Another plea for crackle help

classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

Another plea for crackle help

Matt Savigear-9
Hi all,

I'm still getting crackle on playback on my Tascam US-122 and I have
tried some more stuff. Any help would be much appreciated as this one is
driving me nuts. Here is the info:

* I am using jack 0.103.0
* I get severe crackle, but no reported xruns, on jack playback. It
   seems that recording via jack is fine, because I can export audio from
   ardour and play via another setup (e.g. straight ALSA) and the sound
   is perfect.
* The crackle is present across various jack-based applications.
* I have tried a broad range of jack settings from low to high latency
   and have also tried both the standard and rawusb devices for the
   US-122. I'm not 100% certain of this, but I think the crackle might
   well be reduced if I force jack to 16 bit.
* The interface is fine as it plays back correctly using ALSA.
* Jack is at least partially OK as it works fine with the other card
   in my machine under emu10k1.
* The machine's USB is fine as up until recently it was successfully
   using my Tascam US-428 (plus the ALSA playback works as noted above).
* I have tried a standard low-latency preemptible gentoo kernel and the
   full realtime kernel (rt-sources), both give the same results.
* I have tried various window managers, including my usual Fluxbox.
* Other than being an x86_64 box this machine is set up a closely as
   possible to my other (32-bit) machines, both of which have used the
   US-122 successfully.

Help preserve my sanity! Any input would be most appreciated.

Regards,

Matt.


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: Another plea for crackle help

Marc-Olivier Barre-2
On 8/5/07, Matt Savigear <[hidden email]> wrote:
> * Other than being an x86_64 box this machine is set up a closely as
>    possible to my other (32-bit) machines, both of which have used the
>    US-122 successfully.
>
> Help preserve my sanity! Any input would be most appreciated.
>
> Regards,
>
> Matt.

Hey Matt,

Sounds strange if you manage to get that device working all right on
other similar setups...

What command line do you use to lauch jack ? can you post a small
sample of what you hear on some website for us to download ?
__________________
Marc-Olivier Barre.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: Another plea for crackle help

Quentin Harley
In reply to this post by Matt Savigear-9
Matt Savigear wrote:
> Hi all,
>
> I'm still getting crackle on playback on my Tascam US-122 and I have
> tried some more stuff. Any help would be much appreciated as this one is
> driving me nuts.

Hi Matt,

Most often than not crackle on a digital system is only related to
clipping, or overflow distortion.  Especially if you are not getting any
xruns, it means that the enjin is working fine, but the actual signal is
distorted.  Check your mixer levels again.  I had this on my laptop
where the software would try to drive my outputs way too hot, and
clipping is the natural result.  You need the get your maximum output in
the -1dB to -3dB range.

Ok, so I am telling you something you already know.  Try it though, is
does not always make sense...

Cheers,
Quentin

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: Another plea for crackle help

Matt Savigear-9
In reply to this post by Marc-Olivier Barre-2
Hi Chaps,

Quentin Harley wrote:
 > Check your mixer levels again.  I had this on my laptop
 > where the software would try to drive my outputs way too hot, and
 > clipping is the natural result.  You need the get your maximum output
 > in the -1dB to -3dB range.

I've now been through attenuating the output from the applications which
allow it (ardour and hydrogen specifically) but it doesn't seem to help.
What is apparent, and I don't know if this is relevant, is that I only
get crackle when samples are being played - I had a relatively sparse
drum track and the distortion only occurred in the sample-filled parts
rather than the spaces.

This would lend credence to the clipping theory, except as noted even if
I bring the output right down (by track, drum sample or master output)
it still happens. I wonder if I upgraded some kind of sound library and
picked up a bug. This doesn't seem that likely as I can use audacious
under ALSA and jack, presumably using the same sample decoding path, and
only get distortion under jack.

Marc-Olivier Barre wrote:
> What command line do you use to lauch jack ? can you post a small
> sample of what you hear on some website for us to download ?

Here is the current command line, though I have tried many variations,
plus the relevant part of the following output:

jackd -R -P80 -u -dalsa -r44100 -p256 -n3 -D -Chw:1,2 -Phw:1,2 -i2 -o2

loading driver ..
apparent rate = 44100
creating alsa driver ...
hw:1,2|hw:1,2|256|3|44100|2|2|nomon|swmeter|-|32bit control device hw:1
ALSA/USX2Y: EXPERIMENTAL hwdep pcm device hw:1,2 (aka "rawusb")
configuring for 44100Hz, period = 256 frames, buffer = 3 periods
ALSA: final selected sample format for capture: 24bit little-endian
ALSA: use 3 periods for capture
ALSA: final selected sample format for playback: 24bit little-endian
ALSA: use 3 periods for playback

I'll need to do quite a lot of mucking about to get a recording of the
problem, but I'll try to get something up in the next couple of days.

--
Cheers,

Matt.

"The idea that an arbitrary naive human should be able to properly use a
given tool without training or understanding is even more wrong for
computing than it is for other tools (e.g. automobiles, airplanes, guns,
power saws)." - Doug Gwyn

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: Another plea for crackle help

Vince Werber
Hi Folks...

I had a distortion/crackle when I first stated out... it took me awhile to
figure it  out but the remedy for my situation was to drop the gain on
Qsynth...  When ever I use Qsynth I can't have the gain any hight than "20".

I don't suppose this will help you as you didn't mention Qsynth or
Fluidsynth...

What exactly is the line up that you are using BTW?

Have a good day and a better tomorrow!

vince in Maine




On Sunday 05 August 2007 10:40, Matt Savigear wrote:

> Hi Chaps,
>
> Quentin Harley wrote:
>  > Check your mixer levels again.  I had this on my laptop
>  > where the software would try to drive my outputs way too hot, and
>  > clipping is the natural result.  You need the get your maximum output
>  > in the -1dB to -3dB range.
>
> I've now been through attenuating the output from the applications which
> allow it (ardour and hydrogen specifically) but it doesn't seem to help.
> What is apparent, and I don't know if this is relevant, is that I only
> get crackle when samples are being played - I had a relatively sparse
> drum track and the distortion only occurred in the sample-filled parts
> rather than the spaces.
>
> This would lend credence to the clipping theory, except as noted even if
> I bring the output right down (by track, drum sample or master output)
> it still happens. I wonder if I upgraded some kind of sound library and
> picked up a bug. This doesn't seem that likely as I can use audacious
> under ALSA and jack, presumably using the same sample decoding path, and
> only get distortion under jack.
>
> Marc-Olivier Barre wrote:
> > What command line do you use to lauch jack ? can you post a small
> > sample of what you hear on some website for us to download ?
>
> Here is the current command line, though I have tried many variations,
> plus the relevant part of the following output:
>
> jackd -R -P80 -u -dalsa -r44100 -p256 -n3 -D -Chw:1,2 -Phw:1,2 -i2 -o2
>
> loading driver ..
> apparent rate = 44100
> creating alsa driver ...
> hw:1,2|hw:1,2|256|3|44100|2|2|nomon|swmeter|-|32bit control device hw:1
> ALSA/USX2Y: EXPERIMENTAL hwdep pcm device hw:1,2 (aka "rawusb")
> configuring for 44100Hz, period = 256 frames, buffer = 3 periods
> ALSA: final selected sample format for capture: 24bit little-endian
> ALSA: use 3 periods for capture
> ALSA: final selected sample format for playback: 24bit little-endian
> ALSA: use 3 periods for playback
>
> I'll need to do quite a lot of mucking about to get a recording of the
> problem, but I'll try to get something up in the next couple of days.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: Another plea for crackle help

Chuckk Hubbard
In reply to this post by Matt Savigear-9
I'm no expert.  I do have a suggestion, though: to try with just playback, i.e the -P flag after -dalsa instead of -D.  Hopefully that's not the solution, as it sucks not to have duplex operation, but it would be something to check.  I am able to have duplex if I use onboard sound for playback and my Zoom H4 for capture, which is how I prefer it anyway.
Another distant possibility is that the samples you're using are recorded with low sample rate, or at an odd sample rate, e.g. 16,000 when your playback is 44100.  I don't know if that would slow things down enough to cause distortion, but you could try manually converting the sr in Audacity to several different rates and try with each set of samples.  And I believe sample rate conversion is normally handled by one of the soundfile libraries, so it might not hurt to also try reinstalling them.
Did you compile any of the software you're using on your machine, or are they all binaries from your distro?

-Chuckk


On 8/5/07, Matt Savigear <[hidden email]> wrote:
Hi Chaps,

Quentin Harley wrote:
> Check your mixer levels again.  I had this on my laptop
> where the software would try to drive my outputs way too hot, and
> clipping is the natural result.  You need the get your maximum output
> in the -1dB to -3dB range.

I've now been through attenuating the output from the applications which
allow it (ardour and hydrogen specifically) but it doesn't seem to help.
What is apparent, and I don't know if this is relevant, is that I only
get crackle when samples are being played - I had a relatively sparse
drum track and the distortion only occurred in the sample-filled parts
rather than the spaces.

This would lend credence to the clipping theory, except as noted even if
I bring the output right down (by track, drum sample or master output)
it still happens. I wonder if I upgraded some kind of sound library and
picked up a bug. This doesn't seem that likely as I can use audacious
under ALSA and jack, presumably using the same sample decoding path, and
only get distortion under jack.

Marc-Olivier Barre wrote:
> What command line do you use to lauch jack ? can you post a small
> sample of what you hear on some website for us to download ?

Here is the current command line, though I have tried many variations,
plus the relevant part of the following output:

jackd -R -P80 -u -dalsa -r44100 -p256 -n3 -D -Chw:1,2 -Phw:1,2 -i2 -o2

loading driver ..
apparent rate = 44100
creating alsa driver ...
hw:1,2|hw:1,2|256|3|44100|2|2|nomon|swmeter|-|32bit control device hw:1
ALSA/USX2Y: EXPERIMENTAL hwdep pcm device hw:1,2 (aka "rawusb")
configuring for 44100Hz, period = 256 frames, buffer = 3 periods
ALSA: final selected sample format for capture: 24bit little-endian
ALSA: use 3 periods for capture
ALSA: final selected sample format for playback: 24bit little-endian
ALSA: use 3 periods for playback

I'll need to do quite a lot of mucking about to get a recording of the
problem, but I'll try to get something up in the next couple of days.

--
Cheers,

Matt.

"The idea that an arbitrary naive human should be able to properly use a
given tool without training or understanding is even more wrong for
computing than it is for other tools (e.g. automobiles, airplanes, guns,
power saws)." - Doug Gwyn

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel



--
http://www.badmuthahubbard.com
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: Another plea for crackle help

Matt Savigear-9
Chuckk Hubbard wrote:
> I'm no expert.  I do have a suggestion, though: to try with just playback,
> i.e the -P flag after -dalsa instead of -D.

I've tried this, but I can never get anything to connect properly to
jack if I start it in playback only. Audacious says it "couldn't open
audio device" and ardour silently crashes when I try to open any
project. This has the added annoying side effect of making jack unable
to terminate properly and leaving a [defunct] instance in my process table.

> Another distant possibility is that the samples you're using are recorded
> with low sample rate, or at an odd sample rate, e.g. 16,000 when your
> playback is 44100.

Definitely not - the ardour projects I am working with (or rather,
trying to work with) were recorded at 44.1k on my other machine.

> Did you compile any of the software you're using on your machine, or are
> they all binaries from your distro?

My distro is gentoo, so there are no binaries. Everything is compiled
from source.

--
Cheers,

Matt.

"I think I would say that, no matter what the other conditions, you can
pretty much ruin a relationship by tumble-drying your girlfriend's
baby." - Michael Hopcroft

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel
Reply | Threaded
Open this post in threaded view
|

Re: Another plea for crackle help

Matt Savigear-9
In reply to this post by Matt Savigear-9
Hi Karsten!

Karsten Wiese wrote:
> Nothing related in dmesg output I guess?

Nothing obvious that I can see.

> Anything else connected to the US-122's USB-port? (lsusb would show)

Here is my lsusb output. I have tried shuffling the connectors around,
but then I realised that the but ID isn't hardware so much as ehci vs.
ohci so it makes no difference where things are plugged in. I might try
without the scanner (the Canon one), the only ehci device, next time I
boot, though it didn't used to affect my 428.

Bus 002 Device 006: ID 1604:8007 Tascam US-122 Audio/Midi Interface
Bus 002 Device 005: ID 55aa:b012 OnSpec Electronic, Inc. Mitsumi FA402M
8-in-2 Card Reader
Bus 002 Device 004: ID 045e:001c Microsoft Corp.
Bus 002 Device 002: ID 045e:0040 Microsoft Corp. Wheel Mouse Optical
Bus 002 Device 001: ID 0000:0000
Bus 001 Device 001: ID 0000:0000
Bus 001 Device 003: ID 04a9:221c Canon, Inc.

> Does ALSA piping work fine also? Try:
> $ arecord -Dhw:x -fcd|aplay -Dhw:x -fcd

Yes, this seems to work nice and clear, with playback either through by
emu10k1 or the Tascam unit.

> Do the 32-bit machine's still exist running ok?
> If so, please compare the jackd and alsa-driver versions.
> If they differ: build/run the 32-bit version on the 64-bit box.
> (I mean the 64-bit executable with the working 32-bit-executables version number)

I'll check into this. I know the other machine with the 428 is running
the same versions as I am (mostly alsa 1.0.14 and jack 0.103) but my old
studio machine is running older versions. I'll try those (if I can still
get them) and see what happens. If I can't get them I'll bring the old
one up to date and see if it breaks.

--
Cheers,

Matt.

"There are people who don't like capitalism, and people who don't like
PCs. But there's no-one who likes the PC who doesn't like Microsoft" -
Bill Gates

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Jackit-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jackit-devel