The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
bounce to disk VS. print new stereo track Virtual Instrument Plugins
Old 24th September 2007
  #91
Mastering
 

Quote:
Originally Posted by espasonico View Post
I´m going to try it again:

BTD doesn´t rely on the clock you are using ( at least that´s what I have been able to prove myself on a LE system )

RTD it does rely on the clock you are using so it makes perfect sense that there are diferences between BTD and RTD if mixing with a diferent clock used in recording !!
Forgive me if I misunderstood the motivations of espasonico, but I feel a need to respond in general:

It is very important for people to realize how a Synchronously-clocked DAW works. Repeat after me this mantra: THERE IS ONLY ONE CLOCK. THERE IS ONLY ONE CLOCK. THERE IS ONLY ONE CLOCK. This clock drives ALL the data within the DAW. The data received at the input to the RTD MUST BE IN SYNC WITH THE clock in the DAW. The data at the output of the mix bus also MUST BE IN SYNC with it. It is the law :-). Actually, if not, then you would not even be able to record the source for the RTD at all.

And, REGARDLESS of what the master clock driving the DAW is, the source for BTD is the digital data at the output of the mix bus. The source for RTD is the source that you choose in the input popup menu. If you pick the same mix bus as the source for RTD that would also be used for BTD, then you are getting the same data BY DEFINITION. It's an internal patch in the DAW that can't be screwed up unless the human being picks the wrong patch or source. If the source for RTD is a differrent or additional insert or plugin or analog processor, then it's a moot point, things will sound different... But if not, then the data is the same.

During BTD, the data at the output of the mix bus (or whatever bus you are recording from) is collected sample by sample. It is only a COINCIDENCE that there is a clocked system running at that time and that you happen to be able to listen to it.

During RTD, it's the same thing. You're tapping off the same data stream, if you patch it right!

Now, you wrote, entirely confusing me....

Quote:

So following my theory, if you are mixing something recorded thru a DIgidesign 192 and you are clocking to a BigBen for mixing, it should be a diference between the BTD and the RTD because diferent clocks are involved.
Let me see if I understand. Until you introduced the idea of different clocks for mixing and recording I thought we were on the same page. I thought we were strictly in mix mode and using a constant clock source (either Big Ben or 192).

In other words, do an RTD or a BTD in mix mode, constant clock, right?

I'm making the basic assumption that you are already in mix mode and you are choosing whether to record the output of the mix to a new track OR whether to use Pro tools BTD. I'm also assuming that the clock remains the same for both.

IRONICALLY, if you are mixing totally digitally, in the box, the clock you choose makes ABSOLUTELY NO DIFFERENCE! It ONLY AFFECTS THE LISTENING ephemerally during the bounce! You could change clocks for the BTD and for the RTD, if you are mixing totally digitally, in the box, and there will be no difference in the files created!

As long as there are no analog patch points in the mixdown (no tracks are going out for external analog effecs nor returning for conversion). HOWEVER, the clock does make a difference if you are using any analog patches. HOWEVER (again), as long as you keep that clock the same clock for your BTD versus RTD tests, again, it's a constant and not a variable and so RTD and BTD should still sound the same.

And if they DON'T, then your Pro Tools is broken, and you have to find out why they don't sound the same! The debate between RTD and BTD is not voodoo, this really is rocket science that you can depend on----if there is a difference, then PT is broken.

There are other influences which we have not brought up: Randomizing algorithms! Such as in a reverb that uses modulation to enhance its image. In that case EVERY SINGLE pass will be randomly different. If you test for a null during any two consecutive BTD or RTD passes, you will hear some reverb in the null. This is also normal. As to whether you can hear that difference... who knows.... maybe. But then, if every successive RTD is different from the previous, and you feel you hear differences between them, then, well.... .next post please :-)


BK
Old 24th September 2007
  #92
Lives for gear
 
Jerry Tubb's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
The debate between RTD and BTD is not voodoo, this really is rocket science that you can depend on
So Bob,

I really appreciate all your research and findings about Pro Tools BTD-vs-RTD.

Confirms what I've thought for a number of years, all other things being equal, it's the same.

There might have been a difference a decade ago, but no matter, that's ancient history.


Since you've become quite the expert with Pro Tools, I'd like to issue you a Challenge:

~ Do your next Mastering project with Pro Tools ~

1. Playback from, and Record back to PT.

2. Use your favorite Analog or Digital Processing in the Loop.

3. Use whatever Application you choose to do the final assembly and burn.

I'd be interested in your observations, and since this is a Mastering forum...

There are a number of us here that use PT for our daily mastering work; Andy, Tom, Matt, etc.

Might create some more interesting discussion.


All the Best - JT
Old 25th September 2007
  #93
Mastering
 

Quote:
Originally Posted by Jerry Tubb View Post
So Bob,

I really appreciate all your research and findings about Pro Tools BTD-vs-RTD.

Confirms what I've thought for a number of years, all other things being equal, it's the same.

There might have been a difference a decade ago, but no matter, that's ancient history.


Since you've become quite the expert with Pro Tools, I'd like to issue you a Challenge:
Hi, Jerry. Well, actually, I'm not an expert at all with Pro tools! I've done several mixes with it, I've used it, but do I know all the keystrokes by heart and how to edit with it? Not at all! I'm just quite knowledgeable about how digital audio works, have built several pieces of hardware and software... that's the extent.

Now, back to your proposal:

Quote:


~ Do your next Mastering project with Pro Tools ~

1. Playback from, and Record back to PT.
OK, in two parts.

----------------------

Regarding the SOUND:

HD or LE? With or without plugins? RTAS or TDM plugins?

With what converters? There are so many variables.

Pro Tools HD major weakness compared to other architectures is the truncation going to and from the TDM bus.

If I had to master with Pro Tools HD (and I would NOT prefer to master or edit in PT for reasons mentioned on pages 23 and 24 of the new Second edition of my book) I would:

a---- use only the mix engine, which is 48 bit.

b---avoid TDM plugins and if I understand correctly, RTAS is not succeptible to the same weaknesses, but I would have to test.

c---clock PT internally and feed a top notch external jitter immune DAC, at least Mytek, Cranesong HEDD or Lavry quality. And return through similar quality A/D.

d---return the signal into SADiE or Wavelab after external sample rate conversion for editing and PQ coding.

If we stick to those rules, I am 100% certain that PT HD would give you excellent results as a mastering program. There, do you feel better now :-).

Anyway, in studio A I only have LE available to me for the forseeable future. In Studio B we will have HD fully operational by end of October, I'm sure, as a mix room only. But I don't consider Studio B to be a high enough resolution environment to make the audible comparisons and we've never done full blown mastering in there so there is no listening experience to draw from So this returns us to Studio A and the only comparison I can make is with PT LE. I'll bet you it would perform about as well as SAdiE as an origination platform since it is 32 bit float.

-----------------

Regarding ergonomics:

Pro tools is not very suitable as a mastering platform NOT because of its sound, but because of ergonomic considerations that far outweigh the other issues. Anyone who has used a SADiE, Sonic, Sequoia or Pyramix understands what I am talking about.

For example, what about all the smooth crescendos, decrescendos and micro edits I can perform instantly, conveniently and ergonomically in SADiE using SADiE's ergonomic crossfades? I can't easily get those kinds of edits in PT using the "drag the dots" automation. And I can't intermix 24 with 16 in the same playlist. And I can't intermix interleaved with split either. So please don't ask me to go through the ergonomic torture just to prove that PT sounds good? We already know that PT sounds great if you are conscious of PT HD's limitations and use the basic a through d rules above. If you follow rules A through D I ASSURE you are getting great sounding results! You don't have to ask me to go through hell to prove that, please don't beg me to do that.

And even if you venture into TDM plugin land, if you only use one or two PI's conservatively, I wager that the slight grunge that PT introduces compared to Native programs will be insignificant. Hell, I wouldn't be building a PT mix room if I didn't feel that way. Yeah, for those who were watching me try to do this with Sequoia, I have to tell you that I too have succumbed into the Pro Tools mix camp. It's just the best choice amongst many other partially flawed alternatives for two main reasons:

----works best with external controllers such as Mackie. Fully integrated and ergonomic mixing

----compatible Playlists with what the clients send

If you know what you are doing and avoid too many paths through TDM you can get excellent sounding mixes in Pro Tools HD with the current dithered 48 bit engine.

BK
Old 25th September 2007
  #94
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
b---avoid TDM plugins and if I understand correctly, RTAS is not succeptible to the same weaknesses, but I would have to test.
Bob,

I'm assuming the reasoning has to to with a 24 bit path? Would you avoid going to an external processor (e.g. Weiss, HEDD, or Waves L2 hardware) for this reason?

As with any processing one has to measure the benefits over the pitfalls. There are quite a few TDM plugs that are worthy of a 24 bit conversion to and from. If not, there's usually an RTAS version to use, though I believe the I/O of most are 24 bit anyway on TDM.

I guess a null test of RTAS versus TDM is in order.

As far as the other issues, all good points but can for the most part be gotten around in various ways. My two biggest beefs are can't create a CD and can't run Algorithmix EQs (though there are also other worthy plugs that do not work on a non-PT system).

I'm trying to be more agnostic as far as DAWs, I hate 'em all :-)
Old 25th September 2007
  #95
Mastering
 

Quote:
Originally Posted by masteringhouse View Post
Bob,

I'm assuming the reasoning has to to with a 24 bit path? Would you avoid going to an external processor (e.g. Weiss, HEDD, or Waves L2 hardware) for this reason?
No, it has to do with an undithered and truncated 24 bit path in both directions (to and from TDM) that accumulates distortion each time you add another PI in series.

Contrast that with the architecture of the 48 bit mixer, which works in 48 bits (actually a little less to get a bit of headroom) until it is dithered to 24 bits on the way out. I do not know the hardware limitations of the TDM bus but I can see just a few things that Digi could do in software to lessen the impact.

Regardless, it's a very slow accumulation of grunge which can often easily be covered up by a little analog magic anyway. It's pretty hard to hear the effect of those TDM truncations until you have many PIs, especially if you are pushing them, and then you can't necessarily separate the distortion from the bus from the distortion from pushing those little digital compressors.

It's ok to be DAW-agnostic, but more better to be DAW-aware. When you learn the faults and the weaknesses you can get great product on most any modern DAW. When I get classical mixes that were done in the box in Pro Tools HD that sound harsh I don't immediately blame the TDM bus, but first look towards converters, mike choice and placement and poor engineering techniques.

BK
Old 25th September 2007
  #96
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
No, it has to do with an undithered and truncated 24 bit path in both directions (to and from TDM) that accumulates distortion each time you add another PI in series.
Bob -

Please explain the above further. If I have a TDM plug (e.g. Waves) there is an option to dither to 24, so I would have a signal going to the plug at 48, 48 bit internal processing and dither to 24 bit on a 48 bus (padded with zeroes) on output. This then possibly going to another plug that (should) dither back to 24 on output and pad with zeros on a 48 bit bus. Other than a potential truncation at the start (I don't know how Waves handles input from a 48 bit bus), there does not appear to be any undithered truncation in this scenario.

It would of course be nice if it stayed as a 48 bit path, but this is probably less or as audible as a 32 float being converted to 24. Certainly given an analog path in the equation, masked out entirely?

I think that the question remains, given certain plugs that may be available in PT and not other DAWs, you have to weigh the good with the bad. For example, would the addition of a Crane Song Phoenix or Massey plug outweigh the negligible effect of noise below that produced by an analog chain? If noise and potential truncation distortion at this level is so significant, why use tape or analog processing at all?

There are also issues regarding longevity of a product. When I first bought PT I was weighing it against Sonic. Personally I'm glad I went the way I did. I think that even after the Apocalypse and insects take over the Earth, there will be Pro Tools rigs still around and in use. I will likely add Sequoia or possibly another DAW to my work-flow, but I can't see replacing PT entirely. I think that once you get your TDM system in you may find more use for it than mixing as well. Particularly if you get into any video work.
Old 25th September 2007
  #97
Here for the gear
 

I think the difference between my perspective and others is I question everything. If I'm not getting the results I would like to hear, I'm not going to try and make myself feel better by falling back on the math. The math says it's the same, but I hear differently, and I trust my hearing.

Everone here has the 2 files, could just listen to them musicaly, but won't.
What if I posted them as, "Hey, which one has a better sonic quality?".
Nobdy here except 1 person has even listened to them musicly and posted an opinion. I don't care if you think they sound the same. But why not listen, without some technical test?
Old 25th September 2007
  #98
Lives for gear
 
Jerry Tubb's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
If we stick to those rules, I am 100% certain that PT HD would give you excellent results as a mastering program.
Very thorough reply Bob, thanks.

From your description, sounds like our workflow is a little different.

Which may render the "challenge" moot.

Agree with most of your observations, with the exceptions of using the internal PT clock, the need to "print" to another machine, and the ergonomics ; - )

Perhaps after AES when everyone has a little more free time, more discussion.

Cheers - JT
Old 25th September 2007
  #99
Gear Head
 

TDM/ RTAS NULL Test

On my HD system I just performed a quick Null Test between TDM and RTAS.

Using a typical Rock song:

1 stereo Track with 5 instances of Sony Oxford EQ TDM
1 stereo Track with 5 instances of Sony Oxford EQ RTAS

I recorded the OP of each to seperate stereo tracks.
When I put an instance of the Digirack EQ1 TDM on each of these tracks and phase inverted one they seemed to Null completely when played together.

I then bussed them both to another stereo audio track and recorded the sum of them seemed to "Null". When I used Audiosuite gain to apply SERIOUS gain (65dB) I can quite clearly hear the tune. I checked and when I recorded the sum down the music was playing at -87.1 dB

What does this mean?
Did I leave anything out?

Last edited by F. Davis; 26th September 2007 at 11:06 AM.. Reason: Edited to make more sense
Old 25th September 2007
  #100
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
Quote:
Originally Posted by F. Davis View Post
What does this mean?
Did I leave anything out?
Seems like a valid test to me (I'm assuming that the settings were the same or none used?). It would suggest that there is a difference, but not what the difference is. Could be a difference in the algorithm or math between the two versions, a difference in how the data is passed to and from the bus, or another issue.

Anyone have any ideas on narrowing this down?
Old 25th September 2007
  #101
Lives for gear
 

So What the best to do? or better yet what do you guys do the most? Bounce to disk or Re-track into stereo Recorder or something
Old 25th September 2007
  #102
Lives for gear
 
espasonico's Avatar
 

Quote:
Originally Posted by bob katz View Post
During BTD, the data at the output of the mix bus (or whatever bus you are recording from) is collected sample by sample. It is only a COINCIDENCE that there is a clocked system running at that time and that you happen to be able to listen to it.

During RTD, it's the same thing. You're tapping off the same data stream, if you patch it right!


BK
Ok, maybe the problem is that for RTD I understand going out by using AES/EBU and record to 2 tracks ( or a stereo track ) in the same session and not RTD by using internal busses. As far as I understand and please correct me if I´m wrong, when you RTD by AES/EBU ( or any other digital i/o ) then the clock is not a coincidence.

Please, correct me if I´m wrong.
Thanks
Old 26th September 2007
  #103
Lives for gear
 
MattGray's Avatar
 

Verified Member
Quote:
Originally Posted by masteringhouse View Post
Bob -

Please explain the above further. If I have a TDM plug (e.g. Waves) there is an option to dither to 24, so I would have a signal going to the plug at 48, 48 bit internal processing and dither to 24 bit on a 48 bus (padded with zeroes) on output. This then possibly going to another plug that (should) dither back to 24 on output and pad with zeros on a 48 bit bus. Other than a potential truncation at the start (I don't know how Waves handles input from a 48 bit bus), there does not appear to be any undithered truncation in this scenario.
With the current HD hardware you have a 24bit input, 24bit I/O plug in bus architecture, 48bit summing mixer & a 24bit output path. So you are wrong when you say that the signal going into the plug-ins is 48bit (regardless of the manufacturer), the signal is actually 24bit when it hits the first plug-in regardless of whether it's TDM or RTAS, then depending on the plug-in coding it can either be single precision 24bit internal DSP (most of the Digirack plugs, Focusrite etc.), double precision 48bit DSP (Waves masters bundle, Renaissance Bundle, URS etc.) or in the case of RTAS 32bit float or 64bit float internal DSP. If using all TDM plugs the output of each plugin is truncated to 24bit (unless it's dithered internally on the plugin itself such as Waves Masters). But this is certainly the minority not the majority! More of a concern then dither is the possibility of distortion creeping in if you don't pay very careful attention to the TDM plugs input & output metering especially when dealing with mastering levels. This is not an issue when using RTAS plugs as most have an extra 48db of headroom between inserts to save your ass from unreported clipping or the occasional over.

If using RTAS the first insert input is still 24bit fixed & then each plug hands over from one plugin to the next at 32bit float but will always hand back to the mixer on the last insert as 24bit fixed before re-entering the mix engine. In the case of bussing signals to an auxillary channel, the signal may well be allowed to grow to 56bits (48bits with a 8 bit floating mantissa) inside the summing mixer but when it hits the auxillary it's reduced to 24bit pre fader (depending on the dithered mixer plugin as to whether this is dithered or truncated without dither). This is similar to a master fader except the master fader can be used to reduce the wordlength to fit safely inside 24bits without distortion.

Sorry for the rant...

Matt
Old 26th September 2007
  #104
Lives for gear
 
MattGray's Avatar
 

Verified Member
Quote:
Originally Posted by Jazzmo View Post
I think the difference between my perspective and others is I question everything. If I'm not getting the results I would like to hear, I'm not going to try and make myself feel better by falling back on the math. The math says it's the same, but I hear differently, and I trust my hearing.
No the difference is we listen & if we hear a difference then we look for a logical user error, if we explore all options & still can't explain the difference then it's likely that you may of discovered a bug in the Pro Tools coding. More times then I care to remember it's always been down to operator error. Just remember Pro Tools is no new kid on the DAW block, it's the industry standard & as such they have some of the best people working for them coding the audio engine & checking it to make sure it's operating correctly even under extreme circumstances. Then someone like yourself comes along that doesn't know how Pro Tools works under the hood claiming that the Pro Tools coding is flawed.

Quote:
Everone here has the 2 files, could just listen to them musicaly, but won't.
What if I posted them as, "Hey, which one has a better sonic quality?".
Nobdy here except 1 person has even listened to them musicly and posted an opinion. I don't care if you think they sound the same. But why not listen, without some technical test?
No offense but trust me when I say 'yes there is some audible differences between your 2 files' but we've proved here that it's an operator related error not a Pro Tools flaw. Your test files are flawed as Bob K has explained to you, so why should other people comment on the audible differences?? What's the point?

What if I sent you 2 test files one RTD & the other BTD but done correctly & tested to prove that they null correctly, will you be so bold as to point out the differences for us by listening alone?

Matt
Old 26th September 2007
  #105
Lives for gear
 
MattGray's Avatar
 

Verified Member
Quote:
Originally Posted by F. Davis View Post
On my HD system I just performed a quick Null Test between TDM and RTAS.

Using a typical Rock song:

1 stereo Track with 5 instances of Sony Oxford EQ TDM
1 stereo Track with 5 instances of Sony Oxford EQ RTAS

I recorded the OP of each to seperate stereo tracks.
When I put an instance of the Digirack EQ1 TDM on each of these tracks and phase inverted one they seemed to Null completely when played together.

I then bussed them both to another stereo audio track and recorded the sum of them seemed to "Null". When I used Audiosuite gain to apply SERIOUS gain (65dB) I can quite clearly hear the tune. I checked and when I recorded the sum down the music was playing at -87.1 dB

What does this mean?
Did I leave anything out?
What you've probably exposed is the difference between the 2 signal structures in HD, which is likely to be the TDM truncation distortion which wouldn't be found in the RTAS output. Remembering that 5 TDM inserts will be truncating to 24bit 5 times & I'm pretty sure Sony don't dither internally on their plugs so the result is low level truncation distortion which when amplified by 65db makes the distortion more audible. With the RTAS plugs they are staying at 32bit float between each plug-in insert so you aren't truncating between each step only on the last insert will it be truncated back to 24bit creating a difference between the 2 bounces when you flip the phase.

I've never really had success in Pro Tools when comparing RTAS bounces to TDM bounces even when using the exact same plug in versions. In fact this is the only time I've had null tests fail. I hounded Digidesign about the 24bit TDM plugin bus asking them to make it 48bit on input & output. Their response was that they are aware of the issues this can cause & that they are working on a solution. We'll see if the AES show reveals their answer to this concern.

Matt

Matt
Old 26th September 2007
  #106
Lives for gear
 
MattGray's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
No, it has to do with an undithered and truncated 24 bit path in both directions (to and from TDM) that accumulates distortion each time you add another PI in series.
But as you've pointed out in another thread Bob, the Weiss units aren't dithering to 24bits correctly either, so what makes a Weiss any different from a TDM plugin that truncates in this instance?

Personally I'd prefer to see a 48bit path throughout the whole system then you would only have to worry about dither on the final output.

Matt
Old 26th September 2007
  #107
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
Quote:
Originally Posted by MattGray View Post
With the current HD hardware you have a 24bit input, 24bit I/O plug in bus architecture, 48bit summing mixer & a 24bit output path. So you are wrong when you say that the signal going into the plug-ins is 48bit (regardless of the manufacturer), the signal is actually 24bit when it hits the first plug-in regardless of whether it's TDM or RTAS

Sorry for the rant...

Matt
Matt - I realize this, what I meant was it was coming in off of a 48 bit bus. I should have been more clear but thought that I said this earlier when I mentioned plugs are 24 bit I/O.

Quote:
Originally Posted by MattGray View Post
But as you've pointed out in another thread Bob, the Weiss units aren't dithering to 24bits correctly either, so what makes a Weiss any different from a TDM plugin that truncates in this instance?
I was going to mention this but Matt beat me to this point. I'm assuming Bob uses the Weiss unit regardless because the sound quality that it achieves is superior to any plug that he has operating within the internal bit depth of the DAW he chooses to use, or there are features that aren't available elsewhere. Likewise with analog where the S/N ratio of most outboard devices and tape is closer to 16 bit.

Great sound has to do with lot more than the bit depth and S/N ratio of a device. Would it be technically superior to provide a wider range? Sure. Is it worth sacrificing other features that have a larger component in creating better sound quality? Doubt it.

My rant now complete ...
Old 26th September 2007
  #108
Lives for gear
 
Ben F's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
Regarding ergonomics:

Pro tools is not very suitable as a mastering platform NOT because of its sound, but because of ergonomic considerations that far outweigh the other issues. Anyone who has used a SADiE, Sonic, Sequoia or Pyramix understands what I am talking about.

BK
Well said Bob. I don't understand why anyone would use Protools for mastering after using a dedicated mastering program like Sequoia...other than sticking with the Mac OSX platform. But i digress...
Old 26th September 2007
  #109
Lives for gear
 
Jerry Tubb's Avatar
 

Verified Member
Quote:
Originally Posted by Ben F View Post
Well said Bob. I don't understand why anyone would use Protools for mastering after using a dedicated mastering program like Sequoia...other than sticking with the Mac OSX platform. But i digress...
Seems as tho' you've got some kind of axe to grind Ben?

Why anyone would care enough to criticize someone's choice of DAW is mystifying.

I'm sure us PT guys could ~yet again~ run down the whole list of reasons we work this way... but that would be redundant. And I bet we're all running at full throttle with busy mastering schedules... using Pro Tools.

JT
Old 26th September 2007
  #110
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
fuuck Ben.

(just kidding!!!)

Please no more why PT shouldn't be used for mastering threads. These are about as boring as debates on the loudness war.
Old 27th September 2007
  #111
Mastering
 

Quote:
Originally Posted by masteringhouse View Post
Bob -

Please explain the above further. If I have a TDM plug (e.g. Waves) there is an option to dither to 24, so I would have a signal going to the plug at 48, 48 bit internal processing and dither to 24 bit on a 48 bus (padded with zeroes) on output. This then possibly going to another plug that (should) dither back to 24 on output and pad with zeros on a 48 bit bus. Other than a potential truncation at the start (I don't know how Waves handles input from a 48 bit bus), there does not appear to be any undithered truncation in this scenario.
Yup. That's one of the reasons why a knowledgeable engineer like you may get better results than just anyone "fooling" with TDM. I don't think the option to dither to 24 is built into many Waves plugs. If I recall correctly, however, the Renaissance plugs internally calculate to 48 and then dither to 24 on their output by default, even without telling you on the front panel. This can be tested for with something like Spectrafoo.

You are correct, dithering to 24 bits, ESPECIALLY at 96 K, is a pretty damn good way of getting around the truncation problem.

How much is this a worry? How much is Pro Tools' "sound" compromised by this practice? Very little, and in most practical cases, if you just minimize the number of paths and don't push the plugs too much, the differences will be negligible compared to a fully floating point path, for example.

Would adding an analog path mask out the equation? I don't know the TDM architecture that well, yet, but it seems to me that you can't get rid of the truncation on the way in from the mixer side, and just remember that noise is not the same as adding dither, so that the analog path's noise is not going to "dither the signal that's already been truncated." However, analog noise does cover the distortion slightly and make it more palatable. That is for sure one of the true advantages of analog gear...

Have their been any fair and scientific shootouts between, say, Nuendo, and Pro Tools, running identical plugs, levels and processing? I've read of some, but not sure if they were properly conducted.

Oh, I do agree I made the right choice building a Pro Tools system, there will be lots of uses for it. I just wish that there were other true alternatives to Pro Tools used by a sufficient user base to establish a dual standard. Without competition, the leader tends to stagnate, you know.

BK
Old 27th September 2007
  #112
Mastering
 

Quote:
Originally Posted by MattGray View Post
But as you've pointed out in another thread Bob, the Weiss units aren't dithering to 24bits correctly either, so what makes a Weiss any different from a TDM plugin that truncates in this instance?
You are correct, Matt. The bottom line is that many of these distortions are very small, often covered by noise and the effect is cumulative. I think some of the grungy PT mixes that I receive are from abuse of processing, not from the TDM architecture itself.

As for what distinguishes a Weiss from the typical TDM plug, it is the quality and precision of its internal processing. Truncating from a high quality, Weiss-programmed 40 bit float Sharc down to 24 bits is a lot more palatable than taking a run of the mill plugin and doing the same.

When Daniel gets the 24 bit dither back into the Weiss boxes I plan on doing a real shootout. I wager it will be EXTREMELY hard to hear and prove! Matt, I believe Daniel is mistaken that the latest software has the dither. I've sent him technical measurements. Daniel and I will talk at AES.

This might change greatly if Algorithmix is able to port their EQs to the PT platform.

And Jerry can be perfectly comfortable working in PT for mastering, I just don't think it is the most efficient or ergonomic method. But if he starts in PT and terminates in Wavelab or Sequoia or Sadie he won't lose that much time.

Quote:

Personally I'd prefer to see a 48bit path throughout the whole system then you would only have to worry about dither on the final output.
Yes, but actually it is quite a difficult job to accomplish this with Motorolas. You lose a lot of cycles adding protocol to keep the low and high bytes identified. Probably Digi's 48 bit mix engine does it all in one chip, which is easier to do than if you have multiple chips involved, as in TDM. Natively, a Motorola processor is 24 bits in and out, so you have to send double bytes and identify them with a protocol or a header.

BK
Old 27th September 2007
  #113
Lives for gear
 
Ben F's Avatar
 

Verified Member
Quote:
Originally Posted by Jerry Tubb View Post
Seems as tho' you've got some kind of axe to grind Ben?

Why anyone would care enough to criticize someone's choice of DAW is mystifying.

I'm sure us PT guys could ~yet again~ run down the whole list of reasons we work this way... but that would be redundant. And I bet we're all running at full throttle with busy mastering schedules... using Pro Tools.

JT
It's not criticism, just a difference of opinion and dare I say it a suggestion to stop this endless debate about 48bit vs 32bit float Protools TDM vs LE dithering etc etc. Why stick with a program if you're not happy with the math? Use a full 32bit program like Sequoia and sleep well. Not that clients could ever hear the difference anyway...

I use both Protools and Sequoia every day, plus Logic Studio, Ableton Live, Reaktor, whatever does the job best. And am quite busy as well.
Old 27th September 2007
  #114
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
Does the modulating noise floor of 32 bit float bother anyone? Wouldn't noise (if audible at this level) be less noticable if it were consistent?
Old 27th September 2007
  #115
Mastering
 

Quote:
Originally Posted by masteringhouse View Post
Does the modulating noise floor of 32 bit float bother anyone? Wouldn't noise (if audible at this level) be less noticable if it were consistent?
I don't know how to separate the issues of 32 float's potenial noise modulation from other issues of resolution and purity of tone. Noise modulation is certainly one of the areas that concerned Andy Moorer in his classic paper. I'm certainly impressed with the 40 bit Sharc as implemented by Weiss. A more robust sound quality than most 32-bit float equalizers.

HOWEVER, not all things are created equally! You can do double precision float or other types of calculations within a filter or equalizer and then pass a 32 bit float number out to the system and get great sounding results. Witness the Algorithmix Red. So 32 bit float as an interchange number is not bad at all.
Old 27th September 2007
  #116
Lives for gear
 
masteringhouse's Avatar
 

Verified Member
Quote:
Originally Posted by bob katz View Post
I don't know how to separate the issues of 32 float's potenial noise modulation from other issues of resolution and purity of tone. Noise modulation is certainly one of the areas that concerned Andy Moorer in his classic paper. I'm certainly impressed with the 40 bit Sharc as implemented by Weiss. A more robust sound quality than most 32-bit float equalizers.

HOWEVER, not all things are created equally! You can do double precision float or other types of calculations within a filter or equalizer and then pass a 32 bit float number out to the system and get great sounding results. Witness the Algorithmix Red. So 32 bit float as an interchange number is not bad at all.
Thanks again Bob, your patience and commitment to answering these questions amazes me.

Given the info above one of the things I plan to try today is using a dither plug before output to the converter on my analog chain and the same before any TDM plug. Coming back out of the plug is unfortunately a case of the designer having to implement.
Old 20th October 2007
  #117
Here for the gear
 

Well i've been gone for a while and catching up with this thread.

First off, I'm not trying to prove anything. my impression is most of you guys are mad scientists. I was most amused by the comment "If it was done correctly" ? What was incorrect about what I did? You said you heard a difference, well?

There is a diference, and thats a fact. Why there is is up to you geeks to figure out. If you can't hear it, then I suggest eather you or your monitoring is not up to par, but I doubt that.

I have heard differences between tracking on external and internal hard dirves, between balanced and unbalanced power, between light pipe and aes, between cheap power cords and pro power cords, between having a technical ground and not, between lifting the ground and not. IMHO, the power supply has one of the biggest impacts of the overall sound.

PTs is not broken, digital audio is not broken, it's just to me there is a gray area that's beyond explaination.
Old 20th October 2007
  #118
Lives for gear
 
dave-G's Avatar
 

Verified Member
Quote:
Originally Posted by Jazzmo View Post
There is a diference, and thats a fact. Why there is is up to you geeks to figure out. If you can't hear it, then I suggest eather you or your monitoring is not up to par, but I doubt that.
I think the questions were pointed towards your methods because they could indeed point to the "fact" coming from something other than what you think it's coming from. As the guy who's asserting it's fact, the responsibility of proof is yours.
Quote:
Originally Posted by Jazzmo
I have heard differences between tracking on external and internal hard dirves,
Ummm .. uh ...
Quote:
Originally Posted by Jazzmo
between balanced and unbalanced power
Ok. Sure.
Quote:
Originally Posted by Jazzmo
between light pipe and aes, between cheap power cords and pro power cords
Do you prefer silver or copper?
Quote:
Originally Posted by Jazzmo
between having a technical ground and not, between lifting the ground and not. IMHO, the power supply has one of the biggest impacts of the overall sound
Sure, sure, and no argument.
Quote:
Originally Posted by Jazzmo
PTs is not broken, digital audio is not broken, it's just to me there is a gray area that's beyond explaination.
No offense, Jazzmo .. but given the topic, and some of the examples you cite, I think the "gray area" is your brain.

-dave
Old 20th October 2007
  #119
Gear Guru
 
UnderTow's Avatar
 

Verified Member
Quote:
Originally Posted by Jazzmo View Post
There is a diference, and thats a fact. Why there is is up to you geeks to figure out.
Yes and several possible explanations have been given.

Quote:
I have heard differences between tracking on external and internal hard dirves,
Oh boy...

Alistair
Old 20th October 2007
  #120
Lives for gear
 
beyarecords's Avatar
 

Quote:
Originally Posted by UnderTow View Post
Oh boy...
hehheh
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
philosi / Mastering forum
7
poketman / Low End Theory
7
jcool / So much gear, so little time
6
basement / High end
6

Forum Jump
Forum Jump