Rockbox 3.7 Stable Release Options

[Northpack]

Version 3.7 of the open source digital audio player software has been released.

Some new things since version 3.6:

• New skin format requiring all skins to be updated (details).
• Sansa Clip v1 and v2, Clip+, Fuze v2 and Ipod Nano 2g are now stable release targets.
• Sansa Fuze v1 and E200 v2 now have usb support enabled.
• Ipod 4g through 5.5g, Ipod Mini and Ipod Nano 1g now have usb support enabled and can charge via usb.
• Line out support for Fuzev1 and e200v2
• Greatly improved pitch accuracy on Sansa e200v2, Fuze, Clip and Clip+
• Increase maximum volume on the Fuze v2, Clip v2 and Clip+ to maximum allowed by the hardware
• New plugins: lrcplayer, shopping list, Resistor code calculator.
• Support for WMA Pro codec.

See full release notes: http://www.rockbox.org/wiki/ReleaseNotes37

This post has been edited by Northpack: Oct 31 2010, 18:52

[saratoga]

You beat me to it! I also want to add:

• Optimizations for FLAC, MP3, Vorbis, AAC-LC, AAC-HE, TTA, Cook, Atrac3 and WMA.

Additionally, while we didn't release 3.7 for Android, the port to that platform is now well along and rapidly becoming more usable. Hopefully it'll be ready for the next release.

[Cynic]

Playback of Musepack SV8 files appears to be broken. The system (IAudio X5V) will consistently freeze at the end of some .mpc files and I have to turn it off using the power button.

Musepack files worked fine with release 3.6 but stopped working sometime between the July 11 and October 16 daily builds (the July one still played MPC files correctly but the October one didn't). I was hoping it was some temporary bug that would be fixed on the newest release but as it turns out it didn't. Are the developers aware of this issue?

[saratoga]

Are the developers aware of this issue?

No, no one has reported any issues with MPC, and all my test files decode correctly. File a bug report or its probably not getting fixed.

[Lazpete]

Good news!

What exactly does the line "Greatly improved pitch accuracy on Sansa e200v2, Fuze, Clip and Clip+" mean?

Lasse

[DonP]

Good news!

What exactly does the line "Greatly improved pitch accuracy on Sansa e200v2, Fuze, Clip and Clip+" mean?

Lasse

The OF for at least Fuze and Clip has pitch issues. Varies with model, but my V1 clip played 0.7% fast/sharp.

My v1 e200 (uses a different chip than those listed) is at least 10x closer.

[Northpack]

You beat me to it!

Sorry, I always wanted to file a news item

Good news!
What exactly does the line "Greatly improved pitch accuracy on Sansa e200v2, Fuze, Clip and Clip+" mean?
Lasse

The OF for at least Fuze and Clip has pitch issues. Varies with model, but my V1 clip played 0.7% fast/sharp.

3.7 enables a special PLL clock providing an optimal ratio for 44.1kHz PCM playback on ASMv2 devices (Fuzev2, Clipv2, Clip+). AFAIK all those devices played 1.1% flat with rockbox, with 3.7 the pitch error is reduced to 0.04% (which is more precise than the original firmware).

Edit: here's a measurement showing the pitch error on Clip+ with old/new rockbox and OF

This post has been edited by Northpack: Nov 1 2010, 14:16

[Lazpete]

You beat me to it!

Sorry, I always wanted to file a news item

Good news!
What exactly does the line "Greatly improved pitch accuracy on Sansa e200v2, Fuze, Clip and Clip+" mean?
Lasse

The OF for at least Fuze and Clip has pitch issues. Varies with model, but my V1 clip played 0.7% fast/sharp.

3.7 enables a special PLL clock providing an optimal ratio for 44.1kHz PCM playback on ASMv2 devices (Fuzev2, Clipv2, Clip+). AFAIK all those devices played 1.1% flat with rockbox, with 3.7 the pitch error is reduced to 0.04% (which is more precise than the original firmware).

Edit: here's a measurement showing the pitch error on Clip+ with old/new rockbox and OF

Thx guys, I got it!

Lasse

[saratoga]

Are the developers aware of this issue?

No, no one has reported any issues with MPC, and all my test files decode correctly. File a bug report or its probably not getting fixed.

Thanks for the test file. It appears that some MPC SV8 files have trouble on the X5, H100 and H300. Other files work fine, and everything seems to work normally on other targets. Once this is fixed you'll probably want to update to a post 3.7 build.

[eevan]

There were some problems with SV8 on Fuze. Basically, targets with small RAM can suffer data aborts with SV8. Andree fixed that, at least on Fuze.

[Cynic]

Thanks for the test file. It appears that some MPC SV8 files have trouble on the X5, H100 and H300. Other files work fine, and everything seems to work normally on other targets. Once this is fixed you'll probably want to update to a post 3.7 build.

No problem, thanks for looking into this. I apologise for being lazy and not reporting the bug earlier, it was the first time Rockbox misbehaved since I started using it more than 3 years ago and as such I was confident that someone would notice it and fix it before the next stable release. I'd be happy to help as best as I can.

This post has been edited by Cynic: Nov 1 2010, 20:40

[saratoga]

Looks like this is fixed in the current (post 3.7) build.

[Soap]

Just to quickly chime in...
The plan is, starting this release cycle, to back-port important bug fixes to the stable branch and issue point releases as soon as reasonably possible. I have little doubt this would make it into 3.7.1 how soon that will be (this being the first attempt at such a plan) I can't say.

[Larson]

3.8 Stable is out now!