On Tue, Feb 28, 2012 at 06:16:59PM -0800, Mashiat Sarker Shakkhar wrote:
> ----- Original Message -----
>
> > From: Diego Biurrun <[email protected]>
> > To: libav development <[email protected]>
> > Cc:
> > Sent: Tuesday, February 28, 2012 8:02 PM
> > Subject: Re: [libav-devel] [PATCH] Windows Media Audio Lossless decoder
> [...]
> >> +static void decode_cdlms(WmallDecodeCtx *s)
>
> >
> > And what is cdlms anyway? You have so many Doxygen comments, but
> > this is not explained anywhere.
>
> CDLMS is like a regular LMS filter, but only operates on a single channel
> to reduce intra-channel redundancy. The exact acronym seems to be M$-
> specific; and it expands to "Cascaded Least Mean Squared" (found it in an
> M$ patent).
>
> No need to explain that in the code, anyone who knows audio codecs knows
> what an LMS filter is. (I don't know audio codecs, and even I know what
> an LMS filter is.) The multi-channel variant of the above is called MCLMS
> in this code and in the spec.
>
> [...]
> >> + s->ave_sum[ch] = residue + s->ave_sum[ch] -
> >> + (s->ave_sum[ch] >> s->movave_scaling);
> >
> > Indentation is off.
> [...]
> >> + for (icoef = 0; icoef < s->cdlms[ich][ilms].order; icoef++)
> >> + pred += s->cdlms[ich][ilms].coefs[icoef] *
> >> + s->cdlms[ich][ilms].lms_prevvalues[icoef + recent];
> >
> > Indentation is off.
>
> For a moment, I thought that I understood what you mean, but the two seems
> to disagree. Indent it with reference to what? The '+=' or 's->' or 'pred'?
to s->
i.e.
pred += s->cdlms[a][b][c] *
s->cdlms[a][b][c];
or
pred += s->cdlms[a][b][c]
* s->cdlms[a][b][c];
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
