Audio / Video in Hardy is complete joke

Bug #227067 reported by JamesRichardson
4
Affects Status Importance Assigned to Milestone
jack (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

Binary package hint: jack

Background:
 Use Case: HTPC to watch movies, record audio to mp3, run programs for the kids on the telly. (720p HDTV)
 Hardware: shuttle, High Definition Audio, dual core processor, 4Gb RAM, two sata drives, NVidia GForce 8500 GT.

Expected Outcome:
 - Can record and play back audio
 - Screen is visible (including menus)
 - DVD's play back.

Actual Outcome:
 - Audacity fails to record anything, playback is akin to 78rpm record
 - Screen has huge overscan, rendering top & bottom menus invisible.
- DVD's unwatchable due to pathetic sound.

Research:
 - Overscan just my bad luck - not fixable on linux.
 - Poor sound quality is due to new "feature" using "jackd".
However, jackd cannot play back audio unless:

a) it runs in real-time mode.
 - Real-time mode not possible unless jackd running as root
 - Real-time scheduling in default kernel not sufficient to schedule jackd as often as it needs ( jackd fails as rescheduling time too great)
 - nvidia driver with glx (non-free) not allegedly compatible with rt- kernel.
 - no coherent installation instructions for rt- kernel available.
b) periods/buffer can be increased (also increasing audio latency)
 - alsa support does not properly configure the driver, so setting jack for "Periods/Buffer:4" (in qjackctl) fails with
    "got smaller periods 2 than 3 for playback"

do you really think that normal people can put up with all this.
i only want to watch some video on a pc under my telly.
linux sure has a long way to go if watching a dvd and recording the radio is beyond its ability.

Revision history for this message
Richard Seguin (sectech) wrote :

Thank you for taking the time to report this bug and helping to make Ubuntu better. Please answer these questions:

* Is this reproducible?
* If so, what specific steps should we take to recreate this bug?

This will help us to find and resolve the problem.

Revision history for this message
Richard Seguin (sectech) wrote :

* Marking as incomplete pending requested information...

Changed in jack:
status: New → Incomplete
Revision history for this message
JamesRichardson (james-time4tea) wrote :

- Yes this is reproducible all the time, on my hardware (as described above)

 - To reproduce:

  - Install complete "Hardy" from CD.
  - Apply all updates.
THEN
  - Play audio file (.aup) using "jack" in audacity
OR
 - Play any move file (like a .ogg) in gmplayer

The audio will be extremely poor and suffer from clicks pops and dropouts.

jackd will complain a lot about xruns ( == awful audio)

Nothing better to make an upset user happy than post a standard response and mark as "works for me". thanks!

Revision history for this message
JamesRichardson (james-time4tea) wrote :

marking as "in progress" and reassigning now more information available.

Changed in jack:
assignee: nobody → richseguin
status: Incomplete → In Progress
Revision history for this message
Richard Seguin (sectech) wrote :

Assigning bugs to a particular person is generally not advisable in order to get your problem looked at quickly... My role is to get the new bugs triaged and assigned to the proper group quickly (that way they don't remain open and without attention for a long period). I am removing the assignment and I will continue to help you with your issue so we can find the correct team and to provide them with enough information.

Changed in jack:
assignee: richseguin → nobody
Revision history for this message
JamesRichardson (james-time4tea) wrote : Re: [Bug 227067] Re: Audio / Video in Hardy is complete joke

Richard Seguin wrote:
> Assigning bugs to a particular person is generally not advisable in
> order to get your problem looked at quickly... My role is to get the
> new bugs triaged and assigned to the proper group quickly (that way they
> don't remain open and without attention for a long period). I am
> removing the assignment and I will continue to help you with your issue
> so we can find the correct team and to provide them with enough
> information.
>
>
> ** Changed in: jack (Ubuntu)
> Assignee: Richard Seguin (richseguin) => (unassigned)
>
>
ok fair enough. looking on the web - seems like a whole bunch of people
using jack are having the same problem as me.

i don't belive its a jack specific problem - seems like a mixture of
kernel conf, hardware and the new addition of jack.

cheers

james

Daniel T Chen (crimsun)
Changed in jack:
assignee: nobody → crimsun
status: In Progress → Incomplete
Richard Seguin (sectech)
Changed in jack:
assignee: crimsun → nobody
Revision history for this message
JamesRichardson (james-time4tea) wrote :

ok - i had a chance to do a little investigation. Seems that its the stock jackd thats the problem... its just not up to the job.

Here's what I did.

1) installed rt- kernel
2) my soundcard runs as INT-22, so I did chrt -p 99 `pidof "INT-22"`
3) enabled memlock and rtprio in /etc/security/limits.conf
   @audio soft rtprio 99
   @audio hard rtprio 99
   @audio - memlock 512000
   @audio - nice -19
4) used qjackctl to enable realtime and scheduling priority of 80, as recommended

ok - jack now realtimed, as is the soundcard.

now to get a test program...

i installed libjack-dev and libasound-dev

i then got the source distribution for mpg123 v1.4.2 (from sourceforge)

./configure --prefix=/home/richja/mpg123
(configured oss, alsa, jack)
make
make install

now i have mpg123 ( a very performant and lightweight mp3 player) configured for alsa and jack

so NOT running jack

% mpg123 xxx.mp3
 - picks alsa driver - perfect sound - virtually 0 CPU.

running jack as above

% mpg123 xxx.mp3
 - picks jack driver - instantly audible audio drops, jack messages of: "alsa_pcm: xrun of at least 0.007 msecs", delay of 21329.000 usecs exceeds estimated spare time of 10626.000; restart, also clear audio artifacts. totally unusable.

hope this gives more info.

Revision history for this message
JamesRichardson (james-time4tea) wrote :

oh yeah - once jack is running, the output sample rate of the software has to match the configured jack sample rate or it wont connect.

actually (while running jack) the command:

% mpg123 xxx.mp3
gives:
unable to set up output format (the source is 44100Hz, but the capability matrix shows only 48000Hz, S16)

so you have to run

% mpg123 -r 48000 xxx.mp3

which then exhibits the mentioned audio artifacts.

notably though, the alsa capability matrix shows all (from 8000Hz to 48000Hz, s16, u16, u8,s8,ulaw and alaw)

i found the same thing in audacity, unless the project sample rate == the jack rate the whole thing fails miserably: you don't even get to hear your audio sounding cr*p.

Revision history for this message
Richard Seguin (sectech) wrote :

We are closing this bug report because it lacks the information we need to investigate the problem, as described in the previous comments. Please reopen it if you can give us the missing information, and don't hesitate to submit bug reports in the future. To reopen the bug report you can click on the current status, under the Status column, and change the Status back to "New". Thanks again!

Changed in jack:
status: Incomplete → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.