does anyone have experience with the suitable settings of the ADAV801 ALC Control registers (recovery mode, attack time, release time etc) for a acceptable speech regulation ?
Sorry this question got lost in the shuffle. I imagine at this point you have worked out some suitable settings?
The settings will depend a little on what the application is. For instance, a microphone for a podium used for live speaking would have different settings then audio for a radio transmitter with a cell phone sort of being in between these two extremes. So let me know if you are still working on this problem and more detail about your application. Thanks,
thanks for the reply. The issue is still open. The audio source is a helicopter pilote voice in a noisy environment. Normally a strong signal, but the audio level is not well defined. The problem is also caused by the restricted count of trials when optimizing the settings.
Which recovery mode are you using? If the signal is typically strong maybe you want to us limited recovery mode. I would experiment with the gain counter settings in ALC REG 1 to see how long it would typically take to adapt the signal level. You may want to keep the Recovery Threshold and Attack Threshold farther apart to reduce over modulating the PGA gain.
What behavior are you observing?
at the moment the choosen recovery mode is normal recovery, the attack thresold is set to 0 dB and the recovery threshold is set to -2 dB. The PGA gain register is set to 0x2C.
The test installation includes a Rohde & Schwarz audio test signal generator (for amplitude modulation) and a Rohde & Schwarz audio analyzer; there are still no real voice tests performed.
Although the audio source level is quite disciplined (helicopter pilote voice) the absolute level is spongy because the output level (for me the incoming level) is set by a wide range potentiometer fitted in a panel and which is tuned high-handed (estimated) in despite of a given level specification so it may be usefull to use the ALC function.
My major question concerns the setting of the time relevant parameters like attack timer selection, recovery time selection, attack threshold and recovery threshold to minimize any displeasing modulation effects.
I am still in two minds about the usage of the better recovery mode (there might be spurious transients or noise in the signal).
Retrieving data ...