Yes, 07 firmware: awx_20121207_Mac_8896.elf
Bad is two segudos of music, noise and then silence.
Hi, the last days I stayed in the frozen forest. Without computer and network.
So, I have a clean install of win7 64bit, fresh install of drivers AWSetup_20121004.exe and AB1.2 in UAC1 mode.
Foobar2k preferences for UAC1:
and widgetcontrol
I tested a lot of different flac files and always i have the same symptom: slight interruptions and sometimes metalic sounds.
This is my first UAC1 test from September 2012on win7. First time I see such a problem too.
In the case of 192 on linux:
EDIT: (192 in UAC2on linux)
I tested 30 files, all of them 192.
80 percent of them working badly. How, You must see this movie.
On windows7 in UAC2 mode all this files working corectly.
Tested with pulseaudio and without, always the same problem like on movie.
So, I have a clean install of win7 64bit, fresh install of drivers AWSetup_20121004.exe and AB1.2 in UAC1 mode.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Foobar2k preferences for UAC1:
An externally hosted image should be here but it was not working when we last tested it.
and widgetcontrol
An externally hosted image should be here but it was not working when we last tested it.
I tested a lot of different flac files and always i have the same symptom: slight interruptions and sometimes metalic sounds.
This is my first UAC1 test from September 2012on win7. First time I see such a problem too.
In the case of 192 on linux:
EDIT: (192 in UAC2on linux)
I tested 30 files, all of them 192.
80 percent of them working badly. How, You must see this movie.
On windows7 in UAC2 mode all this files working corectly.
Tested with pulseaudio and without, always the same problem like on movie.
Last edited:
i'm not a windows guy, so cant give win specific help. the UAC1 issue confuses me as ive never had issues with any UAC1 device.
are you using FooBar on both OS? try increasing your buffer length, you have it set to minimum according to your preferences snapshot, this can easily cause dropouts.
are you using FooBar on both OS? try increasing your buffer length, you have it set to minimum according to your preferences snapshot, this can easily cause dropouts.
Hi,
I use Foobar 2k only on windows seven and occasionally. I think the last time in September.
Every day on linux with XBMC player, with pulseaudio and now with alsa too, only in UAC2 mode. The problem exists on linux at 192 with UAC2 and now in windows7 64 on 44.1 with UAC1.
I do not know why is this problem exists on windows. In my free time, I try made fresh installation of windows7 on a separate disk. Very possible that the problem is elsewhere.
I have tested large and small buffer in foobar2000, there is no change.
I use Foobar 2k only on windows seven and occasionally. I think the last time in September.
Every day on linux with XBMC player, with pulseaudio and now with alsa too, only in UAC2 mode. The problem exists on linux at 192 with UAC2 and now in windows7 64 on 44.1 with UAC1.
I do not know why is this problem exists on windows. In my free time, I try made fresh installation of windows7 on a separate disk. Very possible that the problem is elsewhere.
I have tested large and small buffer in foobar2000, there is no change.
When you use pulse audio to play back 192k, you need to ensure that your pulse setting is ALSO 192k.
Search for the file daemon.conf (usually found in /etc/pulse. Google for pulse-daemon.conf). You should make a copy of this file into .pulse/daemon.conf and then edit it so that the sampling rate is 192000. I am at work so have no access to my PC at the moment. I will post my daemon.conf later.
Also you need to give real time priority to you audio software. I will also post the file later.
Alex
Search for the file daemon.conf (usually found in /etc/pulse. Google for pulse-daemon.conf). You should make a copy of this file into .pulse/daemon.conf and then edit it so that the sampling rate is 192000. I am at work so have no access to my PC at the moment. I will post my daemon.conf later.
Also you need to give real time priority to you audio software. I will also post the file later.
Alex
you need to set the default pulse to 32 bit. *For example, this is the setting in my .pulse/daemon.conf file (change default-sample-rate to what you use pulse for. *44100 for playback of CD music, 48000 for most movies, higher rates for HD music in FLAC or wav format etc.):
default-sample-format = s32le
default-sample-rate = 88200
default-sample-channels = 2
default-channel-map = front-left,front-right
default-sample-format = s32le
default-sample-rate = 88200
default-sample-channels = 2
default-channel-map = front-left,front-right
In UAC2 I've been trying various feedback strategies after I found a bug in earlier code. One thing I'd like to experiment with is how often the Host requests sample rate information from the Device. According to Nikolay that's part of the enumeration. Does anyone remember which setting this is?
The recent (20121206 and '07) .elf files use an older version of UAC1 feedback code, to test UAC1 stability primarily on Macs.
Zumbik, could you check if you can replicate the errors you are seeing on the 20121004 firmware? That's the latest "stable" version.
My hunch is that instabilities tend to happen on certain computers, perhaps where there is a larger than normal discrepancy in absolute frequency between available USB clock and the internal XOs in the DAC.
Børge
The recent (20121206 and '07) .elf files use an older version of UAC1 feedback code, to test UAC1 stability primarily on Macs.
Zumbik, could you check if you can replicate the errors you are seeing on the 20121004 firmware? That's the latest "stable" version.
My hunch is that instabilities tend to happen on certain computers, perhaps where there is a larger than normal discrepancy in absolute frequency between available USB clock and the internal XOs in the DAC.
Børge
Hi,
here #1952, I posted output of this command
It looks good to me, but not necessarily be a correct.
I added recommended entries for limits.conf, but there is no improvement.
Thanks Alex.
@Borge
I have exact same problem with audio_widget_20121004_AB-1.2.elf and awx_20121206.elf ...
I do not see any differences between them.
I suppose that the problem is a conflict drivers in win7.
I'll try install fresh copy of win7 on another computer in this weekend and I'll tell you if it works properly.
here #1952, I posted output of this command
Code:
watch -n 1 "cat /proc/asound/card1/stream0"I added recommended entries for limits.conf, but there is no improvement.
Thanks Alex.
@Borge
I have exact same problem with audio_widget_20121004_AB-1.2.elf and awx_20121206.elf ...
I do not see any differences between them.
I suppose that the problem is a conflict drivers in win7.
I'll try install fresh copy of win7 on another computer in this weekend and I'll tell you if it works properly.
hi.Zumbik,
Are u using the AW with another usb device sharing a hub or usb host controller?
what about 176.4k files?
I use Linux and have had no.problems with 192k files.
I have also uploaded firmware audio-widget-2012-11-08.elf which is from the audio-widget (not experimental)
which u can try.
Alex
Are u using the AW with another usb device sharing a hub or usb host controller?
what about 176.4k files?
I use Linux and have had no.problems with 192k files.
I have also uploaded firmware audio-widget-2012-11-08.elf which is from the audio-widget (not experimental)
which u can try.
Alex
what are the differences between this one and the latest (Børge's -12-06) -experimental one?
Alex, did you commit and push to audio-widget?
My latest firmware work has mainly been:
1 - Investigating UAC1 on Macs. I got access to a Mac yesterday, and it looks like the switch from 44.1 to 48 doesn't happen the right way. I'll look through the frequency select code and add debug messages before I get access to that computer again. I have tried two different feedback regulators in UAC1, and both are unstable on 48.
2 - Finding a bug in the UAC2 feedback regulator. The 20121004 firmware added a lot of stability on UAC2. It has a bug in the code which doesn't seem to influence performance. But I'd like to remove the bug regardless! To understand the code I've been toying with alternative regulators, and I must admit I have still not fully understood how the dual-buffer code works.
3 - As part of the UAC2 debug process there have been some reports of certain sample rates not working on Macs. One hunch I had was that the code replacing the Linux Quirk may cause instabilities on Macs. (Hence the difference between 20121206 and '07 firmware.) After yesterday I rather believe there may be a sample frequency change issue.
Demian wrote in a previous post that the frequency set-up protocol is different on Macs and other computers. When I have the time I'll try to translate that to working code.
Børge
My latest firmware work has mainly been:
1 - Investigating UAC1 on Macs. I got access to a Mac yesterday, and it looks like the switch from 44.1 to 48 doesn't happen the right way. I'll look through the frequency select code and add debug messages before I get access to that computer again. I have tried two different feedback regulators in UAC1, and both are unstable on 48.
2 - Finding a bug in the UAC2 feedback regulator. The 20121004 firmware added a lot of stability on UAC2. It has a bug in the code which doesn't seem to influence performance. But I'd like to remove the bug regardless! To understand the code I've been toying with alternative regulators, and I must admit I have still not fully understood how the dual-buffer code works.
3 - As part of the UAC2 debug process there have been some reports of certain sample rates not working on Macs. One hunch I had was that the code replacing the Linux Quirk may cause instabilities on Macs. (Hence the difference between 20121206 and '07 firmware.) After yesterday I rather believe there may be a sample frequency change issue.
Demian wrote in a previous post that the frequency set-up protocol is different on Macs and other computers. When I have the time I'll try to translate that to working code.
Børge