增加音频设备的输出增益

Question

I want to increase my virtual audio device gain. The Volume is turned all the way up but I can barely hear the audio. To do this I think we should look at the gain IOAudioLevelControl::createVolumeControl in IOAudioDevice. This is Soundflower code

 // Gain control for each channel
    control = IOAudioLevelControl::createVolumeControl(SoundflowerDevice::kGainMax,         // Initial value
                                                       0,                                   // min value
                                                       SoundflowerDevice::kGainMax,         // max value
                                                       0,                                   // min 0.0 in IOFixed
                                                       (40 << 16) + (32768),                // 72 in IOFixed (16.16)
                                                       channel,                             // kIOAudioControlChannelIDDefaultLeft,
                                                       channelNameMap[channel],             // kIOAudioControlChannelNameLeft,
                                                       channel,                             // control ID - driver-defined
                                                       kIOAudioControlUsageInput);
    addControl(control, (IOAudioControl::IntValueChangeHandler)gainChangeHandler);

It seems to me that I should change the 5th argument which does some black magic bit shifting but as I'm not sure how this works and how to change it. Thanks

Answer 1

The 4th and 5th parameters are minDB and maxDB, specified as in integer in 16.16 fixed point format. I think the dB here is meant to be dB relative to digital full scale - meaning that 0 dBFS is the max and the min is a negative value. At the very least, your parameters seem to be reversed and the sign of the positive value is wrong.

The black magic as you call it is specifying a value in the 16.16 fixed point format. See https://en.wikipedia.org/wiki/Q_(number_format) .

It is setting the maxDB to 40.5. The 40 coming from bitshifting the 40 up into the integer portion and the 0.5 coming from the 32768. A much more intuitive way of doing this is using the following method:

Multiply the floating point value by 2^n where n is the number of fractional bits (xx.16) and then round the result.

int fixedPtVal = round(40.5 * (1<<16)); 

Back to the issue of what to choose for minDB and maxDB. I couldn't really find any documentation on this but I think you would want to do something like:

minDB = round(-40.5 * (1<<16));
maxDB = 0;