I'm just trying to help Zumbik to get 192k working under Linux. As I said I have had no problems with 192k and 176.4k music at all and I have been using the audio-widget branch firmware.
(I do have problems on 88.2k (and may be other sampling rates, not tested) with the latest commits using the audio-widget-experimental branch though, EVEN AFTER reverting your removing of the linux fix in a local branch.)
So to remove another variable coming from all your changes in the experimental branch, I thought it would be good to let Zumbik try something not so experimental 🙂
(I do have problems on 88.2k (and may be other sampling rates, not tested) with the latest commits using the audio-widget-experimental branch though, EVEN AFTER reverting your removing of the linux fix in a local branch.)
So to remove another variable coming from all your changes in the experimental branch, I thought it would be good to let Zumbik try something not so experimental 🙂
I have found Zumbik's command:
watch -n 1 "cat /proc/asound/card1/stream0"
to be useful so I have just added it to:
watch-card1.sh
in the audio-widget branch 🙂
watch -n 1 "cat /proc/asound/card1/stream0"
to be useful so I have just added it to:
watch-card1.sh
in the audio-widget branch 🙂
So your build of audio-widget was without any code changes?
This is why we call the branch experimental 🙂 I've been toying with a feedback simulator I wrote in Octave and moved the code which did the best job there into the firmware.
Børge
This is why we call the branch experimental 🙂 I've been toying with a feedback simulator I wrote in Octave and moved the code which did the best job there into the firmware.
Børge
Hi all,
1) I can live without the 192k for linux ... yet 😉
2) I seriously wonder about construction of the RS232 interface on AB1.2 board. We have a lot os environments. Probably none of us are not aware what's exactly going on inside AB-1.2 when are having problems in the sound. If I understand correctly, rs232 logs should show where the problem is? How can I build environment for tests?
Borges, whether you make HOWTO?
Then my logs will be yours ... 🙂
1) I can live without the 192k for linux ... yet 😉
2) I seriously wonder about construction of the RS232 interface on AB1.2 board. We have a lot os environments. Probably none of us are not aware what's exactly going on inside AB-1.2 when are having problems in the sound. If I understand correctly, rs232 logs should show where the problem is? How can I build environment for tests?
Borges, whether you make HOWTO?
Then my logs will be yours ... 🙂
Hi Zumbik,
if you populate the RS232 amplifier, set your terminal to 57.6kbaud and plug it in, the debug port should come alive!
First of all you'll need a build environment. See Line 621 and onwards in the Readme file in my signature. Then you can start adding debug messages and building your own code.
I believe a single putchar is harmless because the UART is double buffered. But if you output character #2 before the 1st is sent, things start to take more time because the 1st one hasn't yet left the MCU.
Børge
if you populate the RS232 amplifier, set your terminal to 57.6kbaud and plug it in, the debug port should come alive!
First of all you'll need a build environment. See Line 621 and onwards in the Readme file in my signature. Then you can start adding debug messages and building your own code.
I believe a single putchar is harmless because the UART is double buffered. But if you output character #2 before the 1st is sent, things start to take more time because the 1st one hasn't yet left the MCU.
Børge
Ok, it looks like black magic, yet ... 😉
I understand that it makes sense and will give us the an explanation of trouble?
I understand that it makes sense and will give us the an explanation of trouble?
Before you launch into changing the code please make sure everything else is working right.
I have a collection of different sample rate files (most from Gimmel) tthat are stored as wave file I can play using aplay. This remove many. Moving parts from the chain. If you have pulseaudio you may need to shut it down to access alsa.
You need recent alsa or a recent Kernel for this to work OK. I am using 3.6.6 but I will migrate to 3.7 soon. 2.6.32 would be the oldest I think.
I have had no problems with any sample rate or transistion with 3.6.6 and mpd.
Sent from my myTouch_4G_Slide using Tapatalk 2
I have a collection of different sample rate files (most from Gimmel) tthat are stored as wave file I can play using aplay. This remove many. Moving parts from the chain. If you have pulseaudio you may need to shut it down to access alsa.
You need recent alsa or a recent Kernel for this to work OK. I am using 3.6.6 but I will migrate to 3.7 soon. 2.6.32 would be the oldest I think.
I have had no problems with any sample rate or transistion with 3.6.6 and mpd.
Sent from my myTouch_4G_Slide using Tapatalk 2
Here is where to get the Gimell test files: Gimell Records - Free Trial Downloads 1
I find them very useful tools.
I find them very useful tools.
no. The ALSA version included in Linux kernel 2.6.32 does not support UAC2 (unless an out-of-three, newer ALSA is used). The first "vanilla" kernel which should work is 2.6.38.
I have a headache ...
Is much better, but not perfect.
Now I have kernel 3.7 and alsa 1.0.25 without pulseaudio + audio-widget-2012-11-08.elf
Previously, the command:
give results for 192k:
Now: (kernel_3.7 and alsa_1.0.25)
"Feedback Format" is 7.17 was 15.17 (the same file) Why?
Now we have "Momentary freq." twice.
The first value is variable. I've seen range from 191700 to 192500.
The sound is correct if the first value is constant. Might be different from the 192k, but must be constant. If not, I have a metallic sound.
Summary:
AB1.2 works better with fw audio-widget-2012-11-08.elf for me.
I tested about 40 files in 192k. 90 percent of them is OK.
With fw awx_20121206.elf there is no change, there is often a metallic sound.
Later I'll check other files from Gimell records on linux and windows7. (different formats)
Cheers,
Zumbik
Is much better, but not perfect.
Now I have kernel 3.7 and alsa 1.0.25 without pulseaudio + audio-widget-2012-11-08.elf
Previously, the command:
Code:
watch-n 1 "cat / proc/asound/card1/stream0"
Code:
Every 1,0s: cat /proc/asound/card1/stream0
Fri Dec 7 22:51:31 2012
Audio-Widget QNKTC USB DAC AB-1.2 at usb-0000:00:12.2-1, high speed : USB Audio
Playback:
Status: Running
Interface = 2
Altset = 1
URBs = 2 [ 6 7 ]
Packet Size = 392
[COLOR="Red"]Momentary freq[/COLOR] = 191938 Hz (0x17.fe00)
Feedback Format = 15.17
Interface 2
Altset 1
Format: S32_LE
Channels: 2
Endpoint: 2 OUT (ASYNC)
Rates: 44100, 88200, 132300, 176400, 48000, 96000, 144000, 192000
Data packet interval: 250 us
Code:
Every 1,0s: cat /proc/asound/card0/stream0
Thu Dec 13 19:43:29 2012
Audio-Widget QNKTC USB DAC AB-1.2 at usb-0000:00:12.2-4, high speed : USB Audio
Playback:
Status: Running
Interface = 2
Altset = 1
Packet Size = 392
[COLOR="red"]Momentary freq[/COLOR] = 191977 Hz (0x17.ff40)
Feedback Format = 7.17
Packet Size = 0
[COLOR="red"]Momentary freq[/COLOR] = 192000 Hz (0x18.0000)
Interface 2
Altset 1
Format: S32_LE
Channels: 2
Endpoint: 2 OUT (ASYNC)
Rates: 44100, 88200, 132300, 176400, 48000, 96000, 144000, 192000
Data packet interval: 250 usNow we have "Momentary freq." twice.
The first value is variable. I've seen range from 191700 to 192500.
The sound is correct if the first value is constant. Might be different from the 192k, but must be constant. If not, I have a metallic sound.
Summary:
AB1.2 works better with fw audio-widget-2012-11-08.elf for me.
I tested about 40 files in 192k. 90 percent of them is OK.
With fw awx_20121206.elf there is no change, there is often a metallic sound.
Later I'll check other files from Gimell records on linux and windows7. (different formats)
Cheers,
Zumbik
I have MAX3221EEAE+ (SSOP-16).
In the diagram is laid MAX3221CUE+ (TSSOP-16).
If I understand good, that MAX3221EEAE+ will be good?
Sorry for such a question, but I'm worried about my AB1.2 🙁
Sent from my Note using Tapatalk 2
In the diagram is laid MAX3221CUE+ (TSSOP-16).
If I understand good, that MAX3221EEAE+ will be good?
Sorry for such a question, but I'm worried about my AB1.2 🙁
Sent from my Note using Tapatalk 2
If the chip fits mechanically on top of the footprint it's OK. Just mind the pin1 direction.
Better safe than sorry!
Børge
Better safe than sorry!
Børge
hi there
i noticed something interesting and thought it may help you (or not)...
if i set my player on linux with kernel 3.6.10 and alsa 1.0.26 to oversample via SRC to 96k or 192k i get no issues whatsoever which is kinda interesting to me...
i play a lot of 88.2k 96k 176.4k and 192k files so i probably get the most out of metallic sounds when it occurs 😀
i noticed something interesting and thought it may help you (or not)...
if i set my player on linux with kernel 3.6.10 and alsa 1.0.26 to oversample via SRC to 96k or 192k i get no issues whatsoever which is kinda interesting to me...
i play a lot of 88.2k 96k 176.4k and 192k files so i probably get the most out of metallic sounds when it occurs 😀
That is interesting! Which firmware are you using? This whole distributed control system design is tricky business....
I've decided to buy a used Mac (rather than spend my precious and limited coding time on a Hackingtosh or VM) to test. I hope to get it on line Monday night.
By the way, do you know of good offline SRC programs? I'm tired of sine waves, and want to test with music. I did find a converter which took over my brower settings _and_ did a shitty 44.1->48 :-(
Børge
I've decided to buy a used Mac (rather than spend my precious and limited coding time on a Hackingtosh or VM) to test. I hope to get it on line Monday night.
By the way, do you know of good offline SRC programs? I'm tired of sine waves, and want to test with music. I did find a converter which took over my brower settings _and_ did a shitty 44.1->48 :-(
Børge
i'm using awx_20121206 firmware
if you want to do it on windows you have a plugin for foobar foo_dsp_ssrc (if i remember right... will source it out again)
on linux you have DeaDBeeF and Aqualung both usually come with SRC resampler out of the box
if you want to do it on windows you have a plugin for foobar foo_dsp_ssrc (if i remember right... will source it out again)
on linux you have DeaDBeeF and Aqualung both usually come with SRC resampler out of the box
Hm, good to know the fw you're using!
I'm mainly looking for an offline SRC which gives out .wav.
Børge
I'm mainly looking for an offline SRC which gives out .wav.
Børge
you can do that in foobar if i rember right... you can set DSP chain for convert profile and it spits out FLAC, WAV etc
I would use Dbpoweramp for SRC. It may not be the best but its good and it is the best ripper converter available. Its not free but a modest investment for a lot of utility.
I still have not had an issue with different sample rates with the AB1.2. Lots with other hardware. All on Linux. I can switch between any combo instantly with nary a glitch or play long pieces with no issue.
I still have not had an issue with different sample rates with the AB1.2. Lots with other hardware. All on Linux. I can switch between any combo instantly with nary a glitch or play long pieces with no issue.
For offline the Izotope 64 bit SRC is about as good as it gets. I really wish it could be integrated into foobar for realtime playback, maybe one day.