Any suggestion to support various rates?

[scari]

The driver support only 48k of sample rate.
Since I'm not kernel hacker, I couldn't do much try for it.
Just modifying fp->rates_min to 16000 didn't make it.
Also I tried not using CONTINUOUS and KNOT flag and specified each SNDRV_PCM_RATE_XXXX but no luck.

So could you let me point out where to start?

Anyway, your driver works well /w rpi-3.12.y kernel on raspberry pi. Thanks!

[prisme60]

I am interested with the compatibility with the raspberry pi.
Why do you prefer a lower frequence rate for audio? Perhaps raspberry pi has a too high CPU load? I am aware of an USB problem with raspberry pi, and also the low CPU capability. Do you use overclock?

1. It seems that you need to configure the Fushicai USBTV007 to 16KHz (mono or stereo). But the datasheet of the component is not public. The only way is to use the hardware with a windows software, set the audio to 16 bits, and with USB capture software to spy the setting of registers.

2. Another way (less efficient), is to drop some values (if you remove one value each two values, you will divide the frequence by 2 => 48KHz / 2 = 24KHz)

[scari]

I'm going to do some realtime audio signal processing with raspberry pi + Fushicai USBTV007.

48KHz of sample rate is too high to the processing on bare (700Mhz, no overclock) CPU.
I tried it on overclocked to 900MHz but there were no much improvement.
Though I can down sample it on userspace but I think it is inefficient.

I also aware of an USB problem too. but in normal case, your driver hasn't any noticeable stutter or some. Works really good to me.

I'll give a try to following your suggestion, and let you know what I find.