Skip to contentSkip to conversation

sequencer buffer

DISCUSSION Bernhard Gerlach 2 participants 3 messages Nov 7, 2003 – Nov 13, 2003

lore
Share & tools
lore
#0 lore raw

Hello NG,

again I've got a question regarding the ALSA midi sequencer. I tried to generate Midi time code. For that, I allocated a queue, started it and then filled it up with a couple of midi quarter frame messages. Each messages had got a timestamp containing the (real/absolute) time, at which the message should be sent by the queue. After filling the queue, I drained the buffer and tried to let the process sleep until snd_seq_event_output_pending( handle ) falls below a given value (I tried 500). But I realized, that the buffer size is never decremented, even if the queue outputs events.

The first 7 seconds of output seem to be ok (I control the output on a harddisk recorder which just reads the incoming midi time code and displays it). But then I get "Resource temporarily unavailable" errors when enqueueing another quarter frame event. I think, thats because the output buffer is full. But why isn't an event extracted from buffer when it is sent by a queue?

Thanks for answers,

Bernhard

Signature · 17 lines
---------------------------------
Bernhard Gerlach
Fraunhofer AEMT
Am Helmholtzring 1 (Haus M)
98693 Ilmenau
Germany

mail: [email protected]
---------------------------------



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?   SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/
#1 lore raw
Quoted 2 lines

again I've got a question regarding the ALSA midi sequencer. I tried to generate Midi time code.

you cannot send 100% valid MIDI time code on a Linux system without the high resolution timers patch or some other interrupt source that provides a suitable timing resolution.

MIDI time code is not just a sequence of bytes, it has very precise semantics about when those bytes are to be delivered. there is no mechanism in a default linux kernel on 99%-100% of all hardware today that can implement those semantics.

i know that this isn't directly related to your question, but i wanted to alert you to this issue. although other devices will be able to chase your MTC data stream, some or even all will not lock to it because it is jittery/bursty.

--p

-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it
help you create better code? SHARE THE LOVE, and help us help
YOU! Click Here: http://sourceforge.net/donate/

#2 lore raw

Hi,

late, but its comes: my answer.

Quoted 4 lines · Paul Davis

you cannot send 100% valid MIDI time code on a Linux system without the high resolution timers patch or some other interrupt source that provides a suitable timing resolution.
[..]

Ok. I didn't really think about that. I deferred generating time code, since it is not very important at the moment. Maybe I have to work on that later, but then I'll first have to look for a appropriate timing hardware and something like a realtime/high res timer patch for the linux kernel.

Quoted 4 lines

i know that this isn't directly related to your question, but i wanted to alert you to this issue. although other devices will be able to chase your MTC data stream, some or even all will not lock to it because it is jittery/bursty.

Thanks a lot for this suggestion!

Bernhard

Signature · 17 lines
---------------------------------
Bernhard Gerlach
Fraunhofer AEMT
Am Helmholtzring 1 (Haus M)
98693 Ilmenau
Germany

mail: [email protected]
---------------------------------



-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/