[FFmpeg-devel] [PATCH] Common ACELP routines (2/3) - filters

[Diego Biurrun]

On Sun, Apr 27, 2008 at 01:05:07AM +0700, Vladimir Voroshilov wrote:
> Thank you for your review.

You are welcome.

> --- /dev/null
> +++ b/libavcodec/acelp_filters.c
> +
> + *   Similar filter is named b30 in G.729.

A similar

> + *   G.729 specification says:
> + *     b30 is based on Hamming windowed sinc functions, truncated at +/-29 and
> + *     padded with zeros at +/-30 b30[30]=0
> + *     The filter has a cut-off frequency (-3 dB) at 3600 Hz in the oversampled domain.

long lines, same in multiple other places

> + *   3600 - cut-off frequency
> + *   8000 - sampliny rate
> + *   k    - filter's order

filter order

> +    /* Lookup table contain values corresponding to -2/6 -1/6 0 1/6 2/6 3/6 fractions order.

The lookup table contains values corresponding to -2/6 -1/6 0 1/6 2/6 3/6 fraction order.

> +       Filtering process uses negative pitch lag offset, but negative offset should
> +       not be used in table.

The filtering process uses a negative pitch lag offset, but a negative
offset should not be used in the table.

> To avoid negative offset in table dimension corresponding to
> +       fractional delay following conversion applies:

To avoid a negative offset in the table dimension corresponding to
fractional delay the following conversion applies:

> +    //Make sure that pitch_delay_frac will be always positive

always be

> +    if(pitch_delay_frac < 0)
> +    {

I think Michael prefers braces on the same line.

> +        // Clippin is required to pass G.729 OVERFLOW test

ClippinG

> --- /dev/null
> +++ b/libavcodec/acelp_filters.h
> @@ -0,0 +1,139 @@
> +
> + * \brief Decode of the adaptive-codebook vector (4.1.3 of G.729)

s/of//

> + * \param ac_v [out] buffer to store decoded vector into

s/into//

> + * \param length amount of data to proceed

I think you mean "process" here.

> + * \return 1 if overflow occured, 0 - otherwise

occuRRed

> + * \param speech [in/out] speech data to proceed

ditto

Diego