Sometime back around pd-0.23 or so, i had more-or-less duplex operation happening - that is, i was able to use bonk~ and run some other patches requiring both audio input and output. This was on a Dell PII 350 w/ RH 5.0 and commercial OSS drivers. At some point, Very Bad Things happened to the filesystem on that machine - so i installed RH 6.0, the current OSS package and pd-0.26. Now, audio output works fine, but if i start pd without the -noadc flag the following appears:
OSS: device Nr. 1 on /dev/dsp pd: OSS: requiring extra adc handles pd: OSS: requiring extra adc handles OSS: failed to open /dev/dsp2 read only
and the terminal window is flooded with this message: OSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC buf etc. ad quasi-infinitum.
i've messed around with various other startup flags (buffersize, frags) and none of them help. pd w/ audio out only is great, but in AND out would be outstanding. Has anyone seen similar strangenesses? seems like if it worked once it _should_ be able to work again.
thank you
Bill Sack bsack@toast.net wsack@acsu.buffalo.edu
"'Michael Gilhaney,' said the Sergeant, 'is an example of a man that is nearly banjanxed from the principle of the Atomic Theory. Would it astonish you to hear that he is nearly half a bicycle?'" -Flann O'Brien, "The Third Policeman"
The funny thing about Bill's predicament is that I don't know of any test program in Linux that exercises simultaneous audio input and output -- if Pd fails, I don't know how to set about deciding whether it's a problem with Pd or with OSS (or Alsa or whatnot.) I'm using Pd0.28TEST6 right now and enjoying full duplex, for what that's worth... (available for Linux only so far, from www.crca.ucsd.edu/~msp/Software)
cheers Miller
On Sun, Nov 14, 1999 at 05:03:15PM -0700, Bill Sack wrote:
Sometime back around pd-0.23 or so, i had more-or-less duplex operation happening - that is, i was able to use bonk~ and run some other patches requiring both audio input and output. This was on a Dell PII 350 w/ RH 5.0 and commercial OSS drivers. At some point, Very Bad Things happened to the filesystem on that machine - so i installed RH 6.0, the current OSS package and pd-0.26. Now, audio output works fine, but if i start pd without the -noadc flag the following appears:
OSS: device Nr. 1 on /dev/dsp pd: OSS: requiring extra adc handles pd: OSS: requiring extra adc handles OSS: failed to open /dev/dsp2 read only
...
Bill Sack bsack@toast.net wsack@acsu.buffalo.edu
I have had the same problem finding other programs that do full duplex audio in linux - especially for Alsa. Right now for me, full duplex OSS free crashes my computer pretty hard (this is a known problem with OSS and multiprocessor kernels), reading from my sound card under Alsa seems completely broken for all programs, and the Alsa oss emulation is not full duplex! I think that linux is the best choice for this kind of thing in the long run - I can currently get perfect playback with the smallest buffer setting in PD and the sound card driver which is less than 10ms total I think - but the sound card support has got to get better. Speaking of which, I would really like to hear some reports on the RME card.
Karl
On Sun, 14 Nov 1999, Miller Puckette wrote:
The funny thing about Bill's predicament is that I don't know of any test program in Linux that exercises simultaneous audio input and output -- if Pd fails, I don't know how to set about deciding whether it's a problem with Pd or with OSS (or Alsa or whatnot.) I'm using Pd0.28TEST6 right now and enjoying full duplex, for what that's worth... (available for Linux only so far, from www.crca.ucsd.edu/~msp/Software)
cheers Miller
On Sun, Nov 14, 1999 at 05:03:15PM -0700, Bill Sack wrote:
Sometime back around pd-0.23 or so, i had more-or-less duplex operation happening - that is, i was able to use bonk~ and run some other patches requiring both audio input and output. This was on a Dell PII 350 w/ RH 5.0 and commercial OSS drivers. At some point, Very Bad Things happened to the filesystem on that machine - so i installed RH 6.0, the current OSS package and pd-0.26. Now, audio output works fine, but if i start pd without the -noadc flag the following appears:
OSS: device Nr. 1 on /dev/dsp pd: OSS: requiring extra adc handles pd: OSS: requiring extra adc handles OSS: failed to open /dev/dsp2 read only
...
Bill Sack bsack@toast.net wsack@acsu.buffalo.edu
| Karl W. MacMillan | | Peabody Institute of the Johns Hopkins University | | Network and Telecommunications Services | | karlmac@peabody.jhu.edu | | 410/659-8297 |
I have the same behaviour, but without the ADCbuffOSS business. I've tried both FreeOSS and commercial OSS. From what I can tell the only /dev/dsp channels that are open on my machine are /dev/dsp0:/dev/dsp1:/dev/dsp:/dev/audio
some of which may be aliases I don't really know. What I have been able to determine is that /dev/dsp2 is not available for input or output. I'm going to have a look in the code and see if I can find out why it is looking to /dev/dsp2. Perhaps the most recent OSS changed it's behaviour and is not sending pd the correct info.
-Peter Washburn
Bill Sack wrote:
Sometime back around pd-0.23 or so, i had more-or-less duplex operation happening - that is, i was able to use bonk~ and run some other patches requiring both audio input and output. This was on a Dell PII 350 w/ RH 5.0 and commercial OSS drivers. At some point, Very Bad Things happened to the filesystem on that machine - so i installed RH 6.0, the current OSS package and pd-0.26. Now, audio output works fine, but if i start pd without the -noadc flag the following appears:
OSS: device Nr. 1 on /dev/dsp pd: OSS: requiring extra adc handles pd: OSS: requiring extra adc handles OSS: failed to open /dev/dsp2 read only
and the terminal window is flooded with this message: OSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC bufOSS: drop ADC buf etc. ad quasi-infinitum.
i've messed around with various other startup flags (buffersize, frags) and none of them help. pd w/ audio out only is great, but in AND out would be outstanding. Has anyone seen similar strangenesses? seems like if it worked once it _should_ be able to work again.
thank you
Bill Sack bsack@toast.net wsack@acsu.buffalo.edu
"'Michael Gilhaney,' said the Sergeant, 'is an example of a man that is nearly banjanxed from the principle of the Atomic Theory. Would it astonish you to hear that he is nearly half a bicycle?'" -Flann O'Brien, "The Third Policeman"
-
Bill Sack -
Karl W. MacMillan -
Miller Puckette -
Peter Washburn