Stereo Tool

Hello,
Did you think about inserting a speech detection in stereo tool?
It maybe useful if you then can turn off the Natual dynamics during speech or change the AGC or Multiband timing constants (Attack and release times multipliers).
The simple Algorithm (like in actual Optimods) is to set 2 conditions for the speech detection:
-Signal must be mono
-There is for max 1,5 seconds continous level
The disadvantage is here, that speech with background music will be recognized as music.
But i have heard from a more complicated Dolby- Algorithm for speech detection, maybe that this is free from this disadvantage.
Regards,
Robert Sack

Detecting Speech is small problem... thing is what it should do when detect it.

- Turn ND off for speech: Good idea
- Other things maybe as well (but a bit more complex especially to configure)
- That algorithm is insanely simple... What if you play mono music? (Some old song for example?) What if you have an interview with 2 microphones and it's fed to the processor in stereo? Both things might not be a problem in the real world - I'm not sure.

Hello Hans,
Yes you are right.
This simple algorithm (like in the Orbans) is not optimal.
Perhaps the Dolby- Algorithm is better, i don´t know...

Perhaps the detection could force the loading a part of a preset, so that you can store the speech settings (only some useful audio settings) as a "derivate" of a preset...

Mono songs would be recognized as music even with this simple algorithm (except acapella voices) because the continous level is longer then 1,5s in music.
I forgot to say that this status (both conditions fullfied) must be 2,5s minimum long.
You will recognize, hans, my english is not so very good, sorry for this, but i think you do understand what i mean...

Don't know about detection, but after detecting here's what i think it should do.
Since Speech is always hard to process becasue non-symetrical signal and can make comrpessors/limiters to work not properly, first thing to set Phase Rotator more agressive. Ofc PhR need to be before MB. Also bit faster release speed(s) in Multiband.

Bojcha, you are right.
It´s a good idea to activate (/increase the activity of) the Phase rotator.
It would make also sense to change the Bass clip shape to a softer position during speech.
As i wrote In my opinion also to change the MB Attack and release times makes sense as like to disable the Natural dynamics (ND doesn´t drastical decrease the speech quality, but it´s a bit more consistent if you switch of this).

How does the Omnia 9 do this? It looks very clean even with mono music.

The option to turn off (fade out?) the 19khz pilot with mono (voice?) source is an interesting idea too.
Think this is implemented in the O9? - It would be interesting to hear from those who have used this feature.