TDM Plug In Distortion compared to RTAS

[el Daniel]

Please read this quote from Paul Frindle before making those assumptions, he is referring to Pro Tools yes you have to look out for intersample peaks on the final output post limiter, but please pay particular notice to points 2 & 3 which refer to TDM fixed point plugs.

Why do you think "less capable plug-ins" is the same as TDM?, he's talking about plug-ins internal headroom, not fixed or floating point plug-ins. For example, boost a band in an EQ and cut another one, the boost could take the level over 0dBFS and it would be shown on the output meter clip, then the cut would reduce the level so it could be again in a legal level and no clip would be shown, if the plug-in has no internal headroom it will be clipping internally with the boost and the cut would bring it down the clipped signal and it wouldn't be shown on the meter as a red light. I think I said in my post "unless the plug-in has no internal headroom", so I don't see the incompatibility between what Paul Frindle says and what I said.

"Gotcha. Reconstruction errors are only on the D to A process. Thanks Paul for clearing things up, including the dither on the trim, I was wondering about that too."

Not necessarily. Some processes may have the equivalent of reconstruction within the applications themselves. They may therefore spit out larger value samples than those that went in - without any increase in volume or audible change in the programme!! This was one of the things I demonstrated on the PSW thread.

However since these increases in sample values should show up on the DAW meters you can reduce levels to avoid them as they are not hidden from you - like they may be in a DAC situation.

However since these increases in sample values should show up on the DAW meters you can reduce levels to avoid them as they are not hidden from you

Daniel

[Matt_G]

Why do you think "less capable plug-ins" is the same as TDM?

Because they don't have external headroom they have 24bit fixed outputs just like a DAC. This where intersample peaks can cause problems because they would carry over & come out in any subsequent processing. Not so with the properly designed 32bit or 64bit float plugs.

I think I said in my post "unless the plug-in has no internal headroom", so I don't see the incompatibility between what Paul Frindle says and what I said.

Once again intersample peaks can occur after the 24bit fixed output if the levels are even -0.5dbfs the peaks can still have reconstruction errors through other processes down the chain even though it may not light the meters up.

However since these increases in sample values should show up on the DAW meters you can reduce levels to avoid them as they are not hidden from you

Please note the word "should", that's not an absolute in my book. Also he mentions in point 4. (see below) about the GML EQ which I assume to mean is the MDW EQ which is 48bit DP. Besides this I have personally heard slight distortion from TDM plugs with DP when run close to full scale even though the red lights don't come on.

4. Similarly, being aware that some plug-ins have the equivalent of reconstruction INSIDE their processes and therefore can make a perfectly legal sample value signals (i.e. no red lights) into and illegal sample values (like the DAC above) but actually within the workstation. These may be seen on meters, but confuse people badly and may end up forcing you to make a final mixes much less loud than they could have been. Some plug-ins can do this without any modification to the freq response or loudness - even when set to 'flat' or in 'bypass'. I demonstrated this effect using the noise PT noise generator. The GML EQ is an example of this which was noted in the PSW thread - and it is not unique :-(

[gk]

Hello all,

Thank you for the honest debate and participation in this thread. Sorry to be so slow, but since many of you have called for an official Digi response I thought I’d better chime in!

First, let me say that we are aware of the issue with TDM bus being 24-bit and are considering ways to address it with our future products. That said, most people can work comfortably within the confines of the TDM bus (with the notable exception of some mastering engineers who are required to work at the uppermost limits of headroom). You do have to pay closer attention to gain structure than you would with a floating point system, but hopefully headroom isn’t the only important thing to your work and the other benefits of Pro Tools HD make it one of the best sounding platforms to work on.

With each new version of hardware and software, we strive to make improvements to Pro Tools and address the many different needs of our customers and issues like this (our architecture) are a high priority for us. It's an endless list, but your input helps us determine the order of what gets done, so I encourage you to keep giving us your constructive criticism. There are several ways to address the headroom issue but, as I’m sure you can understand, I can’t give you specific information on how or when we’ll be able to address it.

Thanks for listening and giving us your candid feedback.

Best regards,

Gannon Kashiwa
Professional Products Market Manager
Digidesign

[Iker]

Hi Matt

Interesting discussion...

Just a thought you might want to consider,

on the 32 bit floating format the first bit, is the sign magnitude bit (one=negative, zero=positive or zero), followed by an eight bit exponent and then followed by 23 bits of a normalized two’s complement fractional part of the mantissa. Other formats might invert the position of fraction and exponent.

The fact that you are able to represent huge and the tiniest of numbers doesn't mean you are able to represent all of them. border conditions render interesting situations where the format is pushed to the limit and undesirable things start to happen, and you have those conditions on the zero crossing of a signal...

thanks

[M2E]

Hello all,

Thank you for the honest debate and participation in this thread. Sorry to be so slow, but since many of you have called for an official Digi response I thought I’d better chime in!

First, let me say that we are aware of the issue with TDM bus being 24-bit and are considering ways to address it with our future products. That said, most people can work comfortably within the confines of the TDM bus (with the notable exception of some mastering engineers who are required to work at the uppermost limits of headroom). You do have to pay closer attention to gain structure than you would with a floating point system, but hopefully headroom isn’t the only important thing to your work and the other benefits of Pro Tools HD make it one of the best sounding platforms to work on.

With each new version of hardware and software, we strive to make improvements to Pro Tools and address the many different needs of our customers and issues like this (our architecture) are a high priority for us. It's an endless list, but your input helps us determine the order of what gets done, so I encourage you to keep giving us your constructive criticism. There are several ways to address the headroom issue but, as I’m sure you can understand, I can’t give you specific information on how or when we’ll be able to address it.

Thanks for listening and giving us your candid feedback.

Best regards,

Gannon Kashiwa
Professional Products Market Manager
Digidesign

Hey GK,

Will the new 7.4 address this problem? Or, is it more of a hardware thing? I can't see it being a hardware thing but more a software problem as seeing that Pro Tools HD would benifet if it was 48bit throughout instead of after every plugin it's dithered down to 24bits. Am'I correct in thinking this?
So PTHD could still use the 192i/o but soon as the audio hits the software it would up it to 48bit internally throughout until the last plugin on the Master Fader which would bring it down to 24bit.
Wouldn't that fix the problem and end all this?
If so, why hasn't this just been done? Do we need better or more powerful HD cards? Hey, maybe just take up more slots.

Is this correct thinking?

M2E

[JFreak]

Pro Tools HD would benifet if it was 48bit throughout instead of after every plugin it's dithered down to 24bits. Am'I correct in thinking this? So PTHD could still use the 192i/o but soon as the audio hits the software it would up it to 48bit internally throughout until the last plugin on the Master Fader which would bring it down to 24bit. Wouldn't that fix the problem and end all this? If so, why hasn't this just been done? Do we need better or more powerful HD cards?

Plugs *could* be 48bit throughout the system, but that would naturally mean fewer plugs per chip because double bit depth requires double power. Yes, 48bit would give more headroom, but it would not eliminate the problem: the user who wants to ride the signal on the edge. Besides mastering engineers, nobody should be trying to hit the full scale.

I mean, for mixing, there's no reason to do that. Whatever you record is limited to about 120dB dynamics, or whatever your converters can handle. That's the best case theoretical scenario and real-world signals are usually less dynamic than that. So there is already at least 24dB headroom before full scale for this theoretical and very dynamic signal. Usually that kind of signal gets compressed, which will give you even more headroom after the compressor plug, due to the very nature of compression.

Now if one thinks about this for a second, it is easy to understand that if the signal requires -say- 120dB and the plugins has 144dB to use, then it is not necessary to use the upper 120dB of the plugin "workroom" but it is perfectly okay to stay on the lower 120dB and as long as you have those 120dB it's all right. The problem comes when the mixing engineer thinks that the stuff sounds better when the signal is "almost clipping". Sure, louder tends to sound better, but that's why there's master fader or monitor volume control! You can always monitor louder if you need the volume level, that is a bad excuse for boosting signal level within a plugin.

In most cases it is more beneficial to lower volume or cut something instead of boosting everything you can. After all, mixing is all about balance. It is a whole lot easier to get a good balance *below* the problematic 0dBFS than trying to hit it all the time. Just get the mix done and send it to mastering (or try to do it yourself if you want to) for getting the desired level.

Also...

I don't mind if the next generation Protools has bigger this or larger that. It is always a good thing to have more headroom. It's just that IMO current systems have *enough* everything and it's always up to the operator how to use it.

[M2E]

Plugs *could* be 48bit throughout the system, but that would naturally mean fewer plugs per chip because double bit depth requires double power. Yes, 48bit would give more headroom, but it would not eliminate the problem: the user who wants to ride the signal on the edge. Besides mastering engineers, nobody should be trying to hit the full scale.

I mean, for mixing, there's no reason to do that. Whatever you record is limited to about 120dB dynamics, or whatever your converters can handle. That's the best case theoretical scenario and real-world signals are usually less dynamic than that. So there is already at least 24dB headroom before full scale for this theoretical and very dynamic signal. Usually that kind of signal gets compressed, which will give you even more headroom after the compressor plug, due to the very nature of compression.

Now if one thinks about this for a second, it is easy to understand that if the signal requires -say- 120dB and the plugins has 144dB to use, then it is not necessary to use the upper 120dB of the plugin "workroom" but it is perfectly okay to stay on the lower 120dB and as long as you have those 120dB it's all right. The problem comes when the mixing engineer thinks that the stuff sounds better when the signal is "almost clipping". Sure, louder tends to sound better, but that's why there's master fader or monitor volume control! You can always monitor louder if you need the volume level, that is a bad excuse for boosting signal level within a plugin.

In most cases it is more beneficial to lower volume or cut something instead of boosting everything you can. After all, mixing is all about balance. It is a whole lot easier to get a good balance *below* the problematic 0dBFS than trying to hit it all the time. Just get the mix done and send it to mastering (or try to do it yourself if you want to) for getting the desired level.

Also...

I don't mind if the next generation Protools has bigger this or larger that. It is always a good thing to have more headroom. It's just that IMO current systems have *enough* everything and it's always up to the operator how to use it.

Hey JFreak,

Well, part of the reason I'am wondering this is because I use the Waves SSL G-Buss Compressor and love it. They mention that the more you push level in it the more you get that SSL sound out of it. They also mention that yes, it's ok to clip the plugin....WHAT!!!!!!!!!!! Well, I've been trying it and yes, it does give you that warm as fuzzy feeling inside like the SSL but, what about that strong signal coming out of the plugin and back to the 24bit Master Fader? Something just isn't right here, is it?

Also, when they fix it, before they do, did they not know about this before releaseing HD? The system is based off of plugins, right? Really weird....

M2E

[uno1234]

This is the difference between pro tools digital mixing and SSL digital mixing, and why I would never buy an Icon, but I'm close to buying one of these:

http://www.solid-state-logic.com/music/c200.html

You cannot peak it. Without getting deep into the technical, the console is entirely floating point except on it's hardware outputs. You can mix 12 hot drum signals together, with tons of EQ boost, and you will never peak. There is something like 900 db of headroom when combining signals together. Really you have to talk to an SSL rep about this, but it is pretty incredible. Digi's technical white paper is a joke. While being "accurate," it leaves out a lot.

[M2E]

This is the difference between pro tools digital mixing and SSL digital mixing, and why I would never buy an Icon, but I'm close to buying one of these:

http://www.solid-state-logic.com/music/c200.html

You cannot peak it. Without getting deep into the technical, the console is entirely floating point except on it's hardware outputs. You can mix 12 hot drum signals together, with tons of EQ boost, and you will never peak. There is something like 900 db of headroom when combining signals together. Really you have to talk to an SSL rep about this, but it is pretty incredible. Digi's technical white paper is a joke. While being "accurate," it leaves out a lot.

Yeah but, how much is something like that? I thought that Pro Tools Headroom was the same if not bigger, right?

Little help here....

M2E

[uno1234]

The SSL C200 is around $350,000, completely dependent on your configuration. Yes, it is more expensive. Obviously. But between the dramatically better mixing engine, the amazing eq and compression, and the far far better tactile control surface... it really seems worth the money. Oh, and it doesn't crash. Ever. Imagine that.

The C200 has something like 900 db of headroom in the mix engine. Nothing like Pro Tools.