bounce to disk VS. print new stereo track

[Bob Boyd]

This shouldn't be a problem anymore (emphasis on "shouldn't") but I confess it has been a while since I've used "Bounce to Disk" on any album project. What version of PT are you running? I remember things like this being more of an issue back in the PT MIX system days.

[bob katz]

Originally Posted by Bob Boyd

This shouldn't be a problem anymore (emphasis on "shouldn't") but I confess it has been a while since I've used "Bounce to Disk" on any album project. What version of PT are you running? I remember things like this being more of an issue back in the PT MIX system days.

There is absolutely no difference! In fact, if you add "dub to an external digital recorder via AES/EBU" as a third alternative, there is absolutely no difference among the three choices. This is the biggest urban legend yet. IF YOU HEAR A BONIFIED DIFFERENCE while cutting these files in two different ways, then take the resulting files, they completely null, they are bitwise identical.

Done numerous tests on this precept with alternative files sent in to us by clients. Blind and non-blind.

BK

[gainstages]

That is interesting to hear Bob. I highly respect your opinion, but as you know, this is something that is talked about all the time on this forum and other places - with bounce to disk being something evil that you should never do and a lot of people instead doing the new stereo track.

i've listened before and couldn't tell the difference and just figured my ears must be going bad!

[masteringhouse]

Originally Posted by bangerandmash

if it matters, i am able to hear the difference between the bounce to disk version of a song and the same song being played directly out of pro tools.

thanks in advance for any help.

As Bob mentioned, try taking both of the above files, line them up, and reverse polarity or invert one of the tracks (assuming they are stereo tracks). Send them out to the same interface and let us know if anything shows up on the meter or if you hear anything.

[Bob Boyd]

Originally Posted by bob katz

There is absolutely no difference! In fact, if you add "dub to an external digital recorder via AES/EBU" as a third alternative, there is absolutely no difference among the three choices. This is the biggest urban legend yet. IF YOU HEAR A BONIFIED DIFFERENCE while cutting these files in two different ways, then take the resulting files, they completely null, they are bitwise identical.

Done numerous tests on this precept with alternative files sent in to us by clients. Blind and non-blind.

BK

Yeah, that's where I was leading. I don't know if I did a null test back in my MIX system days but on the current hardware, they will null.

[finetuner]

Unless your project is 24 bit, no dither plugin on the master and bounce output is set at 16 bit.
Then playback will remain at 24 but bounce will truncate.
That's all i can think of.
Just checkin' ? ...

Peter

[lobsterinn]

I have to admit it has been a few years since I did a scientific test of this, but it definitely used to be noticeable- I will try the nulling test asap.

Do the above comments apply to LE systems as well?

[bob katz]

Originally Posted by lobsterinn

I have to admit it has been a few years since I did a scientific test of this, but it definitely used to be noticeable- I will try the nulling test asap.

Do the above comments apply to LE systems as well?

Yes.

And if you hear a difference but the null test shows nothing out, then make sure your listening comparison was done as fairly and accurately as possible, same DAC, A/B comparison, etc.

BK

[bangerandmash]

bob and everyone else thank you for your great input about this matter. i'm still unclear about how to employ the alternative to bounce to disk method. anyone care to run me through it?

also, what i was saying i can hear the differences between were the bounce to disk file and that same song being played straighout of pro tools. bob and others, are you saying that i should not be hearing any difference between these two?

[Jerry Tubb]

Good to hear your findings regarding PT are positive Bob, been my experience as well.

Digidesign Pro Tools seemed to have ironed out all there "math issues" a few years ago.

I think it was with the release of Pro Tools HD, and then the 002 for PT LE.

Probably happened when they rewrote the app to include the higher sampling rates.

A few years ago in a conversation with Bob Ludwig, he stated that PT had finally gotten their math "right".

We use both PT HD & LE everyday, without any hesitation, or significant "issues".

JT

[dave-G]

THE remaining issue with Bounce to Disk is probably only found on slower CPUs with huge, sprawling mixes using all system resources, and insane amounts of automation

... you know ... the average mix.



In those cases, with a PTHD rig, I have observed some occasions where automation was played back with subtle inaccuracies during BTD instead of straight playback/recording to tracks. I use BTD fairly regularly, only worrying about it on those occasions where the project is making the system sluggish. I have a 2nd DAW to record to for those situations, or just make some tracks and record it back.

Maybe with faster intel macs this too is a thing of the past, as it always seemed to me to be a CPU-coping problem. Otherwise/audiowise, BTD is fine.

(real confidence-inspiring, huh?)
-dave

[lobsterinn]

[QUOTE=bangerandmash;1498327]bob and everyone else thank you for your great input about this matter. i'm still unclear about how to employ the alternative to bounce to disk method. anyone care to run me through it?

Make a new stereo audio track and set the outputs to all your other tracks to its' input on via a stereo bus. If you have any program processing or master fader moves you can bus everything through a stereo aux track which then gets bussed into your stereo audio track.

It is a pain, but it (used to?) be worth it.

[dave-G]

Originally Posted by bangerandmash

i'm still unclear about how to employ the alternative to bounce to disk method. anyone care to run me through it?

You could do what lobsterinn says, or you could also do it via AES I/O (well, O/I). If you're running more moderate sessions, btd should work just fine.

-dave

[Bob Olhsson]

I still buss my mixes into a stereo track.

The reason I do this is because I can instantly compare different mixes with the current one by simply switching different playlists between "auto input monitor" and "input only." I can also punch in on a mix or "comp" between mixes which is much faster that fooling around with automation.

[masteringhouse]

Originally Posted by Bob Olhsson

I still buss my mixes into a stereo track.

The reason I do this is because I can instantly compare different mixes with the current one by simply switching different playlists between "auto input monitor" and "input only." I can also punch in on a mix or "comp" between mixes which is much faster that fooling around with automation.

Interesting perspective Bob. Just curious on how you create a stereo interleaved file with this method, or isn't it a concern? Do you export after this?

[MattGray]

Originally Posted by masteringhouse

Interesting perspective Bob. Just curious on how you create a stereo interleaved file with this method, or isn't it a concern? Do you export after this?

I usually do the same thing as Bob O, not because it sounds better but because of the convenience factor. Once I've comp'd my master I export to 24bit - 96k (all my sessions are 24-96) using <cmd+shift+K>.

Matt

[masteringhouse]

Originally Posted by MattGray

I usually do the same thing as Bob O, not because it sounds better but because of the convenience factor. Once I've comp'd my master I export to 24bit - 96k (all my sessions are 24-96) using <cmd+shift+K>.

Matt

Thanks Matt.

So you perform all of your editing and fades in PT, export, SRC/dither to 16, then use this "image" for the final CD?

[MattGray]

Originally Posted by masteringhouse

Thanks Matt.

So you perform all of your editing and fades in PT, export, SRC/dither to 16, then use this "image" for the final CD?

Editing yes but fades & sequencing are done in WaveBurner at 24bit 44.1kHz... MBIT+ dither using Ozone3 as a master insert in WaveBurner.

[bob katz]

Originally Posted by bangerandmash

bob and everyone else thank you for your great input about this matter. i'm still unclear about how to employ the alternative to bounce to disk method. anyone care to run me through it?

also, what i was saying i can hear the differences between were the bounce to disk file and that same song being played straighout of pro tools. bob and others, are you saying that i should not be hearing any difference between these two?

Excepting what Matt Gray brought up about slow automation on slower systems (which I knew about but didn't bother to mention), you are correct, you should not hear any difference among these methods. Below is an outline of how to accomplish each in Pro Tools:

1) New Track. The input for the track should be the output of your master fader. Just make sure the levels do not exceed full scale. Pro Tools meters are notoriously inaccurate. If you see red, I suggest lowering the master fader. Just go into record and play back the rest of the tracks.

2) Bounce to Disk. Self-explanatory.

3) Feed to external. An external digital recorder, or ANOTHER DAW! takes the digital output of your mix that your route to AES/EBU or SPDIF. Record it 24 bit and it should likewise, create a file which is identical to what you hear when you monitor the mix.

[masteringhouse]

There have been some claims that adding a master fader track also changes the sound. I've done some blind tests both myself and with some more critical clients and there seems to be an extremely subtle difference in imaging even with it set to 0 (hope this isn't going into urban legend land).

For mastering there shouldn't be a major reason for using a master fader since it's pretty easy to adjust a stereo track or bus, so to be on the safe side I personally don't use one during the final bounce (as Bob said elsewhere "chicken soup").

Rather than using the master fader track, you may want to also try using a bus as input to a new track, bouncing down from the same aux bus, and external feed as Bob described.

[finetuner]

Originally Posted by masteringhouse

There have been some claims that adding a master fader track also changes the sound.......For mastering there shouldn't be a major reason for using a master fader since it's pretty easy to adjust a stereo track or bus, so to be on the safe side I personally don't use one during the final bounce .....

Exactly, but sometimes i have to add a master track for summing/decoding M/S components back to stereo.
So now you made me curious, and i think a null test should point out if adding a master fader will make any differece.
I'll do a test tomorrow.

Peter

[bob katz]

Originally Posted by finetuner

Exactly, but sometimes i have to add a master track for summing/decoding M/S components back to stereo.
So now you made me curious, and i think a null test should point out if adding a master fader will make any differece.
I'll do a test tomorrow.

Peter

And this result will be different depending onwhether it's HD or LE. Additional calculations can change the sound, but hopefully in a 24 bit environment, it's "below the radar". I've not done the shootouts, and frankly, most times these days I'm working in a floating point environment, and I have NOT heard any degradation when adding a master fader in a floating point DAW.

[Bob Olhsson]

Back in the '90s I did a shootout between a Pro Tools System and a Sonic Solutions system

The Pro Tools (version 4, if I remember correctly) master fader sounded shockingly worse than the Sonic's gain control. Then we stuck a Waves Ren EQ in the ProTools master channel, left the fader at zero and dropped the level using the plug-in. This sounded considerably better than the Sonic!

Unfortunately past shortcomings of Pro Tools have this way of being endlessly repeated many years after the problem has been fixed.

[MattGray]

Originally Posted by masteringhouse

There have been some claims that adding a master fader track also changes the sound. I've done some blind tests both myself and with some more critical clients and there seems to be an extremely subtle difference in imaging even with it set to 0 (hope this isn't going into urban legend land).

For mastering there shouldn't be a major reason for using a master fader since it's pretty easy to adjust a stereo track or bus, so to be on the safe side I personally don't use one during the final bounce (as Bob said elsewhere "chicken soup").

Rather than using the master fader track, you may want to also try using a bus as input to a new track, bouncing down from the same aux bus, and external feed as Bob described.

Tom, it's important to know that regardless of whether you create a master fader or not, there is always a master fader. What I mean is that if you don't create a master fader it's there only hidden from the user & it's operating at unity gain. I read Digi's white paper regarding this. This is why creating a master fader in PT's HD doesn't use any additional DSP resources as it's already allocated.

So creating a master fader is just making it visible to the operator & will not change the sound or add any additional DSP unless you change the gain of that fader or add inserts.

Matt

[masteringhouse]

According to the following document:

"If a master fader is not used in the session, Pro Tools behaves as if there is a master fader behind the scenes set to unity gain. With the dithered mixer plug-in, high-quality dithering is used to deliver the final 24-bit output signal."

http://akmedia.digidesign.com/suppor..._Box_26689.pdf

Possibly there is another dither step (in the dithered mixer) that can be avoided by not using the fader, and may help to explain a potential difference in sound.

[edit] Matt - the above was posted before I saw your reply, but may explain the issue[/edit]

I think that if null testing is being done it may also require using both the dithered and non-dithered versions.

[UnderTow]

Tom, I don't think you are reading the quoted text properly. Matt's post and the text you quote point to the same thing: Adding a master fader (at unity gain) will not affect the sound in any sense whatsoever regardless of using the dithered or non-dithered mixer plug-in. The fader doesn't add dither. Dithering is there depending on whether you use the dithered plugin mixer or not. Adding a master fader at unity gain is only a visual thing.

Alistair

[MattGray]

Bingo!

[masteringhouse]

Originally Posted by UnderTow

Tom, I don't think you are reading the quoted text properly. Matt's post and the text you quote point to the same thing: Adding a master fader (at unity gain) will not affect the sound in any sense whatsoever regardless of using the dithered or non-dithered mixer plug-in. The fader doesn't add dither. Dithering is there depending on whether you use the dithered plugin mixer or not. Adding a master fader at unity gain is only a visual thing.

Alistair

I realize dithering is there in the dithered plug regardless, but my understanding is that each process from 24 to 48 back to a 24 bus invokes a dither process. It's not really completely explained where and when in the white paper from what I've read/understand. By having a "master" fader in the background to me just means that there is some sort of 48 or 56 bit "accumulator" for summing that all of the individual tracks are running through. I would tend to doubt that the processing is exactly the same even if only a visual aid is added to control this "accumulator".

Also the potential issue may have nothing at all to do with dithering or processing the individual channels but in a latency between the two channels by adding a master fader.

Obviously I'll have to do some testing when I get a break and will post results. I really hope that both you and Matt are right and what I and some of my clients have been hearing is some sort of placebo effect, though blind testing seems to prove otherwise. I would rather that it isn't the case for ease of use and flexibilty, so if anything I'm biased in the opposite direction.

Question, for a null test would you recommend:

1. A file with a few 1 sample clicks for lining up the results and removing any potential non-channel latency, followed by pink noise?

2. Given that dither is random, how is it possible to get any two to null completely if they are processeed in different passes?

[masteringhouse]

Hmmm...

Apparently we've been here before:

Pro tools master fader

I'll still test though, one of those times when someone has to call themself out.