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

[Wolfgang Eller]

If a freak is doing an accident like this it's worth to ruin his work.

[Top Jimmy]

Another facet of this issue is that if you import a floating point file with audio above +20dBFS, you cannot use the Audiosuite Gain plugin to decrease the gain of the file. Using AS Gain results in a permanently distorted file.

[XJENSEN]

Yes, these floating point issues with AS gain are known to Avid. I have already reported them way back. Correct and consistent audio processing doesn't seem a priority for Pro Tools.

[midnightrambler]

Whilst I'm in agreement with the posts that say, "Why would you need to push the level that hard?", the fact that this is yet another aspect of the software that's broken for no apparent reason is really, really depressing. It doesn't inspire confidence that other critical things won't get broken down the line.

[seanmccoy]

For post projects I have on several occasions needed to add over 20 dB of gain to ultra-soft field recordings. So this is a potential issue on a practical level.

[Wolfgang Eller]

Originally Posted by seanmccoy

For post projects I have on several occasions needed to add over 20 dB of gain to ultra-soft field recordings.

An ultra-soft field recording is not leveled at 0dbfs.
The problem is if you gain a file 20db over 0dbfs.

[Rich Breen]

Originally Posted by seanmccoy

For post projects I have on several occasions needed to add over 20 dB of gain to ultra-soft field recordings. So this is a potential issue on a practical level.

Hi Sean - the issue is not adding 20dB or more of gain to something - that of course is fine - if it wasn't the world would spin off its axis. The issue occurs when gaining a clip up more than 20dB *over* full scale - the floating point engine should be able to recover signals *hundreds* of dB over "full scale" without clipping, but something weird is going on. Admittedly an edge-case, but still seems like an odd bug.

[seanmccoy]

Originally Posted by Rich Breen

Hi Sean - the issue is not adding 20dB or more of gain to something - that of course is fine - if it wasn't the world would spin off its axis. The issue occurs when gaining a clip up more than 20dB *over* full scale - the floating point engine should be able to recover signals *hundreds* of dB over "full scale" without clipping, but something weird is going on. Admittedly an edge-case, but still seems like an odd bug.

Hey, Rich, good to know. Now can somebody tell me when one might need 20dB over full scale? Talk baby talk.

[daeron80]

Apparently, some of us work only on audio we produced ourselves. But others of us receive wildly horrendous audio from clients, and sometimes have to finish jobs in a massive hurry due to circumstances beyond our control. The floating point engine is there (in part) to handle things when things get crazy. It's broken, and it needs to be fixed.

[Wolfgang Eller]

Originally Posted by daeron80

But others of us receive wildly horrendous audio from clients, and sometimes have to finish jobs in a massive hurry due to circumstances beyond our control.

But still, why is there any need to gain a file 20db above 0dbfs?
Any 32bit float file I get from a client I convert it by import into a 24bit fixed.
For me there is no reason for using 32 float.
I have learned to use gain structure.