[flac-dev] FLAC specification clarification

[Martijn van Beurden]

Okay, what about one of these lines in the specification?

<5> Quantized linear predictor coefficient shift needed in bits (NOTE: this
number is signed two's-complement, but should be positive, as no known
encoders or decoders have support for negative shifts)
<5> Quantized linear predictor coefficient shift needed in bits (NOTE: this
number is signed two's-complement, but should be positive, as many decoders
only support a positive value here)
<5> Quantized linear predictor coefficient shift needed in bits (NOTE: this
number is signed two's-complement, however many decoders only support a
positive value)
<5> Quantized linear predictor coefficient shift needed in bits (NOTE: this
number is signed two's-complement, though all known decoders only support a
positive value)

I just did a quick search to find open-source decoders and check them. All
of them either throw an error or produce an incorrect result.
libFLAC, throws error:
https://github.com/xiph/flac/blob/master/src/libFLAC/stream_decoder.c#L2663
ffmpeg, throws error:
https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/flacdec.c#L380
Rockbox firmware, throws error:
https://github.com/Rockbox/rockbox/blob/master/lib/rbcodec/codecs/libffmpegFLAC/decoder.c#L248
Flac.js, throws error:
https://github.com/audiocogs/flac.js/blob/master/src/decoder.js#L323
DrFlac, assumes positive shift:
https://github.com/mackron/dr_libs/blob/master/dr_flac.h#L2958
Decoder found at https://www.nayuki.io/page/simple-flac-implementation
assumes positive shift
Flac-library-java, throws error:
https://github.com/nayuki/FLAC-library-Java/blob/master/src/io/nayuki/flac/decode/FrameDecoder.java#L300
jflac, assumes positive shift:
https://github.com/nguillaumin/jflac/blob/master/jflac-codec/src/main/java/org/jflac/frame/ChannelLPC.java#L77
Claxon, throws error:
https://github.com/ruuda/claxon/blob/aaaf40bcbfea01b6243e4699358b25b991e0727e/src/subframe.rs#L529

Kind regards,

Martijn van Beurden



Op vr 26 jun. 2020 om 08:57 schreef Thomas Zander <
thomas.e.zander at googlemail.com>:

> On Fri, 26 Jun 2020 at 05:15, Brian Willoughby <brianw at audiobanshee.com>
> wrote:
>
> > That said, there's nothing wrong with adding a note to the specification
> about the common implementations, particularly the reference library. Then,
> future developers will know both the precise specification and still have
> the warning that they risk incompatibility by deviating from the reference
> implementation.
>
> This sounds reasonable to me. The wording of this note should be very
> clear, though.
>
> Thomas
> _______________________________________________
> flac-dev mailing list
> flac-dev at xiph.org
> http://lists.xiph.org/mailman/listinfo/flac-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xiph.org/pipermail/flac-dev/attachments/20200626/3f0a2a7d/attachment.html>