[CODEC] LTC support

[sobotka]

It doesn’t appear that Olive’s master supports LTC? Timecode support across frames for use with including reading and writing metadata, LTC, burn in, etc.

[alcomposer]

https://github.com/x42/libltc good choice?

[sobotka]

That seems the only choice, and indeed appears to work.

[alcomposer]

#629 is only one aspect of this request, setting the timecode of clip from metadata already present in clip.

I interpreted this request as being for LTC as outlined here: https://en.wikipedia.org/wiki/Linear_timecode

Timecode embedded in the audio track of a media file can allow cameras that are not able to receive timecode input to be synced together: https://www.cinema5d.com/get-your-multi-camera-dslr-or-mirrorless-shoots-in-sync-with-ltc-convert/

It would be fantastic, even for my workflow, to send each camera a timecode audio signal, and allow Olive to easily translate that into a start timecode of each media clip. This would allow clips to sync together without having to resort to brute-force audio waveform sync. (I also would like that as well, but its just another way to sync clips).

For cameras that are unable to be synced - (they don't have a Timecode in port) there will be a miss-match between embedded LTC & metadata timecode. So Olive will also need a toggle (probably in clip settings) which timecode to use: Metadata or LTC.

I don't have access to ENG cameras, but I believe that they do embed LTC in either an audio stream- or a burst of timecode in the main audio stream.

For the record I fully understand that Timecode sync is not genlock, and the frames of such multi-camera setups will not be matched. However in reality this is only an issue for Stereo 3D, and to a lesser extent live vision mixing (given the ability for many switchers to compensate for frame timing mismatch).

[sobotka]

Timecode is Timecode, and supporting is supporting.

[alcomposer]

What I am working on currently is not LTC timecode, but metadata timecode. Hence why I made a different request just for that.

[sobotka]

Timecode is timecode. Referencing it as data is part of it. Is it a stamp? What does the stamp represent? Is it an audio stream of synced timecode? What does the timecode represent? It’s all part of that umbrella because ultimately it binds to a frame.

[alcomposer]

There are many NLE's that don't support LTC embedded in audio stream, yet support metadata timecode.

Otherwise this software would not exist: https://www.videotoolshed.com/product/ltc-convert-auxtc/

And yes, I think both solutions should be supported by Olive. The user should be able to decide which timecode to use.

I closed #629, lets use this first request for any future updates, including timecode metadata.

Aspects to this FR:

• Read LTC audio timecode & metadata timecode in imported media
• Write metadata timecode to exported media (allow user to choose custom timecode start- or use selected region timecode)
• Display correct timecode in media viewer (currently media viewer shows 0 timecode start)
• Allow user to select media viewer timecode display (start at 0 or start at media timecode)
• Allow user to choose between using metadata timecode / LTC embedded audio timecode for clip (useful for clip sync in future)

@sobotka please add any more detail if I have missed something out

[alcomposer]

@x42 Are you aware of any ENG/Cinema/Consumer cameras that record LTC audio? Either as full audio track or a bust of timecode for the first frame? (I am not sure the latter exists- but I have seen it mentioned)

My limited personal camera research shows that all the below cameras record metadata as timecode tag on various streams (GH5 & XT3 All streams). The Sony A7Riii looks like it has a data stream that not only has a timecode tag, but also has data inside (204 kb/s).

Sony A7Riii

    Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuvj420p(pc), 1920x1080 [SAR 1:1 DAR 16:9], 15408 kb/s, 25 fps, 25 tbr, 25k tbn, 50 tbc (default)
    Metadata:
      creation_time   : 2019-03-03T04:51:19.000000Z
      handler_name    : Video Media Handler
      encoder         : AVC Coding
    Stream #0:1(und): Audio: pcm_s16be (twos / 0x736F7774), 48000 Hz, 2 channels, s16, 1536 kb/s (default)
    Metadata:
      creation_time   : 2019-03-03T04:51:19.000000Z
      handler_name    : Sound Media Handler
    Stream #0:2(und): Data: none (rtmd / 0x646D7472), 204 kb/s (default)
    Metadata:
      creation_time   : 2019-03-03T04:51:19.000000Z
      handler_name    : Timed Metadata Media Handler
      timecode        : 08:57:23:01

Fuji XT3

    Stream #0:0(eng): Video: hevc (Main 10) (hvc1 / 0x31637668), yuv420p10le(pc, smpte170m/bt709/smpte170m), 3840x2160, 407605 kb/s, 25 fps, 25 tbr, 25k tbn, 25 tbc (default)
    Metadata:
      creation_time   : 2019-03-04T22:50:04.000000Z
      timecode        : 02:39:13:00
    Stream #0:1(eng): Audio: pcm_s24le (lpcm / 0x6D63706C), 48000 Hz, 2 channels, s32 (24 bit), 2304 kb/s (default)
    Metadata:
      creation_time   : 2019-03-04T22:50:04.000000Z
      timecode        : 02:39:13:00
    Stream #0:2(eng): Data: none (tmcd / 0x64636D74), 0 kb/s (default)
    Metadata:
      creation_time   : 2019-03-04T22:50:04.000000Z
      timecode        : 02:39:13:00

Panasonic GH5

    Stream #0:0(und): Video: h264 (High 4:2:2) (avc1 / 0x31637661), yuv422p10le(pc, bt709), 1920x1080 [SAR 1:1 DAR 16:9], 100284 kb/s, 25 fps, 25 tbr, 90k tbn, 50 tbc (default)
    Metadata:
      creation_time   : 2019-02-23T14:45:49.000000Z
      timecode        : 02:23:23:13
    Stream #0:1(und): Audio: pcm_s16be (twos / 0x736F7774), 48000 Hz, 2 channels, s16, 1536 kb/s (default)
    Metadata:
      creation_time   : 2019-02-23T14:45:49.000000Z
      timecode        : 02:23:23:13
    Stream #0:2(und): Data: none (tmcd / 0x64636D74), 0 kb/s (default)
    Metadata:
      creation_time   : 2019-02-23T14:45:49.000000Z
      timecode        : 02:23:23:13

[lutzray]

I'm a git noob... should I start a new feature request or may I clarify/augment this one?

What those manufacturers call timecode isn't timecode (as defined by eg, Wikipedia) and aren't even less SMPTE LTC, as mentioned above by alcomposer. Those are really timestamps , but that doesn't negate their usefulness: what would be great (and I wonder why it's not in every LNE) is this timeline action:

Align Clip Using Metadata Timestamp

This is far simpler to implement than linking to Robin Gareus's libltc excellent library. So, in short: the acronym LTC should be dropped from this Feature Request: they're two different beasts.

Oups... Rereading this thread I realized I want to reopen Set timecode of clip from metadata only #629 Why has it been closed? Haven't FRs less chance to be implemented if we put them in the same bag? As I said, I'm a complete stranger to CVS communities... sorry to add some noise.

[x42]

@x42 Are you aware of any ENG/Cinema/Consumer cameras that record LTC audio?

I've worked with footage from various ARRI cameras in the past that provided LTC (as do most RED). One usually syncs the cameras twice a day with a TC source and drift is negligible, but neither are consumer devices.

With some pro-sumer device I've used an audio-cable from a LTC source to the camera. Fine with a stationary camera, but most camera operators would probably be annoyed.