The Upsampling Your Mix Thread

[Marduchk]

is that a threat my dear man?

I think you are getting a little big for your bridges, mr. RAF

i try to help where i can, if someone gets in my way i treat them equally.... no reason to ban, even if the powers that be made the mistake of giving you that power.

no offense.

wait a sec. bitch all you want, kingston, im not going to derail this thread any further, go ahead, have your fun, prove you are a man, this is actually a nice thread, even with your contributions, thank you for those

[bduffy]

what I don't get there is if I mix this song to "perfection", then send out stems with no effects...it'll be a different mix at that point, won't it? I'd have to apply different EQ and compression and everything could change, no?

no. you take the channel/plugin/bus settings from the original 44.1khz mix. what would be the point if you exported those stems with the effects (at the lowly 44.1khz) already there?

the whole idea is simply to process *the exact same* settings as 96khz. (although I often find myself doing improvements at the 96khz stage).

for example, my drum bus gets exported with the effects off in the bus *but not* in the individual channels of course. common sense dictates these things.

Sorry for my lack of response, I was, er, mixing...

Right: so seeing as I send everything to Group Channels anyway, just export the Drum Group, say, without the effects, but with all the kick, snare, toms, etc channels' effects on, so you'll get that extra boost in quality. I'm sure, like you say, I'd be more than tempted to make some final adjustments, especially with all my effects oversampling! Got it, thanks.

Man, you just can't approach this subject without violent outbursts, eh?

[Kingston]

no reason to ban, even if the powers that be made the mistake of giving you that power.

see again I have absolutely no idea what you're talking about. I guess people like you have to invent things since life is obviously much too boring to you otherwise. on the other hand your comments bear the hallmarks of paranoia, which can be treated with right medication.

no offense.

yeah right. your posting is nothing but that.

[kylen]

actually it was dusted william who REALLY started it.... you can be sure kingston wouldnt have chimed in without already established information.

I was talking about another thread earlier in the week where I bumped into Kingston for a minute and discussed this...anyway thanks to bduffy for making a seperate thread and shouldering the responsibilities for handling the riff-raff in the peanut gallery!

[bduffy]

Hey guys,

When I have done these tests on my machines all of my source sounds were virtual. Virtual synths, drum machines, FX, EQ's, MASTERING.

So when I would bounce to a higher sampler rate in ExT and in LIVE it does treat it like the mix is in a higher sample rate and I don't have to worry about how audio is being processed. If I had audio then I think I would treat it how Kingston said, or just do my whole song from the get go in 96.

I like to do my whole song in 96 anyway and use FREEZE, but that's just me. When I use audio in my tracks that is probably how I would do it.

I then take the mix from the render. Upsample that to 192, then I master it seperately in another app using Ozone. Then I can resample to whatever rate I prefer. I do this so the mastering FX are at a nice and high sample rate even though it kills my machine to do so. I only do this on full finished songs because it is time consuming and can be a pain in the arse.

Ah, I think there's something crucial here that I missed way back last night; all your source sounds are virtual! Shit! I'm mixing guitars, bass, vocals, etc! Audio, audio, audio! Well, I guess that changes things a bit. And now I learn that eXt handles upsampling better than Cubase. So I'll have to look at mixing stems, or planning out where and when I can oversample in my projects. Bouncing out a higher sample rate than the Cubase project isn't going to do shit, but I will try your mastering method one more time on an a mix, see if I notice any difference.

Sorry for being a bit thick on this, guys; a little confusing, especially as I thought Cubase would handle oversampling like the competition. Whoops!

[camsr]

Can you set a higher sample rate in FL than the soundcard can support, only for exporting? Cause there's no option for sample rate at the export window, just bit depth. And I couldn't seem to change it in the options.

[Lagrange]

camsr not on export exclusively but you can switch under the Audio Prefs:


EDIT: arrg I see you can't switch in the options.. Are you using a sb16 card or something?? Don't think I've ever seen that b4.. You could try a workaround using something like ASIO4ALL..

Hey I'm jumpin in kinda late here.. This is really helpfull info!! I was convinced (b4 reading here) that upsampling just created a huge ass khz space above the existing audio that basically did nothing.. It totally makes sense now why some auto tuners sound better at the highest khz/bit rates.. Thanks for this guys!!

As for CPU claut @ 96 or 192. Would it make sense to do add the effects at the lower speed and when your ready to render/mixdown just increase the hosts clock then output?? Would save a huge headache I'd think..

L

[Kingston]

As for CPU claut @ 96 or 192. Would it make sense to do add the effects at the lower speed and when your ready to render/mixdown just increase the hosts clock then output?? Would save a huge headache I'd think..

L

You can, but AFAIK only FL studio and EXT can handle this with grace (maybe ableton live can as well?). It'll make subtle difference with the sound, but rendering time might jump to even half an hour or so. not a big deal if you've ever done 3d rendering work.

keep in mind some (badly coded) plugins don't scale well to higher sample rates, or sound quite different on the same settings.

[Shy]

So, did that Shy dude finally shut up?
Seems so. Maybe he's still searching for a sampler which sounds as good at 44.1 as at 96.
Hint: Shortcircuit and SFZ+ are really good candidates (albeit still not perfect, so even when using them there's a difference).

Did I shut up? Why are asking? Because as I recall from the other thread, you didn't want to hear a single word from me again because I'm <insert thousands of insults> and that's why you can't talk to such an insulting man as myself, even though I didn't even call you a single thing, but never mind, have fun.

Why would I need to search for a sampler? There are too many that aren't a piece of crap, meaning they actually have some form of antialiased resampling. You haven't provided an example with a proper resampler, only with the most crappy aliasing possible, as I suspected you'd do and commented long ago that you probably meant things that have ridiculously crappy internal aliasing to begin with.

[Marduchk]

wow, he got off "thousands of insults" at you?

I would like to see it, link plz

[Lagrange]

Personally I don't care if it took 3 hours to render as long as my mix sounds great after the fact..

So delving a little deeper into this.. What if I was to freeze/apply tracks at this higher rate with the effects (flanger, reverb, saturation/overdrive and auto tuning etc).. Could I essentially reimport/reuse this higher clock audio with the host back @ default khz defeat the point?

Oh yea and what about downsampling back to burnable speed? Do this in a editor? If some upsampling sucks then what what will I loose by downsampling?

This seems like a pretty deep subject.. Gone to look for more info..

L

[bduffy]

Personally I don't care if it took 3 hours to render as long as my mix sounds great after the fact..

So delving a little deeper into this.. What if I was to freeze/apply tracks at this higher rate with the effects (flanger, reverb, saturation/overdrive and auto tuning etc).. Could I essentially reimport/reuse this higher clock audio with the host back @ default khz defeat the point?

Oh yea and what about downsampling back to burnable speed? Do this in a editor? If some upsampling sucks then what what will I loose by downsampling?

This seems like a pretty deep subject.. Gone to look for more info..

Ah, Lagrange...you're getting sucked into the most passionate topic on KVR!

I think the golden rule here is to use the best SRC's possible; so bounce your tracks at the higher sample rate, convert them to 44.1 offline with Audiomove/r8brain, then bring those into your 44.1 project.

I think.

[Kingston]

Oh yea and what about downsampling back to burnable speed? Do this in a editor? If some upsampling sucks then what what will I loose by downsampling?

use this for all resampling tasks, http://www.lcscanada.com/audiomove/ and you don't have to worry about a thing. upsampling sucks? says who and why?

[Sascha Franck]

Why would I need to search for a sampler?

You asked me for an example of sound being degraded at lower samplerates. I posted one.
Now I expect you to post a counter-example of a softsampler that doesn't show that behaviour. That easy. And logical.
I am the one asking for examples now, after your bullshit claims have been proven to be just that: Utter bullshit.
An my question is: Show me a softsampler that doesn't sound better at 96k.

There are too many that aren't a piece of crap, meaning they actually have some form of antialiased resampling.

Name one. And then post an example.

You're not in a too great position to make false claims anymore. So far you have been proven wrong. Show us that you aren't. Tell us about which EQs, compressors, samplers, synths and whatever you use. Share your infinite wisdom.

[Shy]

Why would I need to search for a sampler?

You asked me for an example of sound being degraded at lower samplerates. I posted one.
Now I expect you to post a counter-example of a softsampler that doesn't show that behaviour. That easy. And logical.
I am the one asking for examples now, after your bullshit claims have been proven to be just that: Utter bullshit.
An my question is: Show me a softsampler that doesn't sound better at 96k.

There are too many that aren't a piece of crap, meaning they actually have some form of antialiased resampling.

Name one. And then post an example.

You're not in a too great position to make false claims anymore. So far you have been proven wrong. Show us that you aren't. Tell us about which EQs, compressors, samplers, synths and whatever you use. Share your infinite wisdom.

Sascha, you seem to like missing an important point I've made which is "as I suspected you'd do and commented long ago that you probably meant things that have ridiculously crappy internal aliasing to begin with".
The utter bullshit comes from you, since you did just that. Once you provide one sample made with proper tools (which even you yourself mentioned) you might be taken seriously. "Share your infinite wisdom".