Rockbox: Clip+

[marc2003]

Originally Posted by funman

If you want to change your rockbox build, edit firmware/drivers/audio/as3514.c function audiohw_set_master_vol and change "mix_r = 0x16" and "mix_l = 0x16" into any number. The value ranges from 0 (minimum volume) to 0x1f (31, maximum volume).

i tested this and these are my results. the OF is much louder than rockbox at stock. can't say i noticed with my headphones though.

i used windows sound recorder as saratoga suggested and recorded the clip v2 on full whack and then used foobar2000's replaygain scanner....

[setmenu]

.

[dfkt]

Originally Posted by funman

If you have a clipv1 and a clip+ or clipv2, can you compare OF max volume and rockbox max volume ? (compare OF and rb, and also clipv1 vs the other one).

RB Clipv1 is a tiny bit louder than RB Clip+.
OF Clipv1 is a tiny bit quieter than OF Clip+.
OF Clip+ is clearly louder than RB Clip+.

Absurdly, using the Clip+ with volume 0x1d is the same as using the stock 0x16, no difference in volume. How can that be?

Originally Posted by funman

If you want to change your rockbox build, edit firmware/drivers/audio/as3514.c function audiohw_set_master_vol and change "mix_r = 0x16" and "mix_l = 0x16" into any number. The value ranges from 0 (minimum volume) to 0x1f (31, maximum volume).

I've checked the highest three settings with RMAA. 0x1d should give perfectly acceptable results, 0x1e is also still very good, 0x1f is a bit too much on paper (still only tiny differences at 0db, but still).

RMAA - 0x1d
RMAA - 0x1e
RMAA - 0x1f

[Rockfella]

Can anyone suggest a way to un-rockbox it completely???? I want to get rid of it.....

[saratoga]

Originally Posted by Rockfella

Can anyone suggest a way to un-rockbox it completely????

Yes, just update your firmware:

http://download.rockbox.org/daily/ma...ml#x4-200002.5

[SunnyBoi]

Getting an error while building database with r26037. Tried 3 times and got this error everytime:

*PANIC*
transfer error -18

[Rockfella]

It didnot work with me so i had to manually upgrade sansa firmware, working fine now.

Originally Posted by saratoga

Yes, just update your firmware:

http://download.rockbox.org/daily/ma...ml#x4-200002.5

[marc2003]

Originally Posted by SunnyBoi

Getting an error while building database with r26037. Tried 3 times and got this error everytime:

*PANIC*
transfer error -18

i can confirm this. it's r26028 that introduced the problem. 26027 is fine.

the patch looks to be SD card related but i don't even have one being a clip v2 user???

[saratoga]

Originally Posted by marc2003

i can confirm this. it's r26028 that introduced the problem. 26027 is fine.

the patch looks to be SD card related but i don't even have one being a clip v2 user???

Internal memory is also SD.

[marc2003]

my ignorance knows no bounds.

[ultrauber]

I know it's old new but hey it's r/w now. Too bad it's so unstable...keep up the work.

[kifo]

Originally Posted by marc2003

i can confirm this. it's r26028 that introduced the problem. 26027 is fine.

the patch looks to be SD card related but i don't even have one being a clip v2 user???

Ran into this on r26037 also, fixed by removing all files starting with 'database' in .rockbox, this might be overkill, looks like the only one of these added was database-ignore.

EDIT: Looks like it still crashes when I try to initialize the database, with the same error.

[Timar]

Originally Posted by dfkt

Absurdly, using the Clip+ with volume 0x1d is the same as using the stock 0x16, no difference in volume
...
I've checked the highest three settings with RMAA. 0x1d should give perfectly acceptable results, 0x1e is also still very good, 0x1f is a bit too much on paper (still only tiny differences at 0db, but still).

Why are the TMD/IMD levels so much higher than on the OF (more than two orders of magnitude!!), even if the overall volume on rockbox is lower (on ox1d)!? Seems like a contradiction... are the settings > 0 handled different from the highest volume setting on the OF?

Anyway - I vote for setting the volume to 0x1e, if that value actually gives more volume. It comes at almost no cost!
In fact it would be way better than the current setting if it gives you adequate volume levels before reaching values > 0!!

[funman]

Originally Posted by SunnyBoi

Getting an error while building database with r26037. Tried 3 times and got this error everytime:

*PANIC*
transfer error -18

Check the filesystem with chkdsk.exe / fsck, this error means the storage is accessed beyond its reported capacity.

It could also mean that the capacity is detected incorrectly: what is the capacity of your clip and what is the number of Blocks in System -> Debug menu -> View Disk Info ? (microSD0 is the internal storage and microSD1 is the µSD slot)

[marc2003]

i ran chkdsk and it muttered something about some errors so i ran it again with the /f switch to fix them. i then ran it again to verify and there were no errors. but even after this, i still get the panic message.

i have an 8gb clip v2

Blocks: 0x00F4FC00

[dfkt]

Originally Posted by Timar

Why are the TMD/IMD levels so much higher than on the OF (more than two orders of magnitude!!), even if the overall volume on rockbox is lower (on ox1d)!? Seems like a contradiction... are the settings > 0 handled different from the highest volume setting on the OF?

Anyway - I vote for setting the volume to 0x1e, if that value actually gives more volume. It comes at almost no cost!
In fact it would be way better than the current setting if it gives you adequate volume levels before reaching values > 0!!

From my understanding, 0dB on Rockbox is seen as the loudest possible setting the amp can handle, everything above it is overdrive, and saturation should be expected - so +6dB should give you twice the perceived loudness over 0dB, but at cost of increased distortion. It's the standard way headroom is handled in digital audio, 0dB being defined as the maximum.

At least that's how all my other rockboxed players measured as well, so I believe it is intentional. Having the Clip behave the same way would be fine, IMO.

But as I said, I didn't *hear* any increase in volume, testing stock 0x16 against 0x1d. Did anyone else perceive a higher max volume setting as louder?

[funman]

Originally Posted by marc2003

i ran chkdsk and it muttered something about some errors so i ran it again with the /f switch to fix them. i then ran it again to verify and there were no errors. but even after this, i still get the panic message.

i have an 8gb clip v2

Blocks: 0x00F4FC00

Looks alright (it's a number of 512-byte blocks, so it means ~8GB)

Can you add a line:

panicf("sd%d: %lu + %d > %lu", drive, start, count, card_info[drive].numblocks);

after "ret = -18" in firmware/target/arm/as3525/sd-as3525v2.c and rebuild rockbox?

btw do you use irc? i'm in #rockbox on freenode network

[Timar]

dfkt,

    /* We combine the mixer channel volume range with the headphone volume
       range - keep first stage as loud as possible */
    if (vol_r <= 0x16) {
        mix_r = vol_r;
        hph_r = 0;
    } else {
        mix_r = 0x16;
        hph_r = vol_r - 0x16;
    }

    if (vol_l <= 0x16) {
        mix_l = vol_l;
        hph_l = 0;
    } else {
        mix_l = 0x16;
        hph_l = vol_l - 0x16;
    }

which values did you change exactly? If you change every single 0x16 in the block above to a higher value, it won't have any effect, because you substract from hph what you've added to mix...

Change only the values in the if-condition and the mix_l/r and it will have a tremendous effect! With 0x16 my 56ohm headphones where rather silent @0db, with 0x1e they pound my eardrums into my brain

I'd love to do some measurements. Problem is: I don't have a fitting cable at hand...

[dfkt]

I did exactly what funman suggested. I changed mix_l and mix_r, nothing else.

Originally Posted by funman

If you want to change your rockbox build, edit firmware/drivers/audio/as3514.c function audiohw_set_master_vol and change "mix_r = 0x16" and "mix_l = 0x16" into any number. The value ranges from 0 (minimum volume) to 0x1f (31, maximum volume).

[Timar]

Originally Posted by dfkt

I did exactly what funman suggested. I changed mix_l and mix_r, nothing else.

Ok, would have been to easy a solution I guess Did you check the other values (0x1e, 0x1f) against the original volume?