Stereo Dithered Mixer

[pasanta]

In the unused plugin folder I have the Stereo Dithered Mixer plugin. I've found the section in the manual about it and am still confused. A search here shows the last discussion to have been around version 7 and for HD.

For normal stereo music mixing as it applies to PT 9........and not surround sound.......Is there a compelling reason I should use or even care about this? My mixes are always printed to track/exported or bounced to the same bit and sample rate as the session, then I use Samplitude to even all the mixes out and convert to cd using it's dither if needed.

[Craig F]

you don't want to double dither

[eledu]

I dont technically know why, bat I do hear that when stereo dithered mixer is inside the plugins folder it sounds better, it's like the audio is a little 3D and I can hear stereo instruments place a lot better ( their stereo image). I've even done an absolutely blind test with a great friend engineer sending him three diferents songs, printed with and without the stereo dithered mixer, and him choose always the audio printed with SDM in the plugins folder. Why is that? I don't know.... I think I must read the white papers.
Cheers

[Shifted Music]

on HD Systems people like to use the stereo dithered mixer.

IN fact based on what I read in the users guide the stereo and surround dithered mixers are only for HD systems.

PT 9 defaults to installing the surround mixer because that is the mixer where you can change the pan depth and or pan law.

its not changeable in the stereo mixer.

So just leave the surround mixer installed, its all good.

[boneboy]

The way I understand it is that without the dithered mixer, the moment you lower the fader below 0 it starts truncating the bits making the audio more harsh or brittle, etc. With the dithered mixer, it is doing just that...funneling all the bits as you move the fader below 0 and maintaining warmth, etc.

It is not applying dither to the Master bus. It only applies dither to the fader. You still need to dither down to 16 bits when bouncing from a 24 bit session. I think it's important to use it.

[eledu]

Originally Posted by Shifted Music

PT 9 defaults to installing the surround mixer because that is the mixer where you can change the pan depth and or pan law.

its not changeable in the stereo mixer.

So just leave the surround mixer installed, its all good.

Errr..... I have Protools9 (non HD) and when I open de Session Setup window I see there's PAN DEPTH and I can change it....
And I don't have the Surround Dithered Mixer in my Plugins Folder....
BEst.

[Greg Malcangi]

A little late to be adding to this thread but I thought I'd give it some closure.

More than a decade ago there was a huge thread on the DUC about the audio quality of the mixer in ProTools. There had been a lot of rumours flying around the professional world that the PT mixer didn't compare well to other summing mixers. The discussion got quite technical and some big names got involved. A conclusion was reached that there was a fault with the summing mixer in ProTools. The 48bit output (some plugins were 48bit at the time) on inserts and also on big sessions where the TDM chips had to create invisible submixes to link in more TDM chips, involved using 24bit wide data paths. This meant that anything at 48bit (the mix accumulator or 48bit plugin output) would have the 24 LSBs truncated. Many people thought this explained the weakness in the ProTools summing mixer.

DigiDesign took the whole thing pretty seriously and designed a stereo and surround mixer plugin which dithered at all the points where 48bit output would be truncated by the non-dithered mixer. Digi let us all download the plugins but didn't tell us which was which so we could do a blind test. As a general rule most people couldn't really tell a difference (as Digi had said all through the thread). Digi even setup a double blind test at a top studio with some big names and compared the two mixers with a mix on a Sony Oxford (generally regarded as about the best sounding digital mixer on the planet at the time).

The general conclusion at the end of it all was that there was nothing wrong with the ProTools mixer and that harsh digital recordings were generally a result of newbies not knowing what they were doing or experienced analogue engineers recording too hot for 24bit. However, there were possibly some situations where positive effects of the dithered mixer were just about noticeable and some people swore by it. As time marched on, processing power increased, the plugins got better, we got better at using ProTools and the whole thing became a bit of a non-issue. For a while though, it looked like Digi were in trouble, the competition really jumped on the debate and much of the really serious ProTools bashing started at that time.

As far as I remember, the whole thing was ProTools HD only, in fact I'm not even sure if there was a ProTools LE back then, the current version of PT was 5.1.3 (if I remember correctly)!

My advice is: Close your session, replace the mixer plugin in your Plugins folder with the dithered mixer from your unused Plugins folder, open your session again and have a listen. Or bounce two versions down (one with each mixer) and compare them. The dithered mixer eats up a little more processing power but who knows maybe you'll prefer it, my guess is though that you won't be able to tell the difference.

G

[daeron80]

Good reminder.

Originally Posted by Greg Malcangi

The dithered mixer eats up a little more processing power but who knows maybe you'll prefer it, my guess is though that you won't be able to tell the difference.

It depends on the material. In order for it to make a difference, the total amount of noise in the signal at each point where dither is added must be less than the amount provided by the 24-bit dither itself. Otherwise, the dither is just noise below the noise floor. Irrelevant. That's probably why most people couldn't hear it. It's a rare recording that has that little noise inherent to the tracks themselves.

[Greg Malcangi]

As I remember it, the potential problem wasn't related to recorded noise level in the track but to quantisation error caused by the truncation of the 48bit signal. It is not possible to record anything with a noise floor lower than the 24bit dither level (-138dB approx). Quantisation error can create correlated signals and therefore it can produce artefacts much higher in level than the digital noise floor. That's why we always dither when going from 24bit to 16bit, even though we are raising the noise floor. In effect, the dithering process randomises the quantisation errors, ensuring the errors are de-correlated from the program material. While these quantisation errors can be noticeable going from 24bit to 16bit, Digi's position was that these errors going from 48bit to 24bit would be way below the hearing threshold. In theory that is true but it is remotely feasible that enough correlated errors could interact with the program material to possibly be heard. The chances of that happening though are exceedingly slim. My conclusion and that of many others after extensive testing was that Digi was right all along even though I was one of the doubters to start with.

As I said though, there were some who were completely convinced the dithering mixer was a definite improvement, which is why I suggest giving it a try and making up your own mind.

G

[daeron80]

That's all true. But dither is noise. Granted, it's special noise in the sense that it's shaped to avoid being heard. But any noise in the vicinity of white causes dithering to take place. If you got (whitish) noise, you got dither. So if the noise in your tracks and mix is high enough, it will decorrelate truncation errors. If dither is below the noise floor of your mix, it will be buried by it and make no difference. So says Nika Aldrich, anyway, in Digital Audio Explained, ch 9, if I understand him right. And it fits what I've experienced over the years.