Pro-L - do you use Oversampling? it seems not always better!

[Frankie.T]

i always master at 192k (i upsample any audio i recieve; and downsample again at the very end to whatever format is needed) so there is really no need for oversampling most of the time. unless im clipping, then i do use 2x.
in fact i often use 2x on pro-l aswell, but im not quite sure if im doing myself a service or disservice there, will have to a/b a bit more i think.

Do you realize upsample first, downsample later it's what oversampling does? You are just doing it manually

no. incorrect. as working in a high sample rate, means plugins doesnt have to internally oversample.
i control the upsampling at the very first point, then every bit of processing is in 192k, then i control the very last piece of downsampling as well as the ONLY AA filter. There is simply just one AA filter at the very end. Compared to each time there is oversampling, there is an additional AA filter.
theres a big big difference doing it like this, compared to working in a low sample rate and have each plugin do it itself. plus, there are plugins that doesnt have internal oversampling. either way, its better to do it this way, you just have to upsample once, and downsample once. thus reducing the instances of resampling by alot. less digital artifacts.
also, alot of plugins uses linear phase AA filters, thus each instance of oversampling will give you the nasty artifacts that CAN come using linear phase filters.
with that said I DO use linear phase LP filter at the end of my processing chain, but there is just ONE instance of it. as compared to oversampling a bunch of plugins, where there is many, thus causing more audible digitial artifacts.

What you say about one up/downsampling against multiple OS is correct, i can't argue with that.

I don't know if the final difference is really audible though. I mean technically must be, but i never heard any difference even with few OS plugin in series.

I think that's one of the moment where we could start an infinite math talking or just saying to do whatever sound the best for us, and i'm for the latter

[kmonkey]

and if your song Doesnt suck - would the extra bit of quality matter or not? the logic is a bit flawed saying "if its bad its bad". yes, thats true. but if its good, can it be slightly marginally better? cleaner? less digital noise/artifacts? ive spent countless hours doing a/b testing comparing mastering at 192k vs mastering at 44k, using identical plugin settings.
both masters end up at 44k/16bit. but on one, all processing was done in 192k. time and time again, does the 192k band limited version come on top. very slightly marginally at times, but always does. ive sent out both to clients and have them pick which one they think sounds better. ive asked friends to compare, and each time, the 192k version wins. why? because aliasing and the nyquist "roof" is not an issue anymore. even if its ever so marginal, using things like compressors, clippers, and limiters, in a 192k environment (regardless of any oversampling) just gives you a better quality end result, no matter if you downsample to 44/16 at the very end. this is a FACT. but i digress. this comment would be better suited in the 44k vs 48k thread, as im not really talking about oversampling here. cause oversampling in and of itself can be a bad thing.

It is exactly a mentality of a complicated person such as yourself. You are barking at the wrong tree or should I say you don't see a wood because of the tree. Or less talking more walking. Whatever comes your way.

I don't think anyone can argue with you that technicality is less important. It isn't. What I am trying to point out is that people should strive to reach a point where their artistry is well a real artistry.

Then they can worry about aliasing or oversampling.

OP is a perfect example. He can't hear or perceive a difference, but he worries about what the Plugin maker said in the product documentation. Without knowing that the plugin maker simply offered options for anyone to try and decide. But he asks for someone else to decide about his own decision to be made. On his own artistry.

You can make this technical all day long but it is as simple as: if it sounds good then it is good. For the creator of course.

To give you a better argument: literally, the music heroes who paved the way and defined electronic genres in the 80s or 90s, or even 2000s have been using digital hardware which works at 44khz, and for today's standard they used "outdated" hardware specs. Don't challenge me please to list A-class artists because there are numerous ones literally. Or don't say me that this music sounds outdated because that is subjective + in 2023 many amateurs are still trying to copycat the sound of the 90s.

None of them back then had access to 192khz upsampled plugins. Yes, some used analog mixers, or gear but some used digital ones, and so on and nothing stopped them to become famous. Which is the point I was making. The gold is in art, not in linear-phase oversampling.

[El°HYM]

and if your song Doesnt suck - would the extra bit of quality matter or not? the logic is a bit flawed saying "if its bad its bad". yes, thats true. but if its good, can it be slightly marginally better? cleaner? less digital noise/artifacts? ive spent countless hours doing a/b testing comparing mastering at 192k vs mastering at 44k, using identical plugin settings.
both masters end up at 44k/16bit. but on one, all processing was done in 192k. time and time again, does the 192k band limited version come on top. very slightly marginally at times, but always does. ive sent out both to clients and have them pick which one they think sounds better. ive asked friends to compare, and each time, the 192k version wins. why? because aliasing and the nyquist "roof" is not an issue anymore. even if its ever so marginal, using things like compressors, clippers, and limiters, in a 192k environment (regardless of any oversampling) just gives you a better quality end result, no matter if you downsample to 44/16 at the very end. this is a FACT. but i digress. this comment would be better suited in the 44k vs 48k thread, as im not really talking about oversampling here. cause oversampling in and of itself can be a bad thing.

It is exactly a mentality of a complicated person such as yourself. You are barking at the wrong tree or should I say you don't see a wood because of the tree. Or less talking more walking. Whatever comes your way.

I don't think anyone can argue with you that technicality is less important. It isn't. What I am trying to point out is that people should strive to reach a point where their artistry is well a real artistry.

Then they can worry about aliasing or oversampling.

OP is a perfect example. He can't hear or perceive a difference, but he worries about what the Plugin maker said in the product documentation. Without knowing that the plugin maker simply offered options for anyone to try and decide. But he asks for someone else to decide about his own decision to be made. On his own artistry.

You can make this technical all day long but it is as simple as: if it sounds good then it is good. For the creator of course.

To give you a better argument: literally, the music heroes who paved the way and defined electronic genres in the 80s or 90s, or even 2000s have been using digital hardware which works at 44khz, and for today's standard they used "outdated" hardware specs. Don't challenge me please to list A-class artists because there are numerous ones literally. Or don't say me that this music sounds outdated because that is subjective + in 2023 many amateurs are still trying to copycat the sound of the 90s.

None of them back then had access to 192khz upsampled plugins. Yes, some used analog mixers, or gear but some used digital ones, and so on and nothing stopped them to become famous. Which is the point I was making. The gold is in art, not in linear-phase oversampling.

I will challenge you:

- Post a List of your Top Ten 90s 'famous' Artists.
- Then post a List of the Labels they released on.
- Then post a List of their Mastering Studios & Engineers.

You are Mixing up two different processes here, the creative part (in which you are partly right) and the more technical Mastering Process (in which you are just utterly wrong or maybe unexperienced).

[perpetual3]

Within the context of electronic music at least, the better of an artist one is, the better the sound design, arrangement and mix are. What might be technical considerations in other genres are creative, artistic directions in electronic music. Understanding some of the issues faced in mastering may inform better artistic choices earlier in the creative process.

[Ploki]

I usually oversample @4x, but not always.

I also don't always oversample compressors on masters. The aliasing adds a gritty crips that sometimes works.
Oversampling tends to sound "smoother" and non-oversampled generally sounds more aggressive and unhinged.
With source that's already distorted and aggressive it can work well and makes it sounds really present and popping out of the speakers.

With a source that's more controlled and fluid, it might make it sound more fatiguing and not fit well.

With saturation plugins however, i almost always over sample - because they tend to create harmonics in higher levels than dynamic processors.

[Caine123]

I usually oversample @4x, but not always.

I also don't always oversample compressors on masters. The aliasing adds a gritty crips that sometimes works.
Oversampling tends to sound "smoother" and non-oversampled generally sounds more aggressive and unhinged.
With source that's already distorted and aggressive it can work well and makes it sounds really present and popping out of the speakers.

With a source that's more controlled and fluid, it might make it sound more fatiguing and not fit well.

With saturation plugins however, i almost always over sample - because they tend to create harmonics in higher levels than dynamic processors.

woah Ploki, this is a great reply, and totally understandable without having all the technical background !
i will test it further and check if i hear differences but i keep this in mind while testing!

thanks a lot!

[Ploki]

and if your song Doesnt suck - would the extra bit of quality matter or not? the logic is a bit flawed saying "if its bad its bad". yes, thats true. but if its good, can it be slightly marginally better? cleaner? less digital noise/artifacts? ive spent countless hours doing a/b testing comparing mastering at 192k vs mastering at 44k, using identical plugin settings.
both masters end up at 44k/16bit. but on one, all processing was done in 192k. time and time again, does the 192k band limited version come on top. very slightly marginally at times, but always does. ive sent out both to clients and have them pick which one they think sounds better. ive asked friends to compare, and each time, the 192k version wins. why? because aliasing and the nyquist "roof" is not an issue anymore. even if its ever so marginal, using things like compressors, clippers, and limiters, in a 192k environment (regardless of any oversampling) just gives you a better quality end result, no matter if you downsample to 44/16 at the very end. this is a FACT. but i digress. this comment would be better suited in the 44k vs 48k thread, as im not really talking about oversampling here. cause oversampling in and of itself can be a bad thing.

what you are doing however is oversampling, you're just doing it manually and globally.
Which isn't a bad method anyway, because you avoid intermittent AA filtering. (But can cause more aliasing than having all the steps anti-aliased separately)

Which DAWs had option to enable oversampling locally for the whole channel strip for example.

[mystran]

Purely from the technical point of view, oversampling is generally a no-brainer for every non-linear process with the potential to create aliasing, but peak limiting really is the exception to this rule, because the way things interact with downsampling potentially cancelling some of the limiting forcing slightly more limiting to compensate, in the technical sense for a given peak ceiling you end up with roughly the same amount of distortion either way and even though we might be able to show that the distortion noise with and without oversampling is different, the question is really whether one or the other is better masked in the subjective sense and the best answer one can give from a technical point of view is "it depends."

It's important to remember that even though there might be "aliasing" in the technical sense, most of the time with limiters this "aliasing" is rather more just gritty noise folding down, rather than obvious aliasing partials (unless you're using a limiter as a distortion device on an individual track with clean waveforms), so mostly it's a question of whether one type of gritty noise that's hopefully mostly masked sounds better than another type of gritty noise that's also hopefully mostly masked.

I wouldn't worry about pre-ringing too much, because that ringing is at 20kHz and you need a ridiculously steep filter before the ringing with 50 microsecond period extends so far back in time that it exceeds the premasking time of auditory perception; latency considerations generally discourage such filters anyway. To a certain extend effects like pre-ringing and negligible high frequency loss from oversampling filters can cascade through multiple rounds of oversampling, but for the most part the discussion is rather academic and whatever differences one might hear are more likely to be a result of secondary effects.

[ramseysounds]

Yeah but how does the kick drum sound?!

In all seriousness, it appears to me you’re creating problems so you can find solutions

And nobody yet has demonstrated that there is an AUDIBLE difference to a sound in a track or an overall track. Sometimes too much technology is such a workflow destroyer IMO

[kmonkey]

exactly

[rod_zero]

Is the difference within the perception of humans given psychoacoustics and the physical limits of human hearing?
People overestimate human hearing capabilities all the time.

[plexuss]

Is the difference within the perception of humans given psychoacoustics and the physical limits of human hearing?
People overestimate human hearing capabilities all the time.

Human hearing is one of the most inaccurate senses in terms of representing what is actually happening vs what is sensed. Sight is an even less accurate sense.

[ramseysounds]

@plexuss so is there an audible difference or not, just by listening for example to a track on ipods? Its a yes / no answer

[plexuss]

@plexuss so is there an audible difference or not, just by listening for example to a track on ipods? Its a yes / no answer

Actually, it's more of an opinion. I believe there is empirical data that generally shows that over-sampling sound better. However it depends on the plugin and it depends on the listener. Some listeners can't hear a change, some can, some thinks it sounds better, some thinks it doesn't. So it's not yes or no. Sorry. Such is the nature of audio and human hearing.

[ramseysounds]

So it’s a big fat no then