Cowon D2

[saratoga]

Are binaries of "current build" built with that "IRAM"-optimization ?

Nope.

[mbk1969]

Will be ?

[shotofadds]

Once I've commited it, yes. 

btw. Thanks for updating the table saratoga, I would have done it myself but I can't seem to download parse_testcodec.pl (403 Forbidden on the wiki link). Also how did you calculate those percentages? Mine were effectively "the previous test took xx% longer", which is probably not quite right.

[saratoga]

I just used the script (which indeed doesn't seem to download right now):

$oldfile{$key}->[1]-$newfile{$key}->[1])/$oldfile{$key}->[1]*100+100

so basically (old - new)/old

[shotofadds]

The 6% improvement in MP3 decoding speed seems to have resulted in an improvement in battery life of about 45 minutes (to 18h30m, playing VBR ~190k MP3s from SD). You might get better results with Vorbis / WMA / MPC files, or without an SD card inserted.

I'll get these changes committed soon.

[funman]

this link seems to work (I got here through the revisions of the script)

[shotofadds]

The IRAM change is now committed, although I wouldn't get too excited about it as it doesn't make a huge difference.  Also committed this weekend was a patch from FS#11240 that extends the volume range down to -89dB.

[mbk1969]

I got my nose in Wolfson WM8985 spec and became curious about fully customizable three analog ouputs of this chip. By default Rockbox disables not used outputs, right? And wich output is used in D2 - LOUT2/ROUT2 or LOUT1/ROUT1?

update: D2 with original firmware can play video with line audio signal through the pseudo-usb connector and I think that OUT3/OUT4 outputs are used for that. It can be checked with added  advanced settings for enabling theese outputs (and with such cable)...

[motorbit]

line out is a feature missed forever by some users.
albeit, personally i would not care to much. but i guess it might worth implying it if possible without to much work.

[shotofadds]

It's been a long time since I looked at this. LOUT1/ROUT1 is definitely used for the headphone output, and I *think* LOUT2/ROUT2 is used for the audio portion of the TV-out cable. This could be wrong, of course, and even if correct there might need to be some extra logic to enable it.

The code currently assumes L/ROUT1 for headphones and L/ROUT2 for line:

http://svn.rockbox.org/viewvc.cgi/trunk/firmware/drivers/audio/wm8985.c?revision=26413&view=markup

but I don't have the appropriate cable(s) to test it.

[mbk1969]

Thank you for listing. I will dig in. I have a feeling that instead of muting unused outputs in audiohw_preinit() you can disable them - that will be more powersaveable.
I think I am ready to join the RB as developer... will see...

Can you build patched version with audiohw_set_lineout_vol() called somewhere in initialization stage? It can be tested without cable but with proper measure tools.

[saratoga]

Can you build patched version with audiohw_set_lineout_vol() called somewhere in initialization stage? It can be tested without cable but with proper measure tools.

If you've got a line out cable, maybe you should try it and see.

[mbk1969]

ok, I`ll try

[motorbit]

when i first used rb i hated it - meanwhile i dont use the of anymore.
bit thanks for all that work you had put in.



btw, im expecting a bug with rb:
i often have quite harsh crashes of the player when i plug it onto the usb cable. ill have to use the reset button than to reactivate it.

when i plug my switched of player onto the usb cable with hold engaged, my player will crash most of the time.

when i plug in my switched off player onto the cable with hold disengaged, my player will crash rarely.

when i plug in my switched on player onto the cable it will never crash no mater if  rb is launched or the of.


so, its not a big issue, albeit i often forget to switch off hold, its not a big thing to press the reset button. just wondering if anybody else expiriences this issue. my player is a 16 gig d2+ btw - on of the first d2+ that still do have  the old powermanagement chips.

[buk]

so could anybody tell me which files of the sources' code  should be concerned?

Check out:

/firmware/target/arm/usb-tcc.c

and

/firmware/target/arm/tcc780x/*

Also, /firmware/SOURCES lists most of the D2 specific files if you search it for "D2".

I made a patch regarding USB a while ago (FS#9917).
As said even with those changes USB was unstable. Btw the patch correct a bug in usb_drv_send_nonblocking.