The best oversampling... what should you pick when mixing and mastering? 

So you wanna tell me the x256 oversampling from Standard Clip is unnecessary? 😲

It's important to consider the base rate when choosing oversampling rates. I assume this video was made from the perspective of a 48kHz base rate but it wasn't mentioned. 96kHz projects would only require half the oversampling that a 48kHz project would need to get similar results.

I think its important to think about 4x what? 44.1k? 96k? 200000k? It changes the answer and also the potential for ultrasonic information to cause distortions and lowered headroom. I almost wonder if it makes more sense to think about oversampling in terms of frequency rather than the multiple.

You should make this into a series for replying back to comments you make a video on!

I use 4x oversampling while producing, but 8x or 16x when rendering. in the menu you can select that under "offline"

This is such a cool video, so glad I found your channel recently!

Hi! I was surprised you like using TP limiting, i usually don't like how TP limiting handle the transient, but i have to say one of your last video about how limiter can be disturbing on the low end was very informative and i need to give it a try ! Anyway thanks for what you are bringing for the community !

A question on true peak limiting : Many engineers says it kills transients a bit and regarding facts, they argue that most of todays music .wav files true peak above the 0db. And that "the louder the better". On the other hand, I understand true peak mode is here to prevent distortion after some conversions in the analog domain. I tended not to use true peak mode as most pro music do not care. Tell why I'm wrong, tell me why I'm right ! Great content as always tho

Max attack and min release is a documented and well-known pair of settings for L-2. You wouldn’t push those settings 10db though, but those settings specifically on L-2 are often used for bass scene and similar genres, it gives a very unique sound, and it still can be clean if you mix your track well enough.

I choose oversampling settings by the feel of the sound and the groove and strangely each plugin manufacturer sounds different. Standard clip and Acustica plugins I prefer the sound of 4x. SSL native bus comp I prefer the transients on 2x. TDR precise setting. All other plugins off. I find that most other plugins mess with the groove, I'm not sure if the latency isn't compensated properly or something but the groove sounds tighter when off, especially Lindell plugins I use their API plugins a lot and they sound better when off. I don't care about the technical reasons and aliasing etc, it's how it sounds and feels that dictate all my choices when mixing. Sometimes theory and reality are different, so don't overthink art. If it sounds good it is good.

I'm not sure if you mention it later in the vid, but the way you are doing the test is as if we are assuming that all processing is done on the master/mix bus, or a single bus/track only (which makes sence from a mastering point of view). However it would be diferant for say a mix where saturation etc. is being used at every stage, even if at 4x on the master it's unaudiable, when many layers of it are added together it may later become audiable at 4x, so a higher rate may be nessisary when using it on multiple sources, rather than just on one channel. This is especially true if processing the same source material at multiple points in the chain (track+ bus+ mix bus for example) as each time it will be creating more of the same frequency aliasing distortion (as well as making all the alisasing distortion from before alias it's self too!) :O. It can become quite a mess when pushing it intentionally for experimentation and this could actually make your top end/high mids become a bit hazy in highly processed projects due to so much low level stuff going on beind the music, and aliasing on top of aliasing ontop of aliasing. It's especially noticable when using the same processing over and over again as it produces the same distortion profile (another excuse to use a bunch of diferant saturation types in your mixes ;p). Also the harder you are hitting something with clippers/saturation etc, as well as the base samplerate can also make a big differance to how much alias distortion it produces and how much OS you may need. Oh, and the lower the freqency that's being processed, the more aliasing distortion that will fall into the audiable range at the same OS rate. All things to maybe consider.

Not for nothing, but even on YT I could hear the aliasing at 4x... only at 8x did it go away. Also your take aways on limiting were quite interesting! Thank you for the content! Keep up the good work!

Alias… Anti-aliasing. There’s no word called “Ali-izing”, that’s not what it’s called. Cool demo though!

I like x8 on limiters and clippers if they have the option. Past x8-x16 it doesn't get any better to my ear. But I can hear improvement going from x4 to x8 in some songs especially when the limiter is hitting constantly during the loudest section of the song

Funny thing, I did the exact same thing out of a gut feeling, like you did. Always oversampled my clippers 4 times. Happy I was in the right ballpark. Anyway though. You once said in a video you’re oversampling your L2 by 8x in TP mode to avoid intersample peaks at a -0.2db ceiling , so I’m slightly confused, why you’re now saying that you don’t need to oversample your limiter…

Do I need to use oversampling in the clipper if I am using oversampling in the limiter? I figured I can save DSP if I could use one plugin to clean up all the aliasing from plugins from the mix and mastering process.

Is it only me who often prefers lower / 0 oversampling over higher rates? Sometimes aliasing artifacts feel like less of a price to pay than the slightly smeared transient response I seem to get from oversampling. Perhaps it’s the linear phase filters but I don’t like what it does to my transients a lot of the time

Love Blue Blockers. Highly recommend Felix grey. Theyre a little pricey but I love them- theyre totally transparent too, I had to run some blue light through them to prove to myself they work lol. No more eye strain and they last forever (ive had mine for years now).

So do you prefer to use true peak or oversampling (or maybe both) while limiting? Great content btw keep it up 🙌

If I'm not mistaken, Maximizer oversamples 4x when soft clip is enabled. Otherwise I don't actually know if Maximizer oversamples

4x is probably fine, as the distortion is program dependent.. we will almost never hear static aliasing tones. -- Important because are ears are very good at picking up on things 'not like the others' ; in this case ever changing music, and static tones .. I might try to reproduce this with standard clip, but feeding a dynamic program.

KClip has offline processing. I have it set to x32 atm which I agree is over kill. x4 is plenty x8 max. I belive x32 is unnecessarily slowing down my bounces Nice point about using a limiter in tandem

So isn't it 'safer' to just use maximum oversampling right before rendering? Turning it off while mixing and on while rendering.

what do you think of the newish " antiderivative anti-aliasing" algorithms (ADAA) ? it solves the aliassing problem better than just opersampling alone

Where do I get this shirt though?!?!?!

The true question is, can anyone really tell by listening to the master, when oversampling has been used and when it hasn't?

i tried x256 once on sir and it blew up my house

You can exclude oversampling altogether if your session and rendering sample rate is 96 instead of 48. The end result: Aliasing Distortion and Nyquist ceases to exist.

OMG this shirt is so nice. Where do you got this?

I have a question, when rendering does it matter to activate oversampling in each plug-in that has it? Or is the last plug-in like Masterplan at 4x to 8x enough? Right now I've been activating ovs in each plug-in.

I was analyzing one of your masters on Spotify (The Bad Touch - Pelican, Tribbs). I have the browser audio running through span, and it says the True Peak was reaching 0.5 db. Do you know why this would be the case since you use true peak limiting? 🤔 I'm just wondering if using True Peak even matters if it's still going over anyways.

If you can't hear aliasing at -50db then you shouldn't be mixing or mastering.

i was playing with oversampling yesterday i crushed a mix with standardclip had 1 on 2x and the other on 32x. i then set it up so it would randomly switch between which one was playing. i honestly could not tell when it was switching. i tried to look whenever it was on 32x and got it wrong quite often. it surprised me because ive had mine on 16x

I'd venture as far as 4x is overkill on a clipper. It looks nice on an analyzer, but I'm willing to bet everyone would have an EXTREMELY difficult time to pick out even 1x vs 16x in a blind test. I urge everyone to do this at least once. Can you identify it more than 8 times out of 10? It's a nice reality check, i do it often. Another point that many miss while analyzing aliasing on a scope, is that the artifacts only get audible towards the top end over 8khz or so, and any natural music have much less energy the higher you go up in freq. Even with a cymbal crashing, it is practically nonexistent to have a sound as loud as those analyzer test signals at 10k... or you have much worse problems with your mix than aliasing.... But Standardclip is very easy on CPU, so why not indeed....

Is your session in 441 or 48, looks like 48 but then the signal at 14k is definitely close to the Nyquist at a high level, not practical unless you’re mastering a day at the beach lol. But for example 👍🏼 ok. Mixing with phon values should keep hi frequency to a minimum compared to mids n lows, less risk for foldback. Higher sample should be determined by case, like vocals or guitars especially with sessions under 88.2, or overdriving plugins. 96k sessions get a bad rap for some reason but it’s the best IMO because of Nyquist, time stretches, or pitch related plugins, oh and plugins like Sound Toys and similar are optimal. My 2cents as always and great video performing live test. If you came up with this test on your own then congrats in order, regardless of what I say or others, you’re definitely a must for beginner to intermediate, and entertaining for pros as well. Do a video on Dc offset, and how to fix with eq or rx hum. Also how it’s created and it nasty nature for subs, headroom, and shifting your stereo balance etc etc. ✌️

Is there any moment where oversampling to x1024 is usefull ? (I only saw this option on Acustica Audio plugins)

Wud you spend $249 on Gold Clip when you have Standard Clip? Are the differences enough to justify that price?

What sample rate are you working at?

I always go as high as my system can handle. My transients are still popping though. Should i be doing less that 8x or 16x?