[Flac-dev] Re: process for petitioning for more flac hardware support

Josh Coalson xflac at yahoo.com
Wed Jun 26 13:37:04 PDT 2002

--- Andre <armcc at lycos.com> wrote:
> > 
> > i was wondering how one would go about approaching hardware
> > manufacturers to support flac.
> > 
> One thing to do first might be to make sure that flac is suitable to
> run on lost cost hardware. For example the amount of buffering
> required
> in the decoder.
> If I understand the flac format correctly, all decoders must support
> block sizes upto 65535 samples (ie 256k of decoded stereo data).

FLAC specifies a subset of the whole format that make it
more amenable to small hardware implementations.  The command-
line encoder currently generates subset files, with a max
block size of 4308 samples unless you explicitly give it a
different number.  So while it's possible to have 64k samples
in a block, you have to force a large blocksize when encoding,
then knowingly tell the encoder via another call
(FLAC__stream_encoder_set_streamable_subset(encoder, false))
which is what 'flac --lax' does, that you are straying from the

There are warnings in the docs about the consequences of
--lax but maybe it's not clear enough.

> Maybe it is always possible to play audio from a flac block without
> buffering and decoding the whole block first (I don't know - I
> haven't
> studied the source) but some information covering this type of issue
> on
> the flac website might be a good starting point for persuading HW
> manufacturers.

This is possible as long as you don't care about checking the
data against the CRC before playback.


Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup

More information about the Flac-dev mailing list