Funny that someone would bring up this thread 'The Minute Before I'd Resurrect It'(tm).
Okay, so over at rockbox.org, Buschel has started testing the PortalPlayer chipset's registers for bits that would disable devices that aren't used. So as to save on battery output.
There really is two parts to it : testing the individual bits (which involves building, patching, building, patching... I must have built and installed about 20 different rockbox.mi4 now!) to know what works and what stop the devices from booting. -- And benchmarking the resulting Test Builds.
For example, there's currently a 'PP5020-test-02' build sitting in the forums at rockbox.org that's waiting to be benchmarked by people with H10s and sansa's. (I'll post a link further down).
So what can you do? Well, benchmark it! While you're at it, you might also benchmark the regular svn build (current build) and/or the OF.
Here's how I usually bench (and the results I got, to get you guys salivating a bit at the prospect of this patch getting committed someday soon) :
1) Benching
- Load up a folder with a few good files, preferably the same format (eg, all mp3 192k) and something like a double-album worth so that it can't all be contained in buffer.
- Charge the device 'till it's full.
- Boot it up with hold on to clear settings, you want a controlled environment.
- Choose a sound level and stick with it through the different benches (I chose -20db)
- Start plugin->apps->battery bench
- Start playing the first file in the 'test' folder, set device on hold so it isn't interrupted, and leave it for a few hours.
- Do not disturb. Wanna know if it's finished? Pick up the earphones and give a listen.
When you do another benchmark (from svn to patched, or vice-versa) then try to have the exact same settings and files. So comparisons become meaningful.
(Here's my latest results)
H10, 5gb
- With svn from two days ago, I got 5:27 runtime. Which was better than I thought. Then again, I guess album_art WPSes spin the disc more often than the default one.
- With the PP5020-test-02 patched build, I got 6:49 runtime.
- With the OF (see below) my old battery gave me 7:15 runtime.
Yep, 6:49 vs 7:15 .. That's almost the same !!
(To benchmark the Original Firmware you could plug your H10's output into your computer soundcard (with a male-male cord) and use windows recorder (or other) to record everything. Then when the player is dead, just look at the soundwave's lenght and you'll know how much time the battery gave you).
Ah, right... the link to the forum post.
All PP502x Targets - Device Disable Testing Build (SVNr16150 & FS#8379 v13)
Please, if you have time to bench, do. Especially if you have a 20gb unit (which have NOT been benched by anyone).