| Message ID | CAB0OVGrz2nsYE73NnwjU48BMHvVvYOMgC-DobVVJPc=gFGN68g@mail.gmail.com |
|---|---|
| State | Accepted |
| Commit | ca781761d61de20f28b5c6486c64dbf2c2c4c944 |
| Headers | show |
| Series | [FFmpeg-devel] lavc/pnm: Support decoding gray float pbm images. | expand |
| Context | Check | Description |
|---|---|---|
| andriy/x86_make | success | Make finished |
| andriy/x86_make_fate | success | Make fate finished |
| andriy/PPC64_make | success | Make finished |
| andriy/PPC64_make_fate | success | Make fate finished |
This is variant of PFM images, PBM image format is only fixed point gray.
Am Fr., 12. Feb. 2021 um 00:18 Uhr schrieb Paul B Mahol <onemda@gmail.com>: > > This is variant of PFM images, PBM image format is only fixed point gray. Should I change the commit message or the patch? Thank you, Carl Eugen
On Fri, Feb 12, 2021 at 12:21 AM Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote: > Am Fr., 12. Feb. 2021 um 00:18 Uhr schrieb Paul B Mahol <onemda@gmail.com > >: > > > > This is variant of PFM images, PBM image format is only fixed point gray. > > Should I change the commit message or the patch? > commit message > > Thank you, Carl Eugen > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
Am Fr., 12. Feb. 2021 um 00:24 Uhr schrieb Paul B Mahol <onemda@gmail.com>: > > On Fri, Feb 12, 2021 at 12:21 AM Carl Eugen Hoyos <ceffmpeg@gmail.com> > wrote: > > > Am Fr., 12. Feb. 2021 um 00:18 Uhr schrieb Paul B Mahol <onemda@gmail.com > > >: > > > > > > This is variant of PFM images, PBM image format is only fixed point gray. > > > > Should I change the commit message or the patch? > > commit message Patch applied. Thank you, Carl Eugen
Why you use pbm decoder for probing of pfm files?
Am Mo., 15. Feb. 2021 um 02:39 Uhr schrieb Paul B Mahol <onemda@gmail.com>: > > Why you use pbm decoder for probing of pfm files? The only samples I found have the extension pbm and I simply misunderstood how you designed this since my samples were detected as pbm (because of their extension). Will fix, Carl Eugen
Am Mo., 15. Feb. 2021 um 19:13 Uhr schrieb Carl Eugen Hoyos <ceffmpeg@gmail.com>: > > Am Mo., 15. Feb. 2021 um 02:39 Uhr schrieb Paul B Mahol <onemda@gmail.com>: > > > > Why you use pbm decoder for probing of pfm files? > > The only samples I found have the extension pbm and I > simply misunderstood how you designed this since my > samples were detected as pbm (because of their > extension). > > Will fix, Carl Eugen AVParserContext->codec_ids is limited to five elements, therefore no additional AV_CODEC_ID can be added to the pnm_parser and AV_CODEC_ID_PFM cannot be used for the autodetection. Carl Eugen
then add new parser bellow On Mon, Feb 15, 2021 at 8:16 PM Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote: > Am Mo., 15. Feb. 2021 um 19:13 Uhr schrieb Carl Eugen Hoyos > <ceffmpeg@gmail.com>: > > > > Am Mo., 15. Feb. 2021 um 02:39 Uhr schrieb Paul B Mahol < > onemda@gmail.com>: > > > > > > Why you use pbm decoder for probing of pfm files? > > > > The only samples I found have the extension pbm and I > > simply misunderstood how you designed this since my > > samples were detected as pbm (because of their > > extension). > > > > Will fix, Carl Eugen > > AVParserContext->codec_ids is limited to five elements, > therefore no additional AV_CODEC_ID can be added to > the pnm_parser and AV_CODEC_ID_PFM cannot be used > for the autodetection. > > Carl Eugen > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
Am Mo., 15. Feb. 2021 um 20:18 Uhr schrieb Paul B Mahol <onemda@gmail.com>: > > then add new parser bellow Already working on it: ff_pnm2_parser or ff_pfm_parser? Or another idea? Thank you, Carl Eugen
ff_pfm_parser On Mon, Feb 15, 2021 at 8:29 PM Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote: > Am Mo., 15. Feb. 2021 um 20:18 Uhr schrieb Paul B Mahol <onemda@gmail.com > >: > > > > then add new parser bellow > > Already working on it: > ff_pnm2_parser or ff_pfm_parser? > Or another idea? > > Thank you, Carl Eugen > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
diff --git a/libavcodec/pnm.c b/libavcodec/pnm.c index aad23c7ae2..5af04965e7 100644 --- a/libavcodec/pnm.c +++ b/libavcodec/pnm.c @@ -72,6 +72,7 @@ int ff_pnm_decode_header(AVCodecContext *avctx, PNMContext * const s) s->bytestream[0] != 'P' || (s->bytestream[1] < '1' || s->bytestream[1] > '7' && + s->bytestream[1] != 'f' && s->bytestream[1] != 'F')) { s->bytestream += s->bytestream_end > s->bytestream; s->bytestream += s->bytestream_end > s->bytestream; @@ -82,6 +83,8 @@ int ff_pnm_decode_header(AVCodecContext *avctx, PNMContext * const s) if (buf1[1] == 'F') { avctx->pix_fmt = AV_PIX_FMT_GBRPF32; + } else if (buf1[1] == 'f') { + avctx->pix_fmt = AV_PIX_FMT_GRAYF32; } else if (s->type==1 || s->type==4) { avctx->pix_fmt = AV_PIX_FMT_MONOWHITE; } else if (s->type==2 || s->type==5) { @@ -177,7 +180,7 @@ int ff_pnm_decode_header(AVCodecContext *avctx, PNMContext * const s) if (ret < 0) return ret; - if (avctx->pix_fmt == AV_PIX_FMT_GBRPF32) { + if (avctx->pix_fmt == AV_PIX_FMT_GBRPF32 || avctx->pix_fmt == AV_PIX_FMT_GRAYF32) { pnm_get(s, buf1, sizeof(buf1)); if (av_sscanf(buf1, "%f", &s->scale) != 1 || s->scale == 0.0 || !isfinite(s->scale)) { av_log(avctx, AV_LOG_ERROR, "Invalid scale.\n"); diff --git a/libavcodec/pnmdec.c b/libavcodec/pnmdec.c index 9add5cfc84..4d5ce0bcb5 100644 --- a/libavcodec/pnmdec.c +++ b/libavcodec/pnmdec.c @@ -297,6 +297,30 @@ static int pnm_decode_frame(AVCodecContext *avctx, void *data, } } break; + case AV_PIX_FMT_GRAYF32: + if (avctx->width * avctx->height * 4 > s->bytestream_end - s->bytestream) + return AVERROR_INVALIDDATA; + scale = 1.f / s->scale; + if (s->endian) { + float *g = (float *)p->data[0]; + for (int i = 0; i < avctx->height; i++) { + for (int j = 0; j < avctx->width; j++) { + g[j] = av_int2float(AV_RL32(s->bytestream)) * scale; + s->bytestream += 4; + } + g += p->linesize[0] / 4; + } + } else { + float *g = (float *)p->data[0]; + for (int i = 0; i < avctx->height; i++) { + for (int j = 0; j < avctx->width; j++) { + g[j] = av_int2float(AV_RB32(s->bytestream)) * scale; + s->bytestream += 4; + } + g += p->linesize[0] / 4; + } + } + break; } *got_frame = 1;