I've been investigating further, I haven't figured out where, in the source, the problem occurs.. the alsa input midi buffer is quite large [much larger than 226 bytes],

When testing with amidi, this is the largest message i can send to pd: F0 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F 10 11 F7 so I'm doing: amidi -p hw:2,0 -S "F0 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F 10 11 F7"

I'm using a virtual device to echo the data to an amidi -d instance as well as to pd.. amidi -d gets all messages i send... larger ones like: F0 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F 10 11 12 F7 but pd doesn't display any output at all from [midiin] or [sysexin]

-Alex

On Fri, Jan 16, 2009 at 3:08 PM, Alex x37v.alex@gmail.com wrote:

I made a little test which I've attached.

With alsa I can receive a maximum sysex message of 98 data bytes [100 total including the start and end]. I've tried this with a hardware loopback (plugging my output into my input) and using a virtual device.

With the default midi [portaudio?] I can receive 128 data bytes [130 total].

I've attached my test patch [it requires a counter that I've also attached].

Any idea where these limits are set?

-Alex

On 1/16/09, Alex x37v.alex@gmail.com wrote:

...

I am trying to get a sysex dump into pd, this dump is 226 bytes long. I use alsa for midi. I'm using pd 0.41.4.

I am able to send the sysex data from pd to my device, and the device sends the data back [I can verify the dump with amidi] but once I connect my device's output to pd.. I don't see the data in pd [sysexin] or [midiin]

I've also tried initiating the dump from the device itself (a synthesizer), I get the same results.. amidi -d will show it but I don't see the data on pd's [midiin] or [sysexin] inputs once I connect the device to pd.

So my question is: is anyone aware of a hard coded limit for the length of sysex messages in pd? I'd like to increase this limit if possible..

Thanks,

Alex

Alex wrote:

I've been investigating further, I haven't figured out where, in the source, the problem occurs.. the alsa input midi buffer is quite large [much larger than 226 bytes],

...

...

So my question is: is anyone aware of a hard coded limit for the length of sysex messages in pd? I'd like to increase this limit if possible..

I recall this being spoken of, and I think solved, some time ago, but this was all I could see on a quick search:

http://lists.puredata.info/pipermail/pd-list/2007-07/052402.html

the thread I seem to remember was longer

Simon

On Sat, 17 Jan 2009, Alex wrote:

Ahh, thanks! I just viewed the diff of desire data's s_midi_alsa.c and made the changes I needed for my version.. though, I would eventually like to be able to share this patch that needs this larger sysex input.. so I think I'll try to make a command line option... in the mean time, I've attached the diff [exactly what they do in Desire Data except I changed the name to be a little more descriptive, I think].

You should use diff -u for getting a "unified diff" (+/-) format, which is better because you can then apply it on a different file than the original as long as it is similar enough.

_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - tél:+1.514.383.3801, Montréal, Québec