PT mixer headroom severely reduced to 20dB over 0dBFS. Bug?

[paul_g]

Anyone do a YouTube video on this?

[Tweakhead]

It always helps a bug to get fixed quickly when nearly every signature in a thread is longer than the post it is associated with. Avid techs love this. It helps their "wading process".

[Rich Breen]

Originally Posted by daeron80

Holy smokes. That means the mixer has reverted from floating- to fixed- point type.

I'm not sure that's what it means at all. The 'mixer' seems to be OK as you can increase gain with faders all you want without issue - something specific to this particular method (audiosuite, clip gain) is creating the issue, no? Anyway, not enough time to really go down this rabbit hole right now - again, I don't think this affects my workflow as I can't imagine why I'd ever want to clip-gain something 20dB over full scale, but I'm curious to see what's really happening.

[daeron80]

Originally Posted by Rich Breen

I'm not sure that's what it means at all. The 'mixer' seems to be OK as you can increase gain with faders all you want without issue - something specific to this particular method (audiosuite, clip gain) is creating the issue, no? Anyway, not enough time to really go down this rabbit hole right now - again, I don't think this affects my workflow as I can't imagine why I'd ever want to clip-gain something 20dB over full scale, but I'm curious to see what's really happening.

Could be. The OP said "The same issue occurs using any other method when you go past +20dB within the Protools mixer, like having 2x +12dB trim on a track.

Delving deeper into this, it looks like any sample OVER +20dB is processed as -INF. So it's not digital clipping where anything over +20 is just clipped at +20."

I saw the first statement and jumped to my original conclusion. But the second statement seems to contradict the first and to introduce the idea of an overflow bug in the mixer code. I'm in the middle of a project involving 16 hours of content or I'd do some experiments to see what's really happening to the waveform.

[Farhoof]

Originally Posted by Rich Breen

I'm not sure that's what it means at all. The 'mixer' seems to be OK as you can increase gain with faders all you want without issue - something specific to this particular method (audiosuite, clip gain) is creating the issue, no? Anyway, not enough time to really go down this rabbit hole right now - again, I don't think this affects my workflow as I can't imagine why I'd ever want to clip-gain something 20dB over full scale, but I'm curious to see what's really happening.

I'm not seeing this. When using the mixer alone I get distortion as well. Using track volume, aux volume and masterfader volume, any combination.
(can't use trim volume as it doesn't allow you to go past +12dB)

[JFreak]

I have not even bothered testing this. Seriously guys if you need to boost 20dB over full scale there is something wrong with your mix

[XJENSEN]

Originally Posted by JFreak

I have not even bothered testing this. Seriously guys if you need to boost 20dB over full scale there is something wrong with your mix

The matter of the fact is that something is wrong with Pro Tools.

[wwittman]

What's an example of a reason you'd want anything 20 db over full scale??

[Southsidemusic]

Originally Posted by paul_g

Anyone do a YouTube video on this?

Who cares about utube??? This isn’t Gearslutz and we don’t need a YT video to know its a bug so why would anyone waste their time with that?

[daeron80]

Originally Posted by wwittman

What's an example of a reason you'd want anything 20 db over full scale??

I've done it to test the validity of the assumptions about floating-point mixing. They're correct when the tool is functioning as advertised. But, yeah, other than that, it would just be a comfort that a freak accident isn't going to ruin hours of work.