[FFmpeg-devel] [PATCH] lavfi: Port fspp to FFmpeg

[Michael Niedermayer]

On Tue, Dec 23, 2014 at 03:52:57PM +0530, arwa arif wrote:
> On Tue, Dec 23, 2014 at 12:40 PM, Stefano Sabatini <stefasab at gmail.com>
[...]

> +static void mul_thrmat_c(FSPPContext *p, int q)
> +{
> +    int a;
> +    for (a = 0; a < 64; a++)
> +        ((int8_t *)p->threshold_mtx)[a] = q * ((int8_t *)p->threshold_mtx_noq)[a];//ints faster in C

these should be short not int8_t
short is generally 16bit, int8_t is 8bit
and the data in these arrays is in elements of 16bit not 8


[...]
> +static int config_input(AVFilterLink *inlink)
> +{
> +    AVFilterContext *ctx = inlink->dst;
> +    FSPPContext *fspp = ctx->priv;
> +    const int h = FFALIGN(inlink->h + 16, 16);
> +    const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
> +
> +    fspp->hsub = desc->log2_chroma_w;
> +    fspp->vsub = desc->log2_chroma_h;
> +
> +    fspp->temp_stride = FFALIGN(inlink->w + 16, 16);
> +    fspp->temp = av_malloc_array(fspp->temp_stride, h * sizeof(*fspp->temp));
> +    fspp->src  = av_malloc_array(fspp->temp_stride, h * sizeof(*fspp->src));
> +
> +    if (!fspp->temp || !fspp->src)
> +        return AVERROR(ENOMEM);
> +
> +    if (!fspp->use_bframe_qp) {
> +        fspp->non_b_qp_alloc_size = FF_CEIL_RSHIFT(inlink->w, 4) * FF_CEIL_RSHIFT(inlink->h, 4);
> +        fspp->non_b_qp_table = av_calloc(fspp->non_b_qp_alloc_size, sizeof(*fspp->non_b_qp_table));

the allocation isnt needed when fspp->qp == 0

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

If you think the mosad wants you dead since a long time then you are either
wrong or dead since a long time.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20141223/c7e11d93/attachment.asc>