Stop oversampling everything! (plus, higher sample rates have a reason to be)

[Ploki]

CPU as well

[plexuss]

CPU as well

Totally. This is one of the reasons I switched from Logic, which I've used non-stop since it was CLAB in the 80s, to Reaper: My workflow includes all digital instruments (no recorded audio) and lots of effects and production. In my efforts to get a 96/24 master, I'd often work in 44.1 and then change the project rate to 96k. It got to the point with Logic that this would cause crashes and other issues. I switched to Reaper a couple years ago and now I can switch back and forth with ease speed far more robust than Logic. Also Reaper is generally more efficient than Logic, or so it feels like, so I am less impeded by CPU, thankfully. However, if my project bogs down the machine in 44.1, there is no other way to recover than to start freezing tracks, bouncing down etc. Working in 44.1 and switching to 96 for a master render is my go-to workflow, if it works (sometimes it doesnt).

[metalifuxx]

Every plugin that offers oversampling should have offline oversampling options too

[Tj Shredder]

Wouldn’t it be great, if all DAWs simply could do the offline rendering in 192 kHz if you wanted? And then downsample to whatever you need? Assuming that all plugins work correctly with different sampling rates…
Which DAWs allow that without setting a different sample rate?
Actually the difference is easily audible, especially on recording. A cheap interface turns into a good one if you record in 96 kHz. The analog filters you need for the lower sampling rates need to be too steep, to be without artifacts…
Thanks to the OP for pointing out that oversampling and then downsampling again within a plugin has side effects…

[ampetrosillo]

Every plugin that offers oversampling should have offline oversampling options too

You may oversample also by 16x, but that still requires a brickwall filter at the original sample rate Nyquist frequency. That's not going to solve the pre-ringing issues you may accrue over subsequent repeated over-and-undersampling.

[Ploki]

Wouldn’t it be great, if all DAWs simply could do the offline rendering in 192 kHz if you wanted? And then downsample to whatever you need? Assuming that all plugins work correctly with different sampling rates…
Which DAWs allow that without setting a different sample rate?
Actually the difference is easily audible, especially on recording. A cheap interface turns into a good one if you record in 96 kHz. The analog filters you need for the lower sampling rates need to be too steep, to be without artifacts…
Thanks to the OP for pointing out that oversampling and then downsampling again within a plugin has side effects…

iirc Reaper can run session in any sampling rate regardless of source audio and resamples on the fly.
then you can render at source rate.

Cheap interface wont turn into a good one with 96k sampling rate. It will still have same issues re crosstalk and linearity, and uneven spectral signature

[imrae]

Running whole chains at high sample rates can still lead to aliasing if each process adds harmonics to the supersonic content generated by the previous one. To solve this problem we should also insert filters between the nonlinearities - obtaining equivalent antialiasing to a series of oversampling plugins?

That does remove a bunch of the upsampling filters, so I suppose there could be a small benefit.

[ampetrosillo]

Running whole chains at high sample rates can still lead to aliasing if each process adds harmonics to the supersonic content generated by the previous one. To solve this problem we should also insert filters between the nonlinearities - obtaining equivalent antialiasing to a series of oversampling plugins?

That does remove a bunch of the upsampling filters, so I suppose there could be a small benefit.

The benefits would be twofold: for starters, you can afford to use shallower filters at 20-ish kHz, with less phase rotation (IIR) or pre-ringing (FIR linear phase) in the final passband (<20kHz). (Also, you can choose IIR filters which have less of a sonic impact). Secondly, when you do use oversampling because the non-linearities are too strong to be adequately padded by the added octave of range you have at 96kHz, the pre-ringing effects are most significant in the ultrasonic region (which we don't care about) and reduced in the hearing range.

We aren't very sensitive to phase or pre-ringing, especially in the treble, but one thing is a single steep lowpass filter, another is something like a 60-pole, 100-pole, whatever cumulative filter which might start to be noticeable, and by combining linear-phase and minimum-phase filters an octave above this would be more transparent in the hearing range.

[Erik_Lucas]

Yes.

I've been doing this and advocating it for years. If you can work at 96k and it sounds better at 96k then use 96k. Some plugins dont work properly at anything but 44.1k because the dev didnt code properly in which case there may be no way to work at a higher res. When everything works in 96k, I mix and master in 44.1 and render to 96k and it often sounds superior to lower rates.

In my experience the golden way is to work on your preferred or final sample rate and only use plugins that internally oversample. Without oversampling you won’t rid off aliasing and if you use analogue modelled plugins then this is crucial. Best of both worlds but you need to be selective with your plugins.

This is how I do it.

Question for you all,
If the label I work for asks for the final product to be in 44.1khz, what would you mix/master at? Curious to hear what you all would do.

[chk071]

When I read all this I have to think that I've grown up in the wrong time... the music I grew up with not only had synths which aliased a lot, there was also noise from the recording, noise from the inferior physical recording media, and also had limited bandwidth due to the recording/production equipment.

Don't get me wrong, I get where this obsession with the cleanest sound possible without any wanted or unwanted artefacts comes from. I just wonder if it should really have that kind of significance.

Which, I guess, leads us back to the original topic.

[Ploki]

When I read all this I have to think that I've grown up in the wrong time... the music I grew up with not only had synths which aliased a lot, there was also noise from the recording, noise from the inferior physical recording media, and also had limited bandwidth due to the recording/production equipment.

Don't get me wrong, I get where this obsession with the cleanest sound possible without any wanted or unwanted artefacts comes from. I just wonder if it should really have that kind of significance.

Which, I guess, leads us back to the original topic.

imo no

i sometimes intentionally clip my guitar when recording DI, just slightly, few dB. It's an aliased clip. i don't care

[chk071]

I also often try to add that certain kind of smudge or mud to my synth sounds (mostly by adding noise, or a fast LFO for frequencly modulation), intentionally, to get that kind of "lo-fi" (if you will) artefact-y stuff.

There's a pretty well known German musician and mag writer called Moogulator, who called the Access Virus "schrebbelig" (or maybe he said that the Virus "schrebbelt", I don't remember exactly now). That's a pretty good term for the kind of aliased sound it's quite well known and liked for. IMO.

Again, not saying that that is something you will always, or even should aim for. It's just that a certain amount of "dirt" can be quite pleasant, or even essential.

[plexuss]

Yes.

I've been doing this and advocating it for years. If you can work at 96k and it sounds better at 96k then use 96k. Some plugins dont work properly at anything but 44.1k because the dev didnt code properly in which case there may be no way to work at a higher res. When everything works in 96k, I mix and master in 44.1 and render to 96k and it often sounds superior to lower rates.

In my experience the golden way is to work on your preferred or final sample rate and only use plugins that internally oversample. Without oversampling you won’t rid off aliasing and if you use analogue modelled plugins then this is crucial. Best of both worlds but you need to be selective with your plugins.

This is how I do it.

Question for you all,
If the label I work for asks for the final product to be in 44.1khz, what would you mix/master at? Curious to hear what you all would do.

Im my experience plugins can sound quite different at different rates, beyond what is expected just changing the sample rate. I assume this is due to poor code. I would work on the track in 44.1 as I always do and when ready render at both 44.1 and 96 and choose which sounds better, if either. If the 96 sounded better I'd convert to 44.1 for delivery, otherwise I'd deliver a 44.1 render. I just go with what sounds better...

[pixel85]

Some people like lots of inharmonics, while others don't. Some people prefer a clean sound, while others prefer lofi.

I wouldn't put artefacts like background noise on the same shelf as aliasing. Former is usually natural to the ear and often can be pleasant and beneficial (especially white noise). Plenty of random inharmonic content of aliasing can be accepted by some, and it can destroy the entire pleasure of listening to music.

In my day job, I work at 44.1kHz [because Cubase can't export video from projects in a higher sample rate so I have no choice :p ] and I'm fine with it, but after work, I record all my hardware at 96kHz. Nothing better than, after recording, slowing down by half some analogue synth tracks to get something new (previously unplanned) without loss of quality

[chk071]

I wouldn't put artefacts like background noise on the same shelf as aliasing. Former is usually natural to the ear and often can be pleasant and beneficial (especially white noise).

Most aliasing will probably happen with a lot of sound "content", not just a synth playing sine waves alone (for example), so, it will rather be perceived as additional "smudge", artefact-y whatever. You could say noise.