Ok about to install, are the instructions on page 1 still the ones to use on Diablo on an MMC (don't feel like going through 20 pages but will while waiting for a reply.

OK I installed it along with lardmans Package.

One question I have is I have 2 bluetooth A2DP headphones. If I run the a2dp runfirst twice giving both headsets address' will they both work with this or does the last one overwrite the first one? Is there a file I can manually edit so it enables it on both headsets?

To flash the zImage there's either an on device flasher (anyone got instructions for that?) or the standard Linux desktop flasher (http://maemo.org/community/wiki/flasher_tool_usage/)

Do you suppose there might be a way to run flasher using a unix shell on windows, like cygwin?

Update: Nope - just checked the Cygwin website, and it's pretty clear on this...

CyberCowboy: Yes, the instructions on page 1 are still the best I think. yes, it will overwrite the other setting. The only way that I can see for you to do this is run the "runfirst" when you switch headphones or manually edit /home/user/.asoundrc when you switch. *Sigh* I don't think there was any kind of plan for supporting more than one pair of A2DP headphones on one device. Maybe some kind of trick to use an ALSA mixing module with both pairs listed as outputs? I really don't know...

Yeah, same as the PC flasher.

Thanks GeneralAntilles! That worked!

Just a report on how everything worked. The setup is with a Sony TDM-BT1 digital media bluetooth unit on the home audio. Before the A2DP task was ported to the DSP, music would play for about a second or two, followed by about 1/2 second of silence for the length of the song. Now, with the DSP on, there is only the occasional very short skip. Totally acceptable, by the way, and I'm very grateful to Simon for the code.

Here are some other observations: As someone else has suggested, it's better if the WiFi is disconnected. Also, the music will skip if anything else is done on the tablet (like accessing the GUI, for example). I also noticed that if the N810 is close to the receiver (~1-2 m), the skips are fewer and farther between - presumably because some of the data is dropped in the transmission when the unit is over 5 m away. Also, it seems like the pitch is a tiny bit unstable: on one song (My Song, Keith Jarrett) there is a repeated note on the piano. I noticed that note going slightly sharp, and then slightly flat with a period of a few seconds, almost sinusoidally, which was curious.

Anyway, thanks to all of the developers and those who helped me run the script and figure out how to flash the kernal. This version of A2DP support is a great step forward, in my opinion. If anyone would like, I could write up the steps I took to get this up and running.

Yeah, this is the issue to do with Bluez not having accurate enough timers/the headset not being able to report back where it is. The developers know about this and are working on it. It would also occur if the encoder task was getting too far behind, but I think the DSP task is now more than quick enough - at least for my listening tests, the pitch changes were minor (I assume this is the timer issue) rather than the pretty major ones you can get with the slow encoder.

Thanks, I figured that was the case and I wrote a small script for each headset that does the "runfirst" and puts in the different mac address' and then imported that into the osso-statusbar-cpu. No biggy. I am just one of those odd-balls, I got the two headsets for a total of $40, and couldn't pass them up since the car one also is paired with my cell phone, makes life easier.