[ardour-dev] Strange Issue WIth Automation in ardour-0.9_beta11

[David R. Bergstein]

Paul Davis wrote:
>>I recently ran into some problems whenever automation is enabled on any 
>>track with ardour-0.9_beta11. These were usually indicated as disktream 
>>errors in the ardour log window, and when checking the log generated 
> 
>>from jackd and ardour (this is run as an sudo session on a script in my 
> 
>>home directory), there were playback underruns shown that seemed to 
>>correlate with the log window errors.  A sample of these errors is 
>>copied below:
>>
>>---------------------------------------------------------------------
>>DiskStream Vocals:0: playback underrun (called with 8193 samples but 
>>only 5933 vailable; other half: 0) [speed=8]
>>---------------------------------------------------------------------
> 
> 
> Please notice the "[speed=8]" line. You were apparently
> fast-forwarding when this happened. An 8-fold speed up is quite a lot,
> especially if you have only 7200 RPM disks (or worse). If you weren't
> ffwd-ing, we have some other bug to solve, because ardour thinks you were.

Paul,

Thanks for pointing that out.  I was "playing" a bit with fast forward 
the same day well after the original errors were written to the log 
file.  Here is a more typical set of events from 23Feb04 with 
ardour-0.9_beta11:

transport Starting, sync poll of 1 clients for 2.000000 secs
transport Rolling, 2.000000 sec left for poll
load = 35.2009 max usecs: 4043.000, spare = 6623.000
load = 36.4079 max usecs: 4012.000, spare = 6654.000
load = 37.3161 max usecs: 4077.000, spare = 6589.000
load = 38.3140 max usecs: 4193.000, spare = 6473.000
load = 38.1988 max usecs: 4062.000, spare = 6604.000
load = 37.9818 max usecs: 4028.000, spare = 6638.000
DiskStream Vocals:0: playback underrun (called with 512 samples but only 
0 available; other half: 0) [speed=1]
load = 47.0099 max usecs: 5977.000, spare = 4689.000
transport command: STOP
transport Stopped

> 
> 
>>Another, possibly unrelated, dilemma is an issue I have been 
>>encountering when running latencies (with or without automation) at or 
>>below 3 ms latency (i.e., p=256, n=2 @ 48KHz sampling rate) when running 
>>a 2.4 kernel. In this case, the ardour log window displays a bizarre 
>>error message saying something about a MIDI memory pool error and to 
>>please compile with larger size.  When I run a 2.6 kernel, this error 
>>does not seem to occur. 
> 
> 
> i am sorry to say that this is a long standing problem that we have
> never tracked down. i hope we can solve it before the release of 1.0.
> please make sure its registered with mantis (ardour.org/mantis)

No problem, I can do that . . .

> 
> 
>>sessions), however, the disk stream errors persist whenever automation 
>>is enabled and my system latency is set any lower than 11ms 
>>(p=1024,n=2 at 48KHz).
>>
>>My current workaround is to run the gentoo development kernel 2.6.3-r1, 
>>and use low latencies as low as 3 ms for recording or non-automated 
>>mixdowns, and a setting of 11ms latency or greater if I am mixing down 
>>with automnation.
> 
> 
> how many tracks are you using? a PIII-450 is not a particularly speed
> machine, and it *is* possible that automation is bogging you
> down. however, the problems you mentioned above are not related to
> automation in any direct sense.
> 

Although my hdsp can handle up to 24 channels of ADAT I only am 
currently using a maximum of 8.  Most of the mixes I have so far have 
between 5 and 7 tracks.  IF you take a look at the spare time jackd is 
reporting in the extract pasted above, there seems to be enough headroom 
in the CPU for the work I am doing.  What is interesting is the fact 
that automation in one of the earlier betas (either ardour-0.9_beta9 or 
ardour-0.9_beta10) seemed to work without these issues.

Thanks again for all the fine work you and the ardour development team 
are doing.

Regards,

- David

-- 

David R. Bergstein
Systems Engineer and Blues Musician -
http://mywebpages.comcast.net/dbergstein/
Heart of Blue - bookings on-line at http://www.heartofblue.com
OpenPGP Public Key 0xE1F138CA - For info see http://www.gnupg.org
Key fingerprint = C86E CA2A 4171 AC73 91D7  3DCE 8832 D764 E1F1 38CA