a comparison: mix engine of ProTools LE vs HD-TDM

[mixerguy]

So....

On a different thread a few people came forward and posted comments such as:

Originally Posted by Stoneface

.....Mixes do sound better on HD. If you want proof...try mixing on both systems and see if they are the same. ......

Originally Posted by jslevin

....Almost everyone would agree that Pro Tools LE cannot have a mix engine in the same league with Pro Tools HD. There is no real controversy about it.

JSL

(as an aside - I just love sweeping generalizations such as “almost everyone would agree”.... as if: jslevin has done extensive research, and polled everyone with any knowledge of the ProTools mix engine... and... “ Almost everyone would agree....” )


Well, I decided to do a test of the mix engine..... LE vs HD-TDM

I want to stress - NO PLUGINS WERE USED IN THIS TEST.... no external hardware was used at all. Both rigs were running PT 7.3 (either LE or HD-TDM)

I also did not use the bounce to disk feature at all. I merely had a big, dynamic mix of 43 tracks of audio, (with automation) and routed the mix to a stereo buss, and recorded the buss back to a stereo audio track. 24 bit, 48k. Levels & gain staging were done right.

Because the audio never went to analog - it stayed within ProTools - the clock source DOES NOT MATTER. Similarly, the converters on the rig does not matter, as the mix never left the box while being printed.

I printed the mix to separate stereo audio tracks (once on LE, once on HD-TDM) .... and then did extensive comparisons.....

Sorry - I can’t post clips... my client owns the multitrack - i do not.

I would urge anyone else to do a similar test - as it was very revealing for me.

What I discovered:

1) First and foremost - I could not perceive any significant, appreciable difference between the two mixes. I thought that they might have possibly sounded very, very slightly different, but nothing that I could quantify. I’ll have another listen with fresh ears another day. The TDM mix did NOT sound appreciably better than the mix done in LE. (or visa versa)

2) I noticed that the TDM mix - the stereo audio file - was a few samples LATE as compared to the LE mix. This should not be the case - they should be in perfect sync with each other.... so I’d say that there is a flaw somewhere.

3) I noticed that the TDM mix was 1/10th of a db hotter. (according to the peak display on each audio track) This was repeatable. This is an obvious flaw somewhere.

4) I put a trim plugin on both mixes, and flipped the phase on one, and they did not cancel out. I slowly nudged the TDM stereo mix audio file earlier - one sample at a time, and I could never get the mixes to cancel out totally. The mixes are somehow, slightly different. (No, I didn’t account for what the peak meters on the audio tracks showed me - that the TDM mix was seemingly 1/10th of a db hotter..... perhaps on another day I’ll try again....)

5) My intent was to do a test of JUST THE MIX ENGINE OF LE vs. TDM - and this is why I didn’t use any RTAS or HD-TDM plugins. That is a whole other debate, and something I’m not willing to get into right now.



My conclusions: something (or some things) are flawed - as why would the TDM mix be printed slightly later than the LE mix? and why would there seemingly be a slight level difference?

Digidesign goes to great lengths to show that mixing ITB is basically “perfect” but my tests show different.

OK so I’ve started (yet another) raging debate thread about mixing ITB - and TDM vs LE - everybody - ready - get set - GO!

[thethrillfactor]

Originally Posted by mixerguy

So....

On a different thread a few people came forward and posted comments such as:






(as an aside - I just love sweeping generalizations such as “almost everyone would agree”.... as if: jslevin has done extensive research, and polled everyone with any knowledge of the ProTools mix engine... and... “ Almost everyone would agree....” )


Well, I decided to do a test of the mix engine..... LE vs HD-TDM

I want to stress - NO PLUGINS WERE USED IN THIS TEST.... no external hardware was used at all. Both rigs were running PT 7.3 (either LE or HD-TDM)

I also did not use the bounce to disk feature at all. I merely had a big, dynamic mix of 43 tracks of audio, (with automation) and routed the mix to a stereo buss, and recorded the buss back to a stereo audio track. 24 bit, 48k. Levels & gain staging were done right.

Because the audio never went to analog - it stayed within ProTools - the clock source DOES NOT MATTER. Similarly, the converters on the rig does not matter, as the mix never left the box while being printed.

I printed the mix to separate stereo audio tracks (once on LE, once on HD-TDM) .... and then did extensive comparisons.....

Sorry - I can’t post clips... my client owns the multitrack - i do not.

I would urge anyone else to do a similar test - as it was very revealing for me.

What I discovered:

1) First and foremost - I could not perceive any significant, appreciable difference between the two mixes. I thought that they might have possibly sounded very, very slightly different, but nothing that I could quantify. I’ll have another listen with fresh ears another day. The TDM mix did NOT sound appreciably better than the mix done in LE. (or visa versa)

2) I noticed that the TDM mix - the stereo audio file - was a few samples LATE as compared to the LE mix. This should not be the case - they should be in perfect sync with each other.... so I’d say that there is a flaw somewhere.

3) I noticed that the TDM mix was 1/10th of a db hotter. (according to the peak display on each audio track) This was repeatable. This is an obvious flaw somewhere.

4) I put a trim plugin on both mixes, and flipped the phase on one, and they did not cancel out. I slowly nudged the TDM stereo mix audio file earlier - one sample at a time, and I could never get the mixes to cancel out totally. The mixes are somehow, slightly different. (No, I didn’t account for what the peak meters on the audio tracks showed me - that the TDM mix was seemingly 1/10th of a db hotter..... perhaps on another day I’ll try again....)

5) My intent was to do a test of JUST THE MIX ENGINE OF LE vs. TDM - and this is why I didn’t use any RTAS or HD-TDM plugins. That is a whole other debate, and something I’m not willing to get into right now.



My conclusions: something (or some things) are flawed - as why would the TDM mix be printed slightly later than the LE mix? and why would there seemingly be a slight level difference?

Digidesign goes to great lengths to show that mixing ITB is basically “perfect” but my tests show different.

OK so I’ve started (yet another) raging debate thread about mixing ITB - and TDM vs LE - everybody - ready - get set - GO!

Did you have the latency correction on or off in HD?

[mixerguy]

Originally Posted by thethrillfactor

Did you have the latency correction on or off in HD?

honestly - I can't remember - but - should it matter? I wasn't leaving the box... just audio to a buss, then recorded internally.

It was late by about 4 or 6 samples....

[Curve Dominant]

Originally Posted by mixerguy

(as an aside - I just love sweeping generalizations such as “almost everyone would agree”.... as if: jslevin has done extensive research, and polled everyone with any knowledge of the ProTools mix engine... and... “ Almost everyone would agree....”)

Jay, do you want to clarify anything here, before I chime in?

-E

[mixerguy]

Originally Posted by Curve Dominant

Jay, do you want to clarify anything here, before I chime in?

-E

the other thread is here:

Stock Pro Tools HD Vs. Mytek Pro Tools LE

[Curve Dominant]

Originally Posted by mixerguy

honestly - I can't remember - but - should it matter?

Fvck yeah, dude, it definitely should matter if you were mixing with plugins open as inserts.

???

[Curve Dominant]

Originally Posted by mixerguy

the other thread is here:

Stock Pro Tools HD Vs. Mytek Pro Tools LE

Gotcha. Going off to read that now.

[mixerguy]

Originally Posted by Curve Dominant

Fvck yeah, dude, it definitely should matter if you were mixing with plugins open as inserts.

???

DUDE - THERE WERE NO PLUGINS!!!!!!

i thought i made that perfectly clear!!!

[thethrillfactor]

Originally Posted by mixerguy

DUDE - THERE WERE NO PLUGINS!!!!!!

i thought i made that perfectly clear!!!

What about the trim plug?

That adds samples.

Also was there any automation?

[mixerguy]

Originally Posted by thethrillfactor

What about the trim plug?

That adds samples.

Also was there any automation?

the trim plugs were added only after the mixes were printed... they don't affect the audio that was printed.....and I could SEE that the HD-TDM audio was later when I zoomed way in.

and - yes - as I said in the original post - there was some automation in the mix. Only some gentle fader automation - some moves and rides... nothing crazy at all.

[Curve Dominant]

Originally Posted by mixerguy

DUDE - THERE WERE NO PLUGINS!!!!!!

i thought i made that perfectly clear!!!

My bad, my brother.

[thethrillfactor]

Originally Posted by mixerguy

the trim plugs were added only after the mixes were printed... they don't affect the audio that was printed.....and I could SEE that the HD-TDM audio was later when I zoomed way in.

and - yes - as I said in the original post - there was some automation in the mix. Only some gentle fader automation - some moves and rides... nothing crazy at all.

The reason i mentioned the trim plug is if you are trying to null it out perfectly because the plug itself delays the audio a certain amount it will never null it perfectly.

Its probably best to invert the audio with an audiosuite plug.

The automation does effect the mix somewhat in HD based on the buffer size.

[David Herbert]

Thanks for taking the time to do the comparison. Seems like you controlled for the variables and isolated the mix engines effectively. It's notable that there were some differences. It's also interesting that you didn't hear any subjective difference between the mixes.
Just thought I would throw those comments in before the sh!t storm starts.
David

[mixerguy]

Originally Posted by thethrillfactor

The reason i mentioned the trim plug is if you are trying to null it out perfectly because the plug itself delays the audio a certain amount it will never null it perfectly.

Its probably best to invert the audio with an audiosuite plug.

The automation does effect the mix somewhat in HD based on the buffer size.

I'm one step ahead of you - I put a trim plugin on BOTH audio tracks - but flipped the phase on just one of them.

(you) The automation does effect the mix somewhat in HD based on the buffer size.

[thethrillfactor]

Originally Posted by mixerguy

I'm one step ahead of you - I put a trim plugin on BOTH audio tracks - but flipped the phase on just one of them.

Gotcha.thumbsup

Originally Posted by mixerguy

(you) The automation does effect the mix somewhat in HD based on the buffer size.

Yeah i noticed long ago based on the buffer size in the HD playback engine the automation plays differently which in turn changes the sound of the mix.

[jslevin]

Originally Posted by mixerguy

as an aside - I just love sweeping generalizations such as “almost everyone would agree”.... as if: jslevin has done extensive research, and polled everyone with any knowledge of the ProTools mix engine... and... “ Almost everyone would agree....”

I'm glad you just love it. I admit, it does suffer from the weaknesses of any generalization, e.g., it can be picked to death by an obsessive. But we see hundreds of generalizations just like this on Gearslutz every day, and many of them, like this one, do more good than harm.

I think it's interesting that while you indulge in smug ridicule, you offer no evidence that my statement was incorrect -- none, zero. And here I thought you were all about the evidence. I still think the statement is essentially correct, if perhaps overly broad.

Originally Posted by Curve Dominant

Jay, do you want to clarify anything here, before I chime in?

Only that this guy is apparently talking about pure summing, not mixing in the sense that any professional mix engineer in the world (and almost all amateurs) would define it. (Next up: Proving that two calculators add whole numbers the same!) And hilariously, he does not even manage to prove that they do, in fact, sum the same!

By all means, whoever wants to discuss that 1/15th of a millisecond delay "issue," go right ahead. I am unsubscribing this thread two seconds after I hit post.

JSL

[Curve Dominant]

Originally Posted by mixerguy

the other thread is here:

Stock Pro Tools HD Vs. Mytek Pro Tools LE

OK, I read that thread.

JSLevin is a friend, so I would still prefer he chime in before I do.

EDIT: JSLevin posted the above while I was typing this.

BTW: Jay has a fantastic studio, for whatever that's worth. It is one of the most, if not THE most, respected studios in Philly. It is definitely out of the league of any LE-based home-schmome studio setup. So, let's not bash Jay. The guy not only knows his stuff in spades, but he has consistently put his money where his mouth has been, in regards to making Turtle Studios a premier tracking facility. Everyone I've ever recorded there, walked away feeling lucky that they recorded there. He picks his staff engineers carefully, nurtures their careers, and looks out for them. Jay is a mensch.

I've both recorded at Turtle, and mixed there. I've even mixed one particular song both on his HD system, and on my LE system. I'm tempted to type more, but I'm going to defer to Jay before I jump into any comparison-making, sonics-wise, for now.

I would, however, caution all of us from making too much of this. We need to keep the big picture in mind. In that regard, Jay was correct: the HD TDM mix system "works" better than an LE system does, within the demands of a pro tracking facility. (DISCLAIMER: NOT TALKING ABOUT THE MIXBUS SONICS HERE.)

Sonics are subjective and theoretical. But workflow is concrete, when you've got an ensemble playing which needs to be recorded. I suspect this was what JSLevin was actually referring to. And I suspect he is right.

-Eric Vincent @ Studio Curve Dominant

[zilver]

Originally Posted by Curve Dominant

Sonics are subjective and theoretical. But workflow is concrete, when you've got an ensemble playing which needs to be recorded. I suspect this was what JSLevin was actually referring to. And I suspect he is right.

-Eric Vincent @ Studio Curve Dominant

No, he was stating that "everybody" knows that TDM mixes the tracks better than LE......

Hans

[studjo]

thanks mixerguy for the report

Jo

[mixerguy]

Originally Posted by studjo

thanks mixerguy for the report

Jo

Yer very welcome!



I'd encourage others to do the exact same test - and see if they can hear any sonic differences.

I wanna stress - NO PLUGINS or external hardware was used.... to try to make it a fair test of just the mix engine with LE vs TDM.

[mixerguy]

Originally Posted by David Herbert

Thanks for taking the time to do the comparison. Seems like you controlled for the variables and isolated the mix engines effectively. It's notable that there were some differences. It's also interesting that you didn't hear any subjective difference between the mixes.
Just thought I would throw those comments in before the sh!t storm starts.
David

As i said - I thought that they might have possibly sounded very, very slightly different, but nothing that I could quantify.....

and one was clearly not better than the other.

in my opinion.

[amost]

add plugs and do it...that would be more real world IMO.

Originally Posted by mixerguy

So....


3) I noticed that the TDM mix was 1/10th of a db hotter. (according to the peak display on each audio track) This was repeatable. This is an obvious flaw somewhere.

Well there you have it.
That is going to be the only difference in a perfect, digital mathematical environment.
Everything that is louder sounds better to the human ear.
And of course what we hear is not what's there but what our brain wants us to hear.

I'd think if you can syncronize the output somehow and use the same algorythms which should be the same anyway (except dither) then you should get the same signal.
Of course this means eliminating volume differences at certain stages and is almost impossible so..


Anyone who thinks that outside of ergonomical and minor technical differences the actual MATH is different and therefore the sound, just doesn't know enough about digital audio.

Last but not least: Just because the output is a little hotter, this doesn't warrant a statement that "everyone knows DAW A sounds better than DAW B". I certainly don't. I know they don't even "sound" -different- (as in analogue equivalents).

[pjplayer]

Originally Posted by zwaps

Everything that is louder sounds better to the human ear.

are you saying a 0.1db difference is detectable by the human ear?

[Kiwiburger]

Anyone who thinks that outside of ergonomical and minor technical differences the actual MATH is different and therefore the sound, just doesn't know enough about digital audio.

That's a huge assumption to make.

[studjo]

Originally Posted by amost

add plugs and do it...that would be more real world IMO.

this is a test about the mix engine (and not how one programm handles plug vs the other). Saying HD sounds better than LE is a statement about the mixing process and that has nothing to do with the plugs (or you'd have to say HD handles plugs differently than LE - which is totally true)

just to make it clearer

Jo

[amost]

Originally Posted by studjo

this is a test about the mix engine (and not how one programm handles plug vs the other). Saying HD sounds better than LE is a statement about the mixing process and that has nothing to do with the plugs (or you'd have to say HD handles plugs differently than LE - which is totally true)

just to make it clearer

Jo

I hear ya but to me if you're mixing ITB...how it handles plugs(Mixing/Engine)is way, way, way relative . Everybody uses plugs mixing ITB.

If you're not using plugs...well then you don't need ADC and all that, etc. Okay the mix engine's the same now go mix ITB with no plugs.

YMMV of course.

[studjo]

you're absolutely right - but we first have to agree that the mix engine sounds different or not. If they sound the same you can go to the next step and check how both engines handle plugs and if a TDM plug sounds better than the native version ...
If they don't sound the same, well ...

Jo

[PhilE]

I cant see the point in this test at all- I cant think of anyone who would consider the track mixed just buy tweaking a few levels... you have removed anything difficult from the test- I would be very disturbed if there were big changes between the two systems.

The tests I've done are simple- I have tracked and mixed ITB and OTB on Mix, LE and HD systems... my fave is OTB work based around HD with Apogees. The worst sounding is the LE (both I and O TB) and the middle ground is my Mix rig- OTB mixing. I should strap some Apogees onto my rig and it will close the gap to HD even more.
Surely the only test that matters is to do what I have done- real world tests and everyone decide what suits their needs best- If I did a lot of Hip Hop I think my setup would struggle against even a new Mbox Pro and a bunch of plugs.

[davedarling]

Nice job Mixerguy. I think somehow the discussion went from comparing audio recording, and playback (with conversion taken out of the mix) to mix engine horse power,and features.

I'm an HD guy, but I also have an 002 at home for writing, and editing. When I record something through outboard adda on the 002 and take it into the HD studio,
it sounds (to me) exactly the same as if I would have recorded it through the same adda into the HD rig - i've done it.

who knew people were so passionate about....playback engines ?

see ya - Dave Darling