librespot 0.3.0 to be released (soon)

[Tim Curtis]

I looked at their repo and they have a dev and a master branch. They prolly create releases like we do by tagging the dev branch and then merging dev into master. I think building against the dev sources like you did is the way to do it.

I unzipped the dev branch and searched for 0.3.0 and it exists in the three files below. Looks ok to me.

CHANGELOG.md
Cargo.toml
Cargo.lock

[roderickvd]

Hi guys! First post here.

One of the librespot maintainers @roderickvd has kindly submitted a PR for moOde that will update to upcoming librespot 0.3.0 :-) You can view the PR here https://github.com/moode-player/moode/pull/445

When its released by the librespot project I'll compile it and get it out to the Test Team for eval. If all looks good then it will be include in next moOde release.

Thanks for that. Admittedly it's a bit scratching my own itch as I'm a happy moOde user.

Over at the Librespot Project there's one PR that I forgot to merge before releasing 0.3.0. It causes the last track of an album to be skipped when autoplay is enabled, which is a bit ugly. I hope to collect some more reports about 0.3.0, fix any glaring issues and then release 0.3.1 shortly. If you could let me know before releasing 7.5.0 then we can synchronize and cook up 0.3.1 at least before that.

So in a fit of boredom I built the current dev branch of librespot. It's working ATM as a drop-in replacement in my moOde 7.4.1 player on an RPi4B with no adjustment of existing invocation parameters. Still claims to be v0.2.0 but the binary is about 3 percent bigger. 

Not sure what to look for in testing.

Yes, the existing command line options were kept intact, so it's a drop-in replacement. There are a couple of deprecated ones concerning the Alsa mixer but moOde doesn't use those, as well as a couple of new ones that I introduced in the PR above.

I'm pretty excited about 0.3.0 and think it has a lot to offer. There's a full changelog here: https://github.com/librespot-org/libresp...ANGELOG.md but if you ask me these are the highlights:

1. Dithering has been introduced, which lowers quantization error and intermodulation distortion particularly with 16-bit output to Alsa (which is the default).

2. The entire sample pipeline from the decoder output up to the audio backend is now in 64-bit floating point which means that normalization and volume control are now next to lossless in quality (of course Ogg Vorbis remains lossy, but I'm sure that distinction is clear).

3. There's a cubic volume control which can appeal to those who want a larger "usable" range in the slider.

4. Volume normalization by default now automatically picks track or album mode depending on what's playing.

5. The Alsa backend is a bit more efficient by configuring the buffer size better.

And there's a lot more. Happy listening and thanks for testing.

[roderickvd]

I looked at their repo and they have a dev and a master branch. They prolly create releases like we do by tagging the dev branch and then merging dev into master. I think building against the dev sources like you did is the way to do it.

That's right.

[Tim Curtis]

Compiled and testing 0.3.0 and looking good :-)

The 750 release is nearing completion and so I'd like to ship it at end of October if possible.

-Tim

[roderickvd]

I do believe you have an image of being fast :-) kidding aside that'll be fine, I'll look to get 0.3.1 out of the door next week. Hopefully that won't keep people from testing now so can iron any other wrinks out.

[TheOldPresbyope]

@roderickvd

That's a great set of highlights!

Regards,
Kent

[roderickvd]

Barring any last-minute showstoppers, I plan to release 0.3.1 the coming few days.
Besides fixing that bug I mentioned earlier, it also improves CPU usage of dithering by ~35%.

[Tim Curtis]

Nice. I'll plan to bump to the new version before releasing 750 :-)

[roderickvd]

Just released 0.3.1!

[Tim Curtis]

Nice, I'll compile and test it next week.