uaudio - distorted output

[Alexander Leidinger]

Quoting Gary Jennejohn <gljennjohn at gmail.com> (from Wed, 11 Oct 2017  
06:07:34 +0200):

> On Tue, 10 Oct 2017 20:14:37 +0200
> Alexander Leidinger <Alexander at leidinger.net> wrote:
>
>> Quoting Hans Petter Selasky <hps at selasky.org> (from Mon, 9 Oct 2017  
>> 22:38:32 +0200):
>>
>> > On 10/09/17 21:47, Alexander Leidinger wrote:
>> >>> If you are using multi channel audio equipment with 24-bits,  
>> try >>> to avoid FULL-speed ones!
>>
>> I have now tried every single USB connector, external and internal.  
>> None of them attaches to the EHCI bus. :( The only connector which  
>> Iwas not able to test is labeled "USB SSD" and has a smaller  
>> pin-gridthan those other internal usb-PIN-connectors with cables to  
>> the frontof the case.
>>
>> I think I will now buy a PCIe USB3.0 card. Anything I should avoid?  
>> Anything you recommend? I found this:
>>      
>> www.amazon.de/CSL-Express-Controller-SATA-Stromanschluss-Schnittstellenkart
>> e/dp/B00OBACW0G
>> which seems to have some VLI chip on it. I haven't found anything  
>> usable in terms of vendor/product ID or such.
>> 4 ports may be enough, for an external USB backup HD, the  
>> soundcard, and a video device, but for ~10 EUR difference...
>>
>
> I've been using the CSL B00F9XGPTI XHCI 4-port card which works.  I
> bought it on amazon.de, but I don't know whether it's still available.

I bought the one I had listed. Works. I did a short Amazon review  
which mentions FreeBSD.

>> >> Well... first I want to get the 2 channel 16bit case working...  
>> >> then I will have a look at extending this to 5.1 (most probably  
>> >> 16bit, that's enough to watch action movies).
>> >
>> > Hi,
>> >
>> > Some more ideas:
>> >
>> > 2 channel 16 bit will only work if you turn off bitperfect. The
>>
>> I activated bitperfect as a test because I got the distortion and  
>> was hoping to either gain some lower-latency or at least get rid of  
>> somelayers of code to rule out issues there.
>>
>> > lowest number of channels the device exports is 4 for playback.  
>> That> means the kernel will re-format 2 channels to 4 channels  
>> always and> that is done by the PCM feeder. You'll need to check  
>> that there is a> filter which handle that.
>> >
>> > You might want to install virtual_oss from ports to handle this >  
>> device properly.
>>
>> In the end I want 5.1 output, so if the bus speed doesn't give me  
>> this: no need to waste our time with the slow ports.

It looks like this now (here the second uaudio device is attached):
---snip---
ugen8.1: <0x1106 XHCI root HUB> at usbus8, cfg=0 md=HOST spd=SUPER  
(5.0Gbps) pwr=SAVE (0mA)
ugen8.2: <vendor 0x2109 USB2.0 Hub> at usbus8, cfg=0 md=HOST spd=HIGH  
(480Mbps) pwr=SAVE (100mA)
ugen8.3: <VIA Labs, Inc. USB2.0 Hub> at usbus8, cfg=0 md=HOST spd=HIGH  
(480Mbps) pwr=SAVE (0mA)
ugen8.4: <vendor 0x0d8c USB Sound Device> at usbus8, cfg=0 md=HOST  
spd=FULL (12Mbps) pwr=ON (500mA)
ugen8.5: <VIA Labs, Inc. USB3.0 Hub> at usbus8, cfg=0 md=HOST  
spd=SUPER (5.0Gbps) pwr=SAVE (0mA)
---snip---

What I don't understand is it tells spd=FULL, so the device is not  
doing faster as it theoretically can do on the mainboard-usb device.  
Why does it work better (no audio distortion when connected to the  
xhci device)?

I also tried 5.1 output (data goes via optical connection to the pre-amp/DAC).
dev.pcm.2.bitperfect: 0
dev.pcm.2.buffersize: 0
dev.pcm.2.rec.vchanformat: s16le:2.0
dev.pcm.2.rec.vchanrate: 48000
dev.pcm.2.rec.vchanmode: fixed
dev.pcm.2.rec.vchans: 1
dev.pcm.2.play.vchanformat: s16le:5.1
dev.pcm.2.play.vchanrate: 48000
dev.pcm.2.play.vchanmode: passthrough
dev.pcm.2.play.vchans: 1

Audio comes out of all 5 speakers, but somehow I have the impression I  
get the same output everywhere... I need to download a 5.1 test and  
see. But this is another issue now.

Thanks!
Alexander.

-- 
http://www.Leidinger.net Alexander at Leidinger.net: PGP 0x8F31830F9F2772BF
http://www.FreeBSD.org    netchild at FreeBSD.org  : PGP 0x8F31830F9F2772BF
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: Digitale PGP-Signatur
URL: <http://lists.freebsd.org/pipermail/freebsd-usb/attachments/20171014/b57089af/attachment.sig>