Hi i like editing for music brainz but its annoying when i cant find a jpg for an album cover. i would love it if you guys added webp to your website
Do you have examples of releases where you found cover art as webp files?
I never came across any.
And remember that you must upload the exact 100% correct photo of the actual release, not just a roughly matching image found somewhere.
Google Play showed cover art as WebP files by default. It took some trickery to extract JPEG versions. I’m not aware of any other website or service that uses the format, though we did add support for it in Picard not too long ago.
I come across it every now and again (I add a fair bit of cover art, some historic/I can’t scan the original), and it is annoying to convert. Support would be excellent. I don’t know how straightforward it is, with IA storing the images?
There is already a ticket here I see! If you can vote for it AmH007 that helps, apart from that you can just hope that a developer takes an interest
Their to-do/wishlist is verrryyy long.
For MB, it’s nearly a non-issue. I’m not 100% familiar with MBS, but as far as I can tell, it’d only require adding a new file type row in one of the DB tables, and some lines of JS code to recognise WebP based on file extensions and magic numbers.
On IA’s side, it’s a different story. I tested, IA doesn’t recognise WebP images at all, they remain “Unknown”. I presume it wouldn’t be too difficult for them to add recognition for those, though. The real problem would be the derivation of thumbnails, that would probably require some larger changes on their end.
Most sites that implement webp will respect headers and send you the source image if you don’t have ‘image/webp’ in your accept header. Use curl or something.
Apple Music covers will also download as WebP.
If we’re chiming in in support of additional image formats, I would personally like AVIF support. In my tests it saved 20-30% of the file size (in lossless mode) for 600 DPI booklet scans when compared to the same scans saved as PNG.
I would be interested to know those tricks to extract a jpg from a webp, could you explain here?
If you download the precompiled WebP utilities, you can run
dwebp -o out.png in.webp (might be something like
dwebp.exe /o ... on Windows) to convert it to PNG.
I don’t think that
dwebp supports JPEG output, so you’ll need to run something else afterward if you need that (on Linux, I use
convert out.png out.jpg from ImageMagick).
(And then there are also a bunch of websites of varying dodginess that can do this for you.)
An intelligent website only serves WebP if your browser supports it, and can fall back to JPEG (or potentially another format). If you fetch the URL with an old browser or with a command-line tool like wget, you might get an original JPEG. I use wget for this and it works a treat.
I’m afraid that it would be a re-compressed image, then.
Sorry I’m off-topic:
What’s the point for a website to use WebP, then, if they have to host a duplicate to be fully compatible?
Sounds more efficient to only host most compatible formats, no?
Sorry I don’t know much about WebP.
The more common way I’ve seen to support fallback for older browsers (or ones shipped by Apple ) is to use
<srcset> elements to specify multiple resources and let the browser download whichever one that they think is best.
Serving modern formats results in a better experience for users who can use them (shorter load times, higher quality), but yeah, preserving support for old browsers is a pain.
Bringing this back to the original topic, I agree with @ROpdebee that making the CAA support transcoding between formats could be a lot of work. (Personally, I’ve been annoyed by there not being many native WebP encoding libraries – Go, Rust, and Java all just wrap the
libwebp library, which makes deployment painful. I think that the situation is similar with AVIF.)