bouncng down ruins mixes.
- Thread starter DAS19
- Start date
Steve Henningsgard
New member
That I'm honestly not sure about, but I do see how that could be true. Taking all of those sounds/tracks/effects/whatever, and turning them into a single (or two, I guess, for stereo) sound wave, is likely one of the more critical steps there are.
Anyone have some good info on this?
Anyone have some good info on this?
deejaytrixx
New member
allegedly, you lose very minute amounts of sound quality as well..esp. when doing bit conversions, hence why pro tools recommends dithering even when bouncing files that are already 16bit
Ethan Winer
Acoustics Expert
What school do you go to? Seriously, I'd like to know.
Anyway, that is preposterous and simple to debunk. When you press Play in a DAW program all of the tracks are internally mixed with your plug-ins and send buses etc, then sent to your sound card for playback. When you render a mix to disk it's the exact same data, but written to a Wave file instead. If a rendered mix does not sound exactly the same as when played live, then something is wrong with the DAW software, or something is configured wrong.
Please send a link to this post to whomever at your school told you that, and I'll be glad to explain further if needed.
--Ethan
most of these people are talking about null tests which ive done time and time again. I don't think the math is as simple as 1+1 is 2 and I'm not alone
I think there are reasons they sound diffrent but I asume it's cause there is diffrent math going on from playing plugins back then it is in rendering through them. I think your buffer settings effect the mix as well that to me is why a bounced mix doesn't sound the same as one recorded to a stand alone burner hard disk another daw etc. I think it has to do with the time it takes to render through the plugins with bounce to disk type rendering causing some latency issues with the plugins however minute they might be, but pro tools bounce to disk has always narrowed the stereo spread of any mix ive done from pro tools mix 4 and 5 to ptle 5.5 always a shrinking in the stereo spread
mattr
Resident moody teenager
Are you sure that it isn't due latency issues with plugins when mixing that are introducing delays which widen the stereo image? Then when you bounce to disk, the plugin latency isn't present and so the delays do not manifest, therefore the wideness you thought you had disappears?
Its just a thought, but it might be a playback problem during mixing rather than a problem with bouncing that is making you think your mix is wider than it actually is. Try mixing with higher buffer settings, enable automatic-delay-compensation on plugins, etc, etc
Ethan Winer
Acoustics Expert
That's why I said, "If a rendered mix does not sound exactly the same as when played live, then something is wrong with the DAW software, or something is configured wrong." If a DAW doesn't process playback and a Wave file bounce the same, then it is for all intents and purposes broken. Indeed, how are you supposed to make a mix if you can't tell what it will sound like when rendered?
However, there's another reason people sometimes think they hear a difference when there really is no difference:
A common-sense explanation of audiophile beliefs
--Ethan
NL5
Unpossible!
LOL. No matter how clear cut this one is, some people won't buy it. It's really kinda funny.
Just because two files null 100%, doesn't mean they are the same! ROFLMAO.....
You got the first and last part of that statement correct.
All kidding aside, do you know what floating point is? I hate to break it to you, but 1+1=2 is MORE complicated. There is no "2" in binary LOL. Now, I know PT is a big ol pile of crap as far as audio programs are concerned - and "LE" is pretty much worthless for mixing anything on, but if it's not summing correctly, it is seriously flawed. However, I believe some of the null tests were done with LE and it summed fine.
To the OP. Switch to Reaper. It kicks the crap out of LE in all the areas that actually are different between apps.
gecko zzed
Grumpy Mod
what does null testing have to do with what you hear played back and rendered?
the issue I've had is when listening to my mix it's wider and has better bass, but on bouce to disk I lose some bass definition and the stereo spread shrinks that has nothing to do with null testing
NL5
Unpossible!
Sorry, I got a bit sidetracked. But again, summing is not complicated on a digital system. Even if it was, why would a program sum one way on a mix, then switch to a different summing algo to bounce down. Makes zero sense. I also think taht one of the tests actually nulled a playback digital capture and a bounced mix as well. I'll see if I can find it. Shouldn't be too hard to do yourself. If I had PT I'd do it, but Cubase doesn't have programming flaws like that - it has ton's of others though.
Now, I know some plugins used to use a "preview" type system during playback, but I am pretty sure that stopped awhile back. I can't bounce tracks, so it's never been an issue for me anyway.....
So let me see if I understand what you are trying to state correctly.. You have a mix up in PT ( I assume this is the daw you are using if not please identify it) and you listen to it. I also assume that this is a mix done "in the box" NOT through any outboard gear ie. outboard compressors,reverbs, or console for that matter. It sounds like you want it to, then you bounce it down and it suddenly sounds less wide( regarding stereo imaging) than the mix did prior to being bounced down? Have you tried printing the mix to another track instead of bouncing it down? Does this sound the same as the mix before bounce? Does this printed track null with bounced track? Are you monitoring through a console as you mix?
Randy
Randy
thats what im saying i listen i mix on ptle it sounds good i bounce to disk i lose my stereo spread if i send it to a stand alone 2 track its wider im using the internal monitoring but it does it with my digital mixer as well same with the bottom end i lose bottom end when bouce to disk as well have since pro tools mix 5 and le 5.5
Have you tried printing vs. bouncing and see if they null? I have personally tested this with PT HD 6 and 7 and PT m-powered 6.8, 7.0 7.1 and 7.3. They have all nulled. You will certainly get a slightly different stereo image going through a console usually due to panning law or possibly the console coloring the sound. (if its a decent console it will probably color the sound in a pleasant way) I certainly am not knocking you for your methods so please don't take it that way.... but to claim that somehow what you hear vs. the bounce are somehow different even though they will completely cancel each other out when 180 degrees out of phase fly's in the face of science. I suggest you take this concern to the digidesign user group also known as the DUC.
Randy
Randy
Ethan Winer
Acoustics Expert
I'm pretty sure the real issue is acoustic in nature, as explained in my Believe article. Do you have any acoustic treatment in your room, especially absorption at the side-wall and ceiling reflection points and on the real wall behind you? And what happens if you listen with headphones?
Stereo wideness has nothing to do with summing, and everything to do with content and phase differences between the channels. Here's a simple test you can try: Take a single track panned hard left and bounce it to disk. Then do the same with a single track panned hard right. Do either of the rendered files have leakage etc on the wrong channel?
--Ethan
KineticSound
The VOICE
Did anyone ever come back to this? It seems to have sort of dropped out of the discussion loop, but I'm interested if there's any fact to the assumption that there may be a difference between playing back vs. rendering. 99% of the projects I do end up rendered to .mp3/.wav/.aif and I've never experienced any notable differences from what I heard in playback - but that's not to say the differences were there and just so minor I didn't hear them.
if it wasn't for the fact that it happens in headphones I'd belive that 100% my room sucks it's bare walls no treatment and a small rectangle, but here is the kicker it has happened everywhere i've recorded. With pro tools mix 5 it did it at a college studio through my gear there, but that was a fostex tape transfer from 16 track 1 inch tape to ptle the bass just got lost in the transfer and everyone noticed it not just me. I've even had other engineers tell me the same thing I was in a tdm setup control24 mix system 24bit 888s and we were mixing and I was like the bass seems a lil bit to loud in the mix comming off of genelics not alot, but just a bit the engineer said yea that's cause we will lose the bass slightly when we bounce the mix so I'm not the only one that notices it. I cant explain why it happens but I know I hear it when it does
now cause you all think im nuts im going to find other nuts that agree with me
Sorry, can't agree. I NEVER, after years of trying was happy with a single PTLE bounce. That is even with Waves L1 Maximizer and every dithering option available. It never bounced and sounded like what I was mixing.
With Reaper, that problem was gone on day one. They have different mix engines and I have to assume that they use different equations to get the math done.
There is more than one way to get to 10.
thats one
here is another vs export summing and bouce to BTD
BTW, I've done extensive testing. I can not accurately tell any difference between BTD and doing it this (your) way. back in the old days on a Mix sytem, but on HD I can't tell the difference. It is a lot more convenient for me though - especially when doing stems.
Once upon a time, the Digi mix buss was not all it was cracked up to be. I think that's when the method of recording internally started and when the superstition and distrust grew. HD is well past that and Digi took their lumps, learned, re-wrote code and have (argueably) the best DAW in the world.
Bounce to disk and Internal Layback DEFINITELY sound different and its not just a placebo effect. Here's what happens [this knowledge is from several hands on shootouts, discussions with Digi programmers and as a Digi Certified Op in Music and Post, this was definitely pointed out].
When you select bounce to disk, PT's mix engine architecture is actually taking your mix out of the DAE and into a SEPARATE mix engine. On the way there, it first truncates your data, then begins tossing out bits of information. If you're low on voices, the process is even more detrimental as it needs more power to do the operation and throws out more. The result is something lacking in high end definition, a log jam of a mid range and a cluttered low end.
With internal layback, you're staying in the DAE and avoiding all of that. Essentially, what you mix is what you print. Thats your track. It definitely sounds A LOT better and if you can't hear it, I'm sorry. But its definitely there.
Even beyond the sonic superiority of it, its great for workflow. With QuickPunch engaged, you can easily punch in a part of the print track that all of a sudden clipped because you forgot to fade a region. You don't have to cancel your bounce at the last minute, fix the problem and then wait it out again. You also have the advantage of being able to export different sample rates [a mix for RedBook, DVD-A, etc] as well as different file formats, ALL WITHOUT HAVING TO WATCH IT BOUNCE 6 TIMES! Imagine that!
Ultimately, if you don't hear a difference, you dont hear one. If what you hear from the bounce engine sounds great to you, then thats all that matters.
i could go on and on and on i heard the diffrence then and im still using an old version of ptle maybe they got better as they went and it sounds better now but i know im not crazy when i hear my mix get torn apart by bounce to diskanother posibility
don't know anything about Nuendo,,,, but in pro tools you have to go into a preference setting and change the bounce conversion from 'better' to 'tweakhead' I have no Idea why it doesn't come factory sounding the best, and let you downgrade it if you don't like the conversion time but the sound difference is huge....
I've noticed differences in final mixes on Cubase... I think it has to do with buffer settings, in my instance. I set them pretty high while mixing, for the processor's sake. Then, listening to the final mix, it's definitely different. Happens every time.
---------------------------------------------------------------------------------------------------------------------------------------------------
Don't confuse internal bouncing (or Audio Export, or whatever your DAW calls it) with resampling. When you do a bounce - STAY WITH YOUR PROJECT BIT DEPTH SAMPLE RATE.
Just because your DAW can do both together doesn't mean you should. You shouldn't.
Whenever you resample - say from 24/48 to 16/44 - you are employing an algorithmn which destructively mangles your audio, and you are also forcing your hardware D/A to employ a different filter setting. You will definately hear a difference when you resample - so don't confuse this with summing or internal bouncing - it's a seperate issue.
There are studies and graphs that show that the resampling algo's in most popular DAWs are not very good and introduce artifacts. You will get much better results if you use a standalone resampling app, such as Voxengo R8Brain. Or resample via analog with seperately clocked hardware converters.
Also - be aware that some plugins render differently for realtime and offline. Some will have 'high quality' options for offline that are not used in realtime. And unfortunatley - sometimes a 'high quality' option sounds worse than the lower quality option (I kid you not). Test everything.
Null tests are not always possible - because some plugins have random features that mean that no two bounces will ever null anyway, even A/B'ing with the exact same method. To do proper null tests you need to be working with audio tracks only, with all instruments and effects pre-rendered.
Sorry, can't agree. I NEVER, after years of trying was happy with a single PTLE bounce. That is even with Waves L1 Maximizer and every dithering option available. It never bounced and sounded like what I was mixing.
With Reaper, that problem was gone on day one. They have different mix engines and I have to assume that they use different equations to get the math done.
There is more than one way to get to 10.
thats one
here is another vs export summing and bouce to BTD
BTW, I've done extensive testing. I can not accurately tell any difference between BTD and doing it this (your) way. back in the old days on a Mix sytem, but on HD I can't tell the difference. It is a lot more convenient for me though - especially when doing stems.
Once upon a time, the Digi mix buss was not all it was cracked up to be. I think that's when the method of recording internally started and when the superstition and distrust grew. HD is well past that and Digi took their lumps, learned, re-wrote code and have (argueably) the best DAW in the world.
Bounce to disk and Internal Layback DEFINITELY sound different and its not just a placebo effect. Here's what happens [this knowledge is from several hands on shootouts, discussions with Digi programmers and as a Digi Certified Op in Music and Post, this was definitely pointed out].
When you select bounce to disk, PT's mix engine architecture is actually taking your mix out of the DAE and into a SEPARATE mix engine. On the way there, it first truncates your data, then begins tossing out bits of information. If you're low on voices, the process is even more detrimental as it needs more power to do the operation and throws out more. The result is something lacking in high end definition, a log jam of a mid range and a cluttered low end.
With internal layback, you're staying in the DAE and avoiding all of that. Essentially, what you mix is what you print. Thats your track. It definitely sounds A LOT better and if you can't hear it, I'm sorry. But its definitely there.
Even beyond the sonic superiority of it, its great for workflow. With QuickPunch engaged, you can easily punch in a part of the print track that all of a sudden clipped because you forgot to fade a region. You don't have to cancel your bounce at the last minute, fix the problem and then wait it out again. You also have the advantage of being able to export different sample rates [a mix for RedBook, DVD-A, etc] as well as different file formats, ALL WITHOUT HAVING TO WATCH IT BOUNCE 6 TIMES! Imagine that!
Ultimately, if you don't hear a difference, you dont hear one. If what you hear from the bounce engine sounds great to you, then thats all that matters.
i could go on and on and on i heard the diffrence then and im still using an old version of ptle maybe they got better as they went and it sounds better now but i know im not crazy when i hear my mix get torn apart by bounce to diskanother posibility
don't know anything about Nuendo,,,, but in pro tools you have to go into a preference setting and change the bounce conversion from 'better' to 'tweakhead' I have no Idea why it doesn't come factory sounding the best, and let you downgrade it if you don't like the conversion time but the sound difference is huge....
I've noticed differences in final mixes on Cubase... I think it has to do with buffer settings, in my instance. I set them pretty high while mixing, for the processor's sake. Then, listening to the final mix, it's definitely different. Happens every time.
---------------------------------------------------------------------------------------------------------------------------------------------------
Don't confuse internal bouncing (or Audio Export, or whatever your DAW calls it) with resampling. When you do a bounce - STAY WITH YOUR PROJECT BIT DEPTH SAMPLE RATE.
Just because your DAW can do both together doesn't mean you should. You shouldn't.
Whenever you resample - say from 24/48 to 16/44 - you are employing an algorithmn which destructively mangles your audio, and you are also forcing your hardware D/A to employ a different filter setting. You will definately hear a difference when you resample - so don't confuse this with summing or internal bouncing - it's a seperate issue.
There are studies and graphs that show that the resampling algo's in most popular DAWs are not very good and introduce artifacts. You will get much better results if you use a standalone resampling app, such as Voxengo R8Brain. Or resample via analog with seperately clocked hardware converters.
Also - be aware that some plugins render differently for realtime and offline. Some will have 'high quality' options for offline that are not used in realtime. And unfortunatley - sometimes a 'high quality' option sounds worse than the lower quality option (I kid you not). Test everything.
Null tests are not always possible - because some plugins have random features that mean that no two bounces will ever null anyway, even A/B'ing with the exact same method. To do proper null tests you need to be working with audio tracks only, with all instruments and effects pre-rendered.
Last edited: