On Sat, Mar 07, 2015 at 09:03:10AM +0000, Vittorio Giovara wrote:
> --- /dev/null
> +++ b/libavcodec/tdsc.c
> @@ -0,0 +1,585 @@
> +/*
> + * TDSC decoder
What does TDCS stand for?
> +/**
> + * @file
> + * TDSC decoder
> + *
> + * Fourcc: TSDC
> + *
> + * Codec is very simple:
> + * it codes picture by tiles, storing them in raw BGR24 format or
> compressing
TDSC is very simple. It codes
> + * Every frame is then packed with ZLib.
zlib
> +static av_cold int tdsc_init(AVCodecContext *avctx)
> +{
> + TDSCContext *ctx = avctx->priv_data;
> + int ret;
> +
> + ctx->refframe = av_frame_alloc();
> + ctx->jpgframe = av_frame_alloc();
> + if (!ctx->jpgframe)
> + return AVERROR(ENOMEM);
Care to check both allocations?
> + /* here BE bytestream is used, even though frame is stored LE */
> + switch (cursor_fmt) {
> + case 0x01010004: // old monochrome
> + case 0x20010004: // full colour - BGRA
> + case 0x20010008: // full colour - RGBA
I suggest an enum for these magic values.
> + /* Prepare a packet and send it to mjpeg decoder */
to the MJPEG decoder
Remember that we're currently teaching Sashka how to use English articles ;)
> + av_log(avctx, AV_LOG_ERROR,
> + "Jpeg decoding error (%d) for (%d) frame.\n",
JPEG
> +/* Parse frame and either copy data or decode jpeg. */
same
> +static int tdsc_decode_tiles(AVCodecContext *avctx, int number_tiles)
> +{
> + /* Restore context format and dimensions because of mjpeg overlap */
Same, while I'm at it ..
> + /* Decode alllll tiles in a frame */
aaaaallllll of them? :)
> +static int tdsc_parse_dtsm(AVCodecContext *avctx)
> +{
> + if (action == 2 || action == 3) {
> + ret = tdsc_load_cursor(avctx, action);
> + /* Do not consider cursor errors fatal unless explode */
unless in explode mode
Looks decent overall.
Diego
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
