This is the bug/patch tracker for Rockbox. Click here for more information.
Quick links: Bugs · Patches · Rockbox frontpage
FS#9609 - Gigabeat S FM radio support
Attached to Project:
Rockbox
Opened by Nils Wallménius (nls) - Sunday, 07 December 2008, 20:56 GMT+2
Last edited by Nils Wallménius (nls) - Tuesday, 09 December 2008, 21:49 GMT+2
Opened by Nils Wallménius (nls) - Sunday, 07 December 2008, 20:56 GMT+2
Last edited by Nils Wallménius (nls) - Tuesday, 09 December 2008, 21:49 GMT+2
|
DetailsFm radio support for gigabeat S.
It works but seeking isn't nice yet and should really use the si4700's built in seek functionality. Muting doesn't work for some reason and the keymap can probably be improved. edit: forgot to say that i don't know what audio_set_output_source() should do and im' not quite sure about what i'm doing in audio_input_mux() is correct. 
gbs_fm_v6.diff
(16.9 KiB)
apps/keymaps/keymap-gigabeat-s.c | 13 firmware/export/wm8978.h | 3 firmware/export/si4700.h | 11 firmware/export/audio.h | 1 firmware/export/config-gigabeat-s.h | 7 firmware/SOURCES | 2 firmware/target/arm/imx31/gigabeat-s/audio-gigabeat-s.c | 55 +++ firmware/target/arm/imx31/gigabeat-s/power-imx31.c | 30 ++ firmware/target/arm/imx31/gigabeat-s/fmradio-i2c-gigabeat-s.c | 51 +++ firmware/drivers/audio/wm8978.c | 12 firmware/drivers/tuner/si4700.c | 140 +++++++++- firmware/tuner.c | 15 + 12 files changed, 333 insertions(+), 7 deletions(-) |
This task depends upon
Closed by Nils Wallménius (nls)
Tuesday, 09 December 2008, 21:49 GMT+2
Reason for closing: Accepted
Additional comments about closing: committed
Tuesday, 09 December 2008, 21:49 GMT+2
Reason for closing: Accepted
Additional comments about closing: committed
should be [(write_bytes[7] >> 4) & 3]
The datasheet mentions that some bytes should be not modified when written (first read and then written back).
Doesn't it make more sense to read all the registers (in a si4700_init() for example) and then modify individual bits, instead of using a built in list of registers values?
Fix the array indexing bug funman spotted and implement the suggested si4700_init().
Edit: updated patch without stupid mistake :)
* you should read all the registers to a temporary buffer and copy to write_bytes, because you start reading from register 0xA and write to 0x2
* you should write your copyright in si4700.c ;)
* the #endif comment of the si4700 case in tuner.c is for another tuner
* I don't see the power up bit set in si4700_init() (I see it in si4700_set() however)
* si4700_init() is indented on 5 spaces
Also are the GPIO interrupts needed?
it should remain in the RADIO_SLEEP case in si4700_set.
Make the detection actually work, leave softmute enabled, lower the output
volume to be a little closer to playback, fix a couple of mistakes.
1) The FM radio's volume is quite loud, much louder than normal playback should be.
2) Entering the main menu from any other screen seems to take a second or two longer than it should.
Other than that, great job! :)
1) Fixed a small bug in si4700.c, where returning to the main menu from any other menu or the file viewer would take a second or two longer than it should; this was done by moving fmradio_i2c_read from si4700_get to the RADIO_STEREO case.
2) In audio-gigabeat-s.c, fixed the passthrough volume to be a lower value; originally it was set to 7, which was causing the radio to be much louder than normal playback (passthrough volume was +6dB at this point). Setting it to 5 now sets the passthrough volume to 0dB, which is much more sensible.
3) I set the VolUp and VolDown buttons to work as they do in the WPS on the FM radio screen as well. Please help test this.