fiafcore:Codec

[paulduchesne]

Definition

From the Moving Image Cataloguing Manual pg. 75:

The video compression standard used in the digital video Item, for example, D10, MP4, etc.

From the Moving Image Cataloguing Manual pg. 75:

The audio compression standard used in the audio file, whether it is contained in a wrapper with a video file or as a standalone audio file.

D.7.10 Item Codec type contains a vocabulary for this concept, with the following terms:

AVC (VIDEO)
D10 (VIDEO)
NIL ProRes 422 (HQ) (VIDEO)
S10 (VIDEO)
JPEG2000 (VIDEO)
MPEG-4 H264/AVC Main Profile (VIDEO)
AIFF (AUDIO)
WAV (AUDIO)

Vocabularies

BFI

[Item] xpath .//code_type

term	count
ProRes 422 (HQ)	2095
MPEG-4 AVC	1027
D10	603
JPEG2000	484
TIFF	460
NIL	298
ProRes 422	296
V210	200
Uncompressed	149
FFV1 v3	141
ProRes 4444	43
ProRes 4444 XQ	15
Mixed	7
WAV	7
MPEG-2	4
AVC-Intra 100	2
AVC	2
DV	1
Avid DNxHD	1

Bundesarchiv

[Item] xpath .//ba:Exemplar/ba:CodecBild

term	count
ohne Bild-Codec	3833
DV (SD)	2472
H.264/MPEG 4 AVC	2353
unbekannt	1549
H.262/MPEG 2 XDCam IMX 50	1327
MPEG-2	799
DNxHR 444	794
JPEG 2000	351
DNxHD 444	298
VP 8	87
JPEG 2000 5/3	62
ProRes 422 Proxy	50
DVD MPEG2 TS	46
MPEG-4	37
IMX 50 (XDCAM 422)	17
ProRes 422	17
ProRes	7
ProRes 422 HQ	7
JPEG 2000 9/7	3
H.264/MPEG4 AVC Proxy HQ	2
MPEG-1	1

[Item] xpath .//ba:Exemplar/ba:CodecTon

term	count
AAC	1339
linear PCM	1294
Broadcast PCM WAV	773
PCM	516
ohne Codec	141
Ogg Vorbis	87
AC 3 (Dolby)	20
MPEG-4 Part 3	5
AC3	2
ATRAC (SDDS)	1

Recommendation

Do not include vocabulary for the next iteration.

[heftberger]

Wondering if keeping a controlled list would have the problem of being outdated very quickly. I see two ways:

• treating this like we could treat bit depth (no list, just values)
• go for umbrella terms to contain some level of granularity, like prores, mpeg, h264 (well, we already have 265, 266 theoretically), DNxHD, jpg2000, dv, tiff, ffv1, other?
  What do you think?

[paulduchesne]

Using values rather than entities would make life easier in the short term ("strings not things!"), but I would be concerned that this would allow a plurality of terms for what is ultimately the same thing. As an example, listed synonyms for wd:Q212633: MPEG-4 Part 10, Advanced Video Coding, MPEG-4 Advanced Video Coding, ISO/IEC MPEG-4 AVC, MPEG-4 Part 10, MPEG-4 AVC, H.264/MPEG-4 Part 10, H.264/AVC, H.264, h264, H264, H.26L, AVC, H.264/ISO/IEC 14496-10—MPEG-4 Part 10 Advanced Video Coding, H.264/MPEG-4 AVC... Also noting that presumably tools like MediaInfo already have a standardized vocabulary in place?

My current thinking is whether we should park these granular digital properties for the time being, and look instead to federate ontologies with the OWL version of PREMIS, which already includes some fairly mature structures for streams/containers, PRONOM, etc.

Keen for other opinions on this, as I also do absolutely see the value in supporting queries for, eg, "how much of the community is using FFv1", "is my digitised copy of a film using a lossless/lossy codec as opposed to yours", but I think for now I would label as exclude, unless anyone feels strongly otherwise that this needs to be included from the start.

[heftberger]

I agree: using an existing list would be useful.