Re: [mpc-general] Adding a lossless mode to MPC SV8

Hi,

On Sat, 2003-10-04 at 20:53, Menno wrote:
> > Third, I'd propose to separate entropy coding from the rest=3D20
> > (just as the psymodel would be modularized too). This way I=3D20
> > could implement a nice arithmetic coding layer for those who=3D20
> > want to save maximum space ;-)
>=20
> Arithmetic coding is adaptive, this will not save space, but increase =3D
> needed
> space. Because in an audio coder you only code small data sets (you wan=
t =3D
> the
> file to be randomly accessible) adaptive compression will not work.
> A well tuned set of Huffman tables will probably compress much better =3D
> and
> allow much faster encoding and decoding.

Arithmetic coding, just like Huffman coding, can use either an adaptive
or a static probability model (as long as encoder and decoder are
agreeing).

Huffman coding is a subset of Arithmetic coding, where all probabilities
must be inverse powers of 2. It's faster, but it IS less efficient.

If you use the same model for both, the seekability will be the same,
and Arithmetic coding will compress better. You only lose a bit of
speed.

Plus we can have more flexibility, ie: allowing some rare event to
happen at any time, and only wasting 1/2^32 bit (instead of 1 bit) when
it doesn't happen.

Regards
Jo=C3=ABl


>=20
> Menno
>=20
> >=3D20
> > Regards
> > Jo=3DEBl
> >=3D20
> >=3D20
> > On Sat, 2003-10-04 at 19:22, Christian HJ Wiesner wrote:
> > > Hi,
> > >=3D20
> > > as announced before, i would like to start a discussion=3D20
> > here about the
> > > above written subject.
> > >=3D20
> > > The reason i'm asking is, if we make the jump from SV7 to SV8, we=3D=
20
> > > should
> > >=3D20
> > > - guarantee backwards compatibiliy, means that old SV7=3D20
> > files should be
> > > convertable to SV8 ( with special tools ) in a lossless=3D20
> > process, but=3D20
> > > still be fully spec compatible SV8 files
> > >=3D20
> > > - make sure the SV8 specs offer a lot more functionailities=3D20
> > than SV7,=3D20
> > > so
> > > that MPC SV8 can be used for hi-quality movie encodings as=3D20
> > well as for=3D20
> > > DVDs, where space is maybe not a big issue anymore
> > >=3D20
> > > - consider to check if its possible to allow peeling for MPC SV8, s=
o
> > > that users can store their MPC files with rather high=3D20
> > bitrates on DVD or=3D20
> > > CD media, but can peel the files down in bitrate so they could be=3D=
20
> > > transferred onto portable players ( of course, this is NOT=3D20
> > a lossless=3D20
> > > process anymore )
> > >=3D20
> > > - add a loss less mode to SV8, and the speciality here is=3D20
> > that mixing
> > > lossy and lossless modes in one and the same file should be=3D20
> > possible,=3D20
> > > just like in the upcoming Wavpack4 from David Bryant ( who=3D20
> > i am copying=3D20
> > > on this email, i dont wnat to 'steal' his idea without at=3D20
> > least telling=3D20
> > > him :-)  )
> > >=3D20
> > >=3D20
> > > To make one thing clear out front, i dont know if Frank has thought
> > > about all these things in his current specs, because i am=3D20
> > sorry i am not=3D20
> > > capable reading technical specs at all, but from a user=3D20
> > perspective i am=3D20
> > > convinced these features could lift MPC on a new level, and=3D20
> > make a big=3D20
> > > differentation from other formats.
> > >=3D20
> > > What you think ?
> > >=3D20
> > > Christian
> > > MPC and matroska project admin
> > >=3D20
> > >=3D20
> > > http://mpc.corecodec.org
> > >=3D20
> >=3D20
> http://mpc.corecodec.org
>=20
>=20
> http://mpc.corecodec.org
>=20

http://mpc.corecodec.org