The Upsampling Your Mix Thread

[Lagrange]

Basically, we're talking about imparting that last 10% of clarity to a mix or master. What I'm doing, right this second, is bouncing down basic stems of my 48/32 song, with the bus channel's FX turned off, then I upsample those stems to 96khz, bring them back into a 96khz project, load the FX (that I've saved) back onto those channels, maybe do a little extra mixing there (kind of inevitable) and export a 96khz mix from there. Now you have a higher-quality mix that the mastering plugin will have to work at 96k too, in more detail, so you'll get a better quality 44k master at the end.

Phew! Hope that helps.

Yes it does!! You can breath now

Thanks!

L

[bduffy]

Actually, yours was less windy! Well done.

[Lagrange]

smaller lung capacity!

[bduffy]

smaller lung capacity!

Maybe you should upsample them?

[frequency_algorithm]

If you currently work at 88/24, then you're fine; you're already enjoying the benefits of your effects working at a higher sample rate, which is the whole point of this excercise: for the last stage of mixing and/or mastering, you get the mix's sample rate up, so the effects are working in higher detail, and you will benefit from a higher-quality sound.

I think of the anlaogy like this: Your mix is a story that you've written, and now it's going to be read aloud in front of an audience; you have the choice of it being read by an English teacher, who will do a good enough job, or you can get experienced vocal talent to do it, adding that extra flourish to the chapters!

Basically, we're talking about imparting that last 10% of clarity to a mix or master. What I'm doing, right this second, is bouncing down basic stems of my 48/32 song, with the bus channel's FX turned off, then I upsample those stems to 96khz, bring them back into a 96khz project, load the FX (that I've saved) back onto those channels, maybe do a little extra mixing there (kind of inevitable) and export a 96khz mix from there. Now you have a higher-quality mix that the mastering plugin will have to work at 96k too, in more detail, so you'll get a better quality 44k master at the end.

Phew! Hope that helps.

This would be cool, but why not just work in 96khz in the first place?

First of all, Cubase conversion is FINE. Obviously, none of you are really THAT worried about sound quality or you would be using better gear, like a IZ RADAR system or at least apogee stuff, tracking to tape, mixing on analog, etc.

The difference in sound quality will be basically neglegible. An "awesome" boost in sound quality from conversion algorithms are not always from a higher sample rate. You could upsample the crap out of something, but it will not change your headroom and it may not make it sound better.

Poor sound quality comes from harsh aliasing, phase smearing, distortion, poor summing and stuff like that. Besides, if you are all the way in 44.1/32 land, and you get your mix perfect...when you change over to 96k you may be pretty far from where you started, basically making a bunch of re-adjustments that may not make your mix sound that better. And when you dither down from that 96k or higher sampling rate, again the resulting sound may be a different picture than what you were looking at. Some engineers have learned to compensate for this difference, but I find it a hassle.

From real-world experience in mixing I truly find it difficult to be too far away from the sound I want while recording it or mixing it. This is true when using computer conversion, as opposed to working with pro quality A/D/A converters.

But yes, it helps fader resolution and fx resolution to have high sample rates, but only if the plug-ins have dither going in and coming out. And again, if you have low or user-quality plug-ins (specially reverbs) the resulting improvement will be insignificant, or may even be worse.

Without the proper dither applied to and effect going in and out, you might as well leave your mix the way it is.

If your music already sounds pretty good, just leave it cause all this computer conversion won't really affect your mix all that much.

There are, however, other ways you can improve a mix, like using excellent sounding pro quality conversion, mixing on analog consoles, etc.

Sorry that I didn't have time to get into the ballistics on upsampling, dither, etc., but this is a lot of technical typing that I am not getting paid for, I hope yall appreciate it.

[Lagrange]

Here we go again.. ARRRG How could I forget..
A header is a header is a header..

L

[bduffy]

If you currently work at 88/24, then you're fine; you're already enjoying the benefits of your effects working at a higher sample rate, which is the whole point of this excercise: for the last stage of mixing and/or mastering, you get the mix's sample rate up, so the effects are working in higher detail, and you will benefit from a higher-quality sound.

I think of the anlaogy like this: Your mix is a story that you've written, and now it's going to be read aloud in front of an audience; you have the choice of it being read by an English teacher, who will do a good enough job, or you can get experienced vocal talent to do it, adding that extra flourish to the chapters!

Basically, we're talking about imparting that last 10% of clarity to a mix or master. What I'm doing, right this second, is bouncing down basic stems of my 48/32 song, with the bus channel's FX turned off, then I upsample those stems to 96khz, bring them back into a 96khz project, load the FX (that I've saved) back onto those channels, maybe do a little extra mixing there (kind of inevitable) and export a 96khz mix from there. Now you have a higher-quality mix that the mastering plugin will have to work at 96k too, in more detail, so you'll get a better quality 44k master at the end.

Phew! Hope that helps.

This would be cool, but why not just work in 96khz in the first place?

First of all, Cubase conversion is FINE. Obviously, none of you are really THAT worried about sound quality or you would be using better gear, like a IZ RADAR system or at least apogee stuff, tracking to tape, mixing on analog, etc.

The difference in sound quality will be basically neglegible. An "awesome" boost in sound quality from conversion algorithms are not always from a higher sample rate. You could upsample the crap out of something, but it will not change your headroom and it may not make it sound better.

Poor sound quality comes from harsh aliasing, phase smearing, distortion, poor summing and stuff like that. Besides, if you are all the way in 44.1/32 land, and you get your mix perfect...when you change over to 96k you may be pretty far from where you started, basically making a bunch of re-adjustments that may not make your mix sound that better. And when you dither down from that 96k or higher sampling rate, again the resulting sound may be a different picture than what you were looking at. Some engineers have learned to compensate for this difference, but I find it a hassle.

From real-world experience in mixing I truly find it difficult to be too far away from the sound I want while recording it or mixing it. This is true when using computer conversion, as opposed to working with pro quality A/D/A converters.

But yes, it helps fader resolution and fx resolution to have high sample rates, but only if the plug-ins have dither going in and coming out. And again, if you have low or user-quality plug-ins (specially reverbs) the resulting improvement will be insignificant, or may even be worse.

Without the proper dither applied to and effect going in and out, you might as well leave your mix the way it is.
If your music already sounds pretty good, just leave it cause all this computer conversion won't really affect your mix all that much.
There are, however, other ways you can improve a mix, like using excellent sounding pro quality conversion, mixing on analog consoles, etc.

Sorry that I didn't have time to get into the ballistics on upsampling, dither, etc., but this is a lot of technical typing that I am not getting paid for, I hope yall appreciate it.

I don't work at 96khz because my computer can't handle it, and I'm only just exploring using these higher sample rates now. It can only handle a few stems with a few FX, and even then, I have to bounce those pretty early on.

Second, I never said upsampling will make your audio sound better; if anything, it should sound exactly the same. If you got that from what I wrote, then I guess I didn't explain it clearly enough. It's the plugins' processing we're talking about. This has been discussed already, if you go back through the thread.

[mauseoleum]

It's plugs + summing. But it's also softsynths + samplers. When the latter are in the spiel, the diff is like wowowowow.

And, the difference can be distinct even after going back to cd rate, even on mp3.

Btw. I played a bit more with eXT, and discovered a "feature" so to speak (1.4) if trying to "oversample" the render (44k project to 88k render):

- if one renders from sequencer and has arpeggiators hooked to the midi channel, one has to manually compensate for diff in samplerate at export (i.e. adjust speed of the arp from say 16 to 8 !!! ). Which brings back the point of validity of X2 factor, which I stated in the other thread (re. tracktion, starting here http://www.kvraudio.com/forum/viewtopic ... c&start=45 and next page):

- in order to "oversample the host" and to all inner workings to align properly, I (still) assume that whole-number proportion should be preserved (but isn't neccessary when applying offline conversion).

In other words, at least with eXT1.4, you'd get a funny syncopated mess if you export 44k at say 96k and have arps hooked to midichannels.

I think somebody should transfer this to the developer. Such behaviour should be prevented/corrected in XT2 (imho).

[kylen]

... Obviously, none of you are really THAT worried about sound quality or you would be using better gear, like a IZ RADAR system or at least apogee stuff, tracking to tape, mixing on analog, etc.

...Sorry that I didn't have time to get into the ballistics on upsampling, dither, etc., but this is a lot of technical typing that I am not getting paid for, I hope yall appreciate it.

We're worried bout quality - that's why we're taking the trouble to get as much as we can with what we have at the moment. It takes a few $$$ to get the necessary analog stuff, I have some good AD/DA already (Benchmark, Mytek) maybe an analog mastering compressor down the road. But you misunderstood what we're doing anyway...

I wouldn't worry about not getting paid for your comments!

[frequency_algorithm]

my bad. But what you're doing is still useless. just go back to mixing and making music.

[kylen]

my bad. But what you're doing is still useless. just go back to mixing and making music.

It's OK - we're havin a good time though. The digital plugs sound and work better at the higher sample rate...

[frequency_algorithm]

my bad. But what you're doing is still useless. just go back to mixing and making music.

It's OK - we're havin a good time though. The digital plugs sound and work better at the higher sample rate...

duh....any engineer or plug in manufacuter will tell you that.

but you still need dither.

[mauseoleum]

No prob. You just don't get it. This one for you. http://www.weiss.ch/p2d/p2d.html You like hardware better, scroll down.

[kylen]

my bad. But what you're doing is still useless. just go back to mixing and making music.

It's OK - we're havin a good time though. The digital plugs sound and work better at the higher sample rate...

duh....any engineer or plug in manufacuter will tell you that.

Well OK - that's what we're doing then, what's the prob?

[frequency_algorithm]

No prob. You just don't get it. This one for you. http://www.weiss.ch/p2d/p2d.html You like hardware better, scroll down.

what about weiss? what's your point. you still need dither.