Audio problem on I2
Rob
Just started experiencing an audio problem, neatly described back in
1999 here :
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&q=indigo2+mp3+buzz&sa=N&tab=wg
Whats odd is i'm running 6.5.17m. The buzz kicks in after a undefined
length of time (usually less than 2 minutes) when I'm not using the
machine (but before blank screensaver kicks in at 20 minutes). It's
totally random too,
Play back of mp3's with mpg123 or xmms makes no diff.
Play back in xmms results in the uv meter freezing at the last correct
played part of the mp3
Audio queue flickering block freezes during the buzz, errors do not
increment
The buzz can occasionally stop of it's own accord, but usually i move
the mouse or hit any key to stop it
obligatory disc-lamer : play's ok in linux...but i aint got a speaker
in my peecee.
Never seen this before, having played mp3's on this machine before on
earlier M releases of IRIX. Have never played mp3's on this release of
IRIX. Anyway, it's ruining chillin' to Massive Attack. Can any of you
share an experience ?
BZZZZZZZZZZZZZZZ
Cheers,
Rob
mike
with the audio, I was never able to find a way of fixing it, neither were the people that I
talked to about the issue.
Rob
> Is this a 250mhz R4400 ?
Yes it is.
> Seems there is a CPU bug
effffin great!
I've played mp3's on this before with no buzz, so I'm not beat.
It really does take the piss tho. I'd just got this machine to 'maxed
up' status, and now of all the luck this.
pffffff
Rob
And people chastise peecees.
mike
>
> It really does take the piss tho. I'd just got this machine to 'maxed
> up' status, and now of all the luck this.
I dunno, this problem was the excuse I gave myself for getting an 195mhz
R10k Max Impact Indigo2, Im not regretting it, although I would
recommend getting an octane now :)
cheers!
mike
Rob
> Playing audio creates BZZZ sound on R4400 250 2MB; stops on mouse/keyb
> movement.
What seems clear here is that issuing a keypress or mouse movement
fixes the problem until the is no keyb/mouse input again
AFAIK mouse/keyb generates a hardware interrupt.
Is it possible to generate such an interrupt programatically ? I'm
currently searching thru stuff, but I'm not a C programmer and my
knowledge in this regard and therefore search criteria are probably
not getting me the best google responses. I however am looking into a
way of having some kind of script/c prog running in the background
which when detects no mouse/keyb input for a specific amount of time
generates the equivalent interrupts (or other side effects) to the
system to see if this problem can be worked around.
If I manage it, I'll post it up in here so all can use it.
Anybody with anything, even simply better terms to search on, please
step forward.
Cheers
Rob
Bryan Manternach
> And people chastise peecees.
If it makes any difference to you, I have never had any luck with
eSound or similar MPEG audio players on IRIX. XMMS has never been
stable in any way for me, on any IRIX machine except the Indy.
I've been using "mxaudio" on IRIX for quite some time, and it
was fine even on a 250 MHz R4400 Indigo2 Impact. It's shareware,
so you shoudl send inthe ~$10 registration fee, but IMO, it was
worth it for a solid MP3 player for IRIX.
--
"Windows 95: 32 bit extensions and a graphical shell for a 16
bit patch to an 8 bit operating system originally coded for a
4 bit microprocessor, written by a 2 bit company, that can't
stand 1 bit of competition."
---------------------------------------------------------------
Bryan 'SMASH' Manternach - ECS/TREX Systems Administration Team
SGI, 1200 Crittenden Lane, MS 30-4-175, Mountain View, CA 94043
E-mail: mailto:sm...@sgi.com Office: (650)933-6856
Pager: mailto:sma...@pager.sgi.com Mobile: (408)209-3099
David Evans
Bryan Manternach <sm...@sgi.com> wrote:
>
>If it makes any difference to you, I have never had any luck with
>eSound or similar MPEG audio players on IRIX. XMMS has never been
>stable in any way for me, on any IRIX machine except the Indy.
>
But as with all anecdotes, things get weird quickly. I've had the
XMMS from SGI's freeware collection work reliably for ages on R10k Indigo2
and R5k O2. These were under various flavours of 6.5. This was using the
IRIX AL output plugin IIRC.
--
David Evans (NeXTMail/MIME OK) dfe...@bbcr.uwaterloo.ca
PhD Student, Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/
University of Waterloo "Default is the value selected by the composer
Ontario, Canada overridden by your command." - Roland TR-707 Manual
lisp
> But as with all anecdotes, things get weird quickly. I've had the
> XMMS from SGI's freeware collection work reliably for ages on R10k Indigo2
> and R5k O2. These were under various flavours of 6.5. This was using the
> IRIX AL output plugin IIRC.
I find certain MP3s cause xmms to bus error and dump core. I tried
recompiling the MP3 module, but that made things worse.
--
(lisp) - http://www.thecommune.org.uk/~lisa
SBS#37(with oak leaf cluster) two#40 DOGMUK
Suzuki GS125ES
My other car's a cdr
Q...@quickdekay.net
>
> I find certain MP3s cause xmms to bus error and dump core. I tried
> recompiling the MP3 module, but that made things worse.
Which mp3s? You might want to try
http://quickdekay.net/tmp/xmms-1.2.7-compiled.tar.gz ..
Everything in the tar resides under /usr/local/xmms; so as long as
you don't already have this directory, no files will be overwritten.
You need the libraries gtk, gdk, gmodule, gthread and glib other than
the normal X and xmms libraries. There's a possibility that my binaries
will not work with the /usr/freeware/lib ones; the libraries mentioned
are also custom compiled libraries. If you (or whoever) needs those
libraries, just mail me.
I've never had xmms crash on me, hope it helps.
--
Khamba Staring
"I want a sitting duck serving as a scapegoat for my guinea pig!"
César Blecua Udías
>
> Is this a 250mhz R4400 ? Seems there is a CPU bug with
> those that will always cause a buz with the audio, I was
> never able to find a way of fixing it, neither were the
> people that I talked to about the issue.
Some people told me they never experienced this "freeze" problem with
R4400-250MHz I2, but I'm not sure if they used the machine for
applications where the problem is noticeable (cursors stopping to blink
in some text editors, gr_osview freezing, as well as the mentioned audio
buzz).
Or maybe their machines were getting interrupts very often from
somewhere, so that the freeze never happened...
I thought that perhaps some hardware revisions might fix the problem,
and so I posted a message at comp.sys.sgi.hardware in November 12th
asking whether it's possible to run a R4400-250MHz without this "freeze"
problem. Unfortunately, I got no response.
César
cesar...@ono.com
Rob
I have a work around.
It's not nice by any stretch of the imagination, but it works for me.
http://thecave.reality.dsvr.net/irix/indigo2_lockup_workaround/
It's a UDP client server which satisfies the one thing - it keeps the
indigo2 busy enough to not lockup and buzz. Run the server on your
indigo2 and the client on something nice and stable - like a linux
box. The steady stream of nonsense will keep you Indigo2 happily
running mp3's all night long.
I too have now started getting bus errors with XMMS, on some mp3's
more than others. mpg123 dies on them too, and it might be something
to do with vbr mp3's, however i've been watch XMMS happily adjusting
itself. Ah well. I will catch up to these ever moving goal posts.
XMMS has always been good on both my Octane and my Indigo2 using the
libIRIX.so audio plugin. It's only since 6.5.17m I have started
getting the buzz and the bus errors - infact its only today i started
getting the bus errors. I'm using the libsgial.so plugin now, however
changing it doesn't seem to make any dif to this.
RAM was recently checked out, tested with memtest and IRIX has never
died with this RAM. The bus errors are probably bugs in XMMS, or some
quirk of IRIX it doesn't handle well.
Ah whatever, I'm happy the machine's stopped buzzing.
Incidently does Bryan Manternach (sm...@sgi.com) ever get the
lockup-buzz from ``mxaudio'' ? mxaudio did run nice on my Indy, and
was less resource intensive.
Anyway, I hope this workaround is helpful to someone else.
Rob
Tony 'Nicoya' Mantler
(Rob) wrote:
: I too have now started getting bus errors with XMMS, on some mp3's
: more than others. mpg123 dies on them too, and it might be something
: to do with vbr mp3's, however i've been watch XMMS happily adjusting
: itself. Ah well. I will catch up to these ever moving goal posts.
The standard id3v2 library is a mess of spaghetti-C++, and is full of bugs. This
is likely the source of your crashes.
If you find an mp3 that xmms crashes on, try stripping the id3v2 tags and
playing it again. Most likely it'll fix the crash. I think it's also possible to
disable id3v2 support, while retaining old (somewhat more stable) v1 support.
Cheers - Tony 'Nicoya' Mantler :)
--
Tony "Nicoya" Mantler - Renaissance Nerd Extraordinaire - nic...@apia.dhs.org
Winnipeg, Manitoba, Canada -- http://nicoya.feline.pp.se/
Pete Plank
to find a home for it yet ... for the price of shipping it's yours if
you want to give it a try.
r...@dsvr.net (Rob) wrote in message news:<b8993cd5.02121...@posting.google.com>...
Rob
> In article <b8993cd5.02121...@posting.google.com>, r...@dsvr.net
> (Rob) wrote:
>
> : I too have now started getting bus errors with XMMS, on some mp3's
> : more than others. mpg123 dies on them too, and it might be something
> : to do with vbr mp3's, however i've been watch XMMS happily adjusting
> : itself. Ah well. I will catch up to these ever moving goal posts.
>
> The standard id3v2 library is a mess of spaghetti-C++, and is full of bugs. This
> is likely the source of your crashes.
>
> If you find an mp3 that xmms crashes on, try stripping the id3v2 tags and
> playing it again. Most likely it'll fix the crash. I think it's also possible to
> disable id3v2 support, while retaining old (somewhat more stable) v1 support.
Well done that man. You are absolutely correct. Stripping the v2 id3
from the file and XMMS and no more bus errors.
http://freshmeat.net/projects/v2strip/?topic_id=120
$ /home/rob/bin/v2strip *
File '01 The Orb - Hidden[FFWD] (Fripp, Fehlmann, Weston, Paterson).mp3'
does not have any valid ID3v2 tag
File '02 The Orb - Lucky Saddle.mp3' does not have any valid ID3v2 tag
Stripping 03 The Orb - Drone.mp3... Done
File '04 The Orb - Hempire.mp3' does not have any valid ID3v2 tag
Stripping 05 The Orb - Collossus.mp3... Done
Stripping 06 The Orb - What Time Is Clock.mp3... Done
File '07 The Orb - Can Of Bliss.mp3' does not have any valid ID3v2 tag
Stripping 08 The Orb - Elevenses.mp3... Done
File '09 - Meteor Storm.mp3' does not have any valid ID3v2 tag
File '10 The Orb - Buckwheat & Grits.mp3' does not have any valid ID3v2 tag
Stripping 11 The Orb - Klangtest.mp3... Done
Stripping 12 The Orb - Suess wie eine Nuss.mp3... Done
Consistant bus error on track 5. I wonder what the charactors in it were :)
Rob
yi ge mei you wei ba
>
> Some people told me they never experienced this "freeze" problem with
> R4400-250MHz I2, but I'm not sure if they used the machine for
> applications where the problem is noticeable (cursors stopping to blink
> in some text editors, gr_osview freezing, as well as the mentioned audio
> buzz).
>
> Or maybe their machines were getting interrupts very often from
> somewhere, so that the freeze never happened...
>
> I thought that perhaps some hardware revisions might fix the problem,
> and so I posted a message at comp.sys.sgi.hardware in November 12th
> asking whether it's possible to run a R4400-250MHz without this "freeze"
> problem. Unfortunately, I got no response.
>
> César
> cesar...@ono.com
I watched for replies to your message and never saw any, either.
I was interested because I was getting a freeze / total loss of
user control. Usually it appeared as I was moving a scrollbar.
The scrollbar would stop moving. Then apps wouldn't respond to
requests to change focus. Then the cursor would freeze. Telnetted
top showed the cpu running at 100% utilization and xsgi gobbling
up 80+% of cpu time.
This would happen at random, about once a day. I distrusted just
about every app I own, but never could pin down a culprit.
Then I got a larger hard drive and moved the one in the Octane to
an O2. Suddenly the Octane quite freezing : and the O2 began. It's
only been a week, but the evidentiary finger is pointing very hard ....
Fujitsu hard drive OEM for Compaq. I _knew_ that drive was gonna
be Trouble the moment I laid eyes on the Compaq label :-(
César Blecua Udías
I think we're talking about different problems. The freeze I was talking
about is a temporal "pause", which doesn't cause any harm except for
audio playing (and a huge date delay, of course), and it happens on
R4400-250MHz only. The whole system gets paused for a few seconds, until
an interrupt happens.
César
cesar...@ono.com
yi ge mei you wei ba
>
> I think we're talking about different problems. The freeze I was talking
> about is a temporal "pause", which doesn't cause any harm except for
> audio playing (and a huge date delay, of course), and it happens on
> R4400-250MHz only. The whole system gets paused for a few seconds, until
> an interrupt happens.
>
Ahh. You had a glitch. I had a _freeze_ ! May as well keep it in mind
anyway - if you ever get totally random senseless freezes, think hardware.
Two weeks on a new hard drive now without one recurrence.