How to do USB audio

Couldn’t resist having a quick delve before quitting. Noticed that one behaviour of the UCA202 seems to be that the first time you try to use it after installing the modules it returns rubbish. Have to unplug it and plug it in again. That may have muddled my initial tests. Maybe that was why it didn’t work and I blamed my program then proceeded to alter it.

Also realised that the EnumerateRates listing is bigger than I thought because I’m currently only reading though the first set of ‘play’ results and I’ll have to see if the ‘record’ ones follow in the rest of the buffer. More tomorrow…

Jim

Well, I may have done something wrong, however… As soon as I got the UCA202 I tried that, and it didn’t work. I had to explictly give the endpoint and alternative values when making the call.

There is no way to pass the endpoint and alternate settings to openOut or OpenIn. I have no idea what you are doing.

Have you tried reading the documentation?

Raik.

Thanks for that – it pointed me in the right direction. I finally got Angstrom linux working but couldn’t get the version you pointed to working. I think the main problem has been the SD cards I’ve been using I finally got a 2gb image of angstrom working by using the 2gb card out of my satnav.

At the moment the test sine wave works where I configured sound – don’t know if it’s the right place :-) but I can’t get music to play in gnome player. My high res DAC is undetected too so it may not be useful even when I get it working :-(

If those that are modifying the video system to auto detect the screen mode. RISC OS defaulted to 640 × 480 on my HD Ready TV – can’t change it easily so I’ve left it, first thoughts not very good. But it’s better than this linux which picks a screen mode which is 40mm to the left and 10mm too big at the top. I can’t figure out easily how to fix that either. But at least with RISC OS I’m not being asked to fill in text fields with the field names off screen :-(.

Angström: It was a old link from my bookmarks ;-)
Have you try in terminal

dmesg (To look at your DAC is detected?)

and

alsamixer (To activate.)

It works with the Pandora Angström. The Beagleboard one I never really use. If I find the time, I will compare.

A couple of things I noted from this site

http://www.diyaudio.com/forums/digital-source/188902-xmos-based-asynchronous-usb-i2s-interface.html

They seem to be getting a serial number from the device which would perhaps identify it from other gear that is plugged in.
Also note that high data rates consume nearly 500mA which could create problems unless external power is used in some form.
This device has a means of reflashing it’s firmware, I didn’t spot the way this is done
These and other Audio2 devices describe there connection as asyncronous, but I guess it is asynchronous via syn isocronous rather than having anything to do with the asynchronous USB mode.
There is synchronous audio also to confuse the issue.

Jim, it sounds like the UA202 is a fully working option for people happy with 96K data rates?

Thanks, Ron

The Beagleboard and the Pandora Angström are not really comparable. The Pandora-Angström much better and more stable.
Therefore, I “used” on the BeagleBoard OpenSuse . But “use” is too much to say here. Tried maybe.

Jim, it sounds like the UA202 is a fully working option for people happy with 96K data rates?

The UCA202 is only 32k/44.1k/48k 16 bit for play, and only rates up to 48k 16 bit for record. And at present it may not work nicely at rates other than 44.1k due to the mysterious ‘crackles’ <ahem> Colin will doubtless fix soon. 8-]

Their UMC202, etc, ranges do 96k/24 and are class 2. So seem much more promising. However I haven’t yet managed to get any response from Behringer wrt some questions I have about the range. Also they don’t use domestic phonos for inputs to ADC but a mix of XLR and TRS jacks. Fine for semi-pro home music recording ‘studios’ but no the norm for domestic hifi. Also their line levels are based on studio specs which can be an awkward match. So I’m trying to find out more or find a dealer who can help. Otherwise I guess I’ll end up buying one ‘blind’ and then play with it.

WRT power drain. I’m always using the UCA202 via an external hub with its own mains power to ensure I’m not dragging current out of the ARMiniX. But as yet I have no comment to make on anything like mains ripple, etc, this may add to the audio. Look at that once the more basic problems are sorted! :-)

Jim

Have you tried reading the documentation?

Yes. Have been. That doesn’t guarantee I always understand it though! Nor that reality always seems to have read the documentation, either. :-)

I think the problem I had was that the UCA202 seems happy to crash and then refuse to communicate if, say, I have it plugged in when I load Colin + your modules. So I was finding it would not play and thinking my program was wrong. That then lead me to try various changes until it started working. But I need to spend some more time on this later today. Have something else to do this morning, but I’ll check this afternoon.

Jim

Thanks. That looks more up to date but how did you copy it to an SD card?

I’ve been using win32DiskImager an my laptop. That worked for .img files. The image you pointed out is a .raw file. I changed that to a .img file and used win32DiskImager. Looking at the partitions on the disc the card now has 2 partitions as expected but the first one isn’t fat32 formatted.

Do you know if I need to do something else?

Colin, if you want some info on how to sort out ALSA on a Linux box then

http://www.audiomisc.co.uk/Linux/ALSA/NoMoreSilence.html

and

http://www.audiomisc.co.uk/Linux/ALSA/ALSAforUsers.html

may help. Note you can use the aplay -l command to list what is connected and to play test wave files. It then plays ‘direct’ so you don’t need to worry if a player program is working and setup.

Jim

Rename the raw-image to img shouldt work.
In WIN you can also use USBit to write direct from/in a archive.

USBit gave the same result as win32DiscImage 2 partitions first (small one) not FAT formatted. Do you know if the first one has to be FAT formatted I would have thought so.

In WIN you can look in the RAW e.g. with 7zip (right mousebutton and open). Both Partitions are from Linux (ext?). The ready card is working. USBit has a small bug (sorry I forgot) the Statusbar is out but the card is not ready. Click at the USBit Window and the statusbar came back. Not sure if this works with the opensuse Image. Normally I use DD with Linux e.g.

gunzip -c /Path/…/Image_Name.img.gz | sudo dd of=/dev/mmcblk0

or

dd if=/drivename/…/Image_Name.raw of=/dev/mmcblk0

Hmmm… it shows that it has 2 linux partitions. The version of angstrom I can get to work has a FAT partition first. You must be booting from a linux partition which doesn’t seem to work here.

Not sure. I can try but it is weekend :-(
I found this todo but I think I have not do so, I can’t remember ;-) It is long time ago and I use a older Version but also with two ext-Partitions.

I found that page thanks. It looks like I’ll need to get the serial connection to work – such fun :-)

Does USBAudio_EnumerateRates give output for both play and record when the same device does both of these? I’m currently either getting odd results because my ‘probe’ program is buggy (likely!) or because it does this in a way not explained. Example below.

The UCA202 has many play and many record settings and rates. If I run my probe program at present it gives the following:

 USB Audio Probe 0.5
 ===================

 Scans USB to find and report on the details of Audio devices

 Using USBAudio_EnumerateDevices to find out what is connected
 -------------------------------------------------------------

 List of devices by ID = USB17,USB13,USB14

 Device 1 => USB17
 Device 2 => USB13
 Device 3 => USB14
 Devices counted = 3
 

 Now examine each device in detail
 =================================

 Device  1 details
 ------------------

ID [USB17]
Maker  = Burr-Brown from TI              
Device = USB Audio CODEC 
 Number of resolutions available =  6 
 play using  16 bits per sample using  2 bytes per subframe
 play using   8 bits per sample using  1 bytes per subframe
  rec using  16 bits per sample using  2 bytes per subframe
  rec using   8 bits per sample using  1 bytes per subframe
Rates buffer size used/needed = 336
 play  ch = 2  res = 16  bytes/ch = 2  lpcm  number of rates = 3
 rates are:    32000    44100    48000  samples/sec
offsets 20 bytes = 5 ints
 Next 0 1 16
 Next 2 1 0


 Device  2 details
 ------------------

ID [USB13]
Maker  = Halide Design
Device = SPDIF Bridge
 Number of resolutions available =  1 
 play using  24 bits per sample using  3 bytes per subframe
Rates buffer size used/needed = 24
 play  ch = 2  res = 24  bytes/ch = 3  lpcm  number of rates = 4
 rates are:    44100    48000    88200    96000  samples/sec
offsets 24 bytes = 6 ints

[snipped the rest here]

I get the results for the play rates but it doesn’t list the record settings. Which should include a whole series of rates like 8000, 24000, etc, as well as the usual suspects. I’d thought that these might follow on from the ‘play’ values but start with an x80, as the buffer size used is large (336 bytes) and only a few are used for the ‘play’ details above. But the format seems not to be as I expected. I get a zero next which implies another separate list of ‘play’ values. I’ll try parsing that using the recipy, but it is a surprise to me as the details for the SWI don’t say this may occur.

The ‘next’ values show the 6 bytes after the nominal end of the ‘play’ results as integer values. You can see the zero where I guessed a 128 should appear.

The above was with three devices connected, but I’ve snipped the above once the results for the second (play only) have been given.

My guess now is that I’ll get ‘play’ results for 1ch and then for 8bit, etc, so there may me a series of such ‘play’ and ‘rec’ sections in whatever order the device maker decided. Does that sound right? If so the API should probably warn about this. The implication is that you have to keep parsing until you reach the end of the buffer size returned to catch any more sections. (Following on from that I guess that there may be sections for some devices which are neither ‘play’ or ‘record’ but something else if the maker dreams up something! :-) ]

Jim

Yes, if I loop the parsing it does list all the play and rec details OK. Afraid the API doc didn’t make this clear to me, but makes sense.

FWIW I don’t have any devices that give a range of sample rates. All mine list specific sets of values. So I’ve not (yet) been able to examine that.

Jim

Jim, I just noticed that !USBDescripters has the C source included, and the program appears to be broad in it’s capabilities. I’m only a part timer, so I would start at the main() (inside main.c) and trace the operation adding comments until I found out whats doing what, sometimes adding a printf() here and there acts like a TRACE, but this program is printing regularly anyway.

One thing with the devices such as UCA202, as it uses an all in one chip, (combined USB interface and DAC) I’d expect identical usb functionality from anything that uses this chip wether it is Behringer, Muse or a DIY kit.
The Sept EPE pcm2902 project has a switch so that a line in can bypass the op-amp stages for the XLR mic inputs.
The article describes four main USB endpoints.
1 A control end-point which receives control commands from the PC Host
2 An HID endpoint which allows inputs to the chip to generate keypress events on the host PC, to control muting, volume etc.
3 An isochronous IN endpoint (recording)
4 An isochronous OUT endpoint (playback)

(2) could be clearer, I’m assuming that it reports the state of controls, that can be set by using (1) or perhaps (1) is exclusively for setting rates, I dont know. HID isn’t needed for basic operation anyway.

EDIT: one difference I can see with this chip is that it uses a technique to
use the incoming isochronous (1ms?) frame rate to lock to the internal 98MHz clock. This might be causing complications.
It is called SpAct by TI.
Why it would work at 44100 and not others I dont know, maybe 1ms framerate
is being maintained better by the RISC OS driver at that rate?
Also, a cure for some OHCI isochronous problems here:

http://www.diyaudio.com/forums/digital-line-level/231138-asynchronous-usb-dac-clock-source.html

@Colin
In the night I have try my Beagleboard Linux card. OpenSuse failed with a discerror but Im sure I never use a serial connection. The Angstrom demo works but no alsamixer aviable. The Angstrom Genome also works. I remember I use the bootfiles from the Demo. There is a alsamixer and my USB-audiodevice is aviable. But no player inside. My networkadapter is working so I will try to installiert a player via opkg.

There is a alsamixer and my USB-audiodevice is aviable. But no player inside.

Does aplay let you play a 48k/16 or 44.1k/16 lpcm wave file? If so, that’s probably enough for simple tests as it will allow playback and let you assess the system modules and kernel. Having a desktop player is useful in normal use, but can add a layer of things that might not work out-of-the-box, so confuse the issue.

Jim

Jim, I just noticed that !USBDescripters has the C source included, and the program appears to be broad in it’s capabilities.

I have been looking at other programs to some extent. But my main aim in writing programs at present is to try and ensure I know how to use the swis, etc, so I can explain them in due course to others. I find that easiest by writing my own code and making my own idiotic mistakes. Helps me spot where things have been taken for granted. Slow, and can mean I’m a pest who keeps asking ‘daft’ questions. But I find it works best for me in the end.

Jim

Raik.

It’s looking like a waste of time.

I got Angstom Gnome working as you suggested and after it autologged in as root was presented with a black blank screen with a grey bar across the bottom with nothing on it. But that wasn’t going to stop me :-) my screen has the top and left missing :-) so I clicked around outside the top of the screen until I discovered a bar which eventually I was able to move to the left – yippee! I could see something. It’s surprising how many windows open at the top left. Can’t find an xterm yet but I run some program and it said xterm not available.

I’m stuck with the offscreen display the monitor configuration gives me no display options.

Also I can’t right click on the desktop without creating a new folder ie the first item on the menu is selected before I can do anything.

I plough on.

Jim. :-)

Your online tutorials failed at the first hurdle with the words ‘type in’. The question was where and how do I get to where. On the linux version I tried yesterday I eventually found xterm and could run alsamixer – an older version that you showed – but no aplay

@Colin
Weekend, only a smartphone :-(. I hope I remember …
Open your created folder and search via Strg (Ctrl) + F …
Alsamixer you find in /usr/bin
Terminal you find in usr/share/applications
Displaysettings you can find via Display or Monitor search …

Your online tutorials failed at the first hurdle with the words ‘type in’. The question was where and how do I get to where. On the linux version I tried yesterday I eventually found xterm and could run alsamixer – an older version that you showed – but no aplay

Sorry. Classic case of taking things for granted in that I didn’t think to explain about terminals. If you can find alsamixer then the ALSA system is nominally installed. aplay should be fundamental to that.

To check: Open an xterm and type in man aplay and press return. In any ‘normal’ linux install that should start listing the manual for the aplay command. (Press the space bar for the next screen worth. And press ‘q’ to quit the manual.)

BTW if you find the xterm window text too small, you can change things. e.g. by using xterm -fn 10x20

I can’t tell if the Linux you’ve installed is worth your experimenting with as I have no idea how much it may have been hacked down or altered. But if you have alsamixer I’d expect aplay to be present. If it isn’t then I suggest you give up as it may be harder for you to make sense of the hardware using that Linux distro than using RO! A linux version that lacks aplay is probably hacked to the point of being crippled so far as audio is concerned.

If you want to persist, remind me of the Linux distro name and where you got it. I’ll see if I can find out some more details. But TBH it seems like this will be a waste of time and will simply divert effort.

Jim