.com/video/-jCwIsT0X8M/w-d-xo.html
I still oversample distortion plugins if they have oversampling. If they don't have oversampling, I don't worry.
Looks like I may go back to mixing at 48K. I wonder how it will effect UAD DSP? Can't wait to get back to the studio to find out
It appears that it works with UAD but does eat DSP and it won't let you run different oversampling. Example: On track 7 you oversample the UAD 1176 and then on track 27 you want to run a non oversampled Lexicon 480, it won't let you. They have to match...ugh...back to 96k
Great work as always Dan. I really appreciate the time and effort you put into these videos. Incidentally, can you make a new version of the backing track for this video where you replace the "Bum Bum Buuum" with "Dan Dan Daaan"?
That first alias picture looks cool af haha
Hello Dan, first off, great vidoe as always. Since Reaper does linear phase, from my limited understanding, that's not always needed. Does it take advantage of minimal phase oversampling which in turns takes a lot less cpu power. Something to think about. That's one of the reasons I love SIR's Standard Clip.
thanks for another amazing video! I have been thinking about trying reaper for some time, and i wanted to ask does reaper have the option to use post fader effects without having to send it to another channel?
No, unfortunately. But you can easily nest a track inside another to work around it. Or automate pre FX gain if that's the issue.
@@DanWorrall i use airwindows console when mixing, and i love the effect it has on my tracks but unfortunately i cannot use my faders because they require unity gain to work properly. the only daws i’ve found that allow post fader effect slots are cubase, ardour, and harrison mixbus. i’ve tried ardour and mixbus and they don’t seem to be stable enough to migrate over to, and i’ve been avoiding cubase just because of the price but i think that’s going to end up being my best choice.
hopefully you talking about twin 2 means twin 3 is around the corner and you got some info behind the scene
Hey, good video and good new feature :)
So, right now I am working on a project in 96khz with Reaper, all audio was recorded at 96. There are amp simulations, distortion plugins and just basically a big load of effects.
Now I was wondering if there is any need for me to oversample? I watched your other video about sample rates, but I am still not 100% sure what the best solution is for me.
Also, because the project is in 96kHz, it's very CPU demanding, so it reacts pretty slow, which ist still tolerable but not ideal.
When I choose 48kHz in project/sound driver settings it will resample all media, right? I tried this and when I just change the sample rate without any further actions the sound is very different and definitely worse, much less subtle of a difference then in your examples here, and I think it was aliasing.
I then tried to stick to 48 but oversample every plugin that has the ability to do oversampling for a test. Which were a small proportion of all plugins (the fabfilter and some izotope stuff), and it didn't make a difference, or not big enough of a difference. (maybe something like the not oversampled Decap is the problem in that case, which would make sense)
So two questions, would it be less CPU demanding to have the session downsample the media material to 48kHz by Reaper and then oversample most FXs and are there any pitfalls regarding sound quality? (for instance do I need to worry about resampling quality in Reaper or do any Low Passing manually)
And secondly, how can I achieve the best quality for playback and rendering, would I need to care about oversampling plugins when I would choose to run/render the session at 96?
Would be extremely helpful if you or anyone else can give me a hint about this.
Best Regards
Oh ! Was this a hint ?!!😀 Is there a Twin update in the making ?!😍
My thoughts exactly! :D
Seems like a great addition. Would love to see the over sampling compared to FL studios too. Which has had 2 4 8 and 16x in many plugins for years. Not it's compressors though unfortunately.
It's not even in their agenda for FL 21 and has never been discussed in the forum
@@anissbenthami It was requested and discussed on the forums a little while back. And while I do not expect it for FL21, we can only hope to get it at some point. :)
guys wake up new Dan Worrall just dropped
Big thanks for you video i a have question if we have file 48K 24 we suppose with Aliasing if converted to 44.1 16b the alising still there ?
Of course. Why would it suddenly disappear? If it's present in 48k, it will be present in 44.1k. 24 bit or 16 bit doesn't matter.
Sorry I missed this one. Yes, the answer above is correct. Once aliasing has been added there's currently no way to separate it from the wanted audio again, which is why you need to oversample and catch it before it actually reflects down into the audio range.
Thanks! An extremely useful series of videos, we are waiting for the continuation. Big about tests, oversampling, aliasing, analog summing, transcates, dittering and other horror of digital.I would really like to understand which is better 24/96 or 64(2.8 MHz / 1 Bit)..
Great stuff! Hope to see a separate option for "offline", like Meta-Plugin and some others.