[Carlo Sirtori]

Wow, thanks so much for these awesome advices.

I’m starting to think that we will give up the option of using an external recorder. The director wants to see the rushes with a rough edit: that would put us one day behind no matter what.

What you say about the QT container that most of the recorders use to wrap DNxHD is very useful information, nevertheless, I better keep that in mind for the future.

I guess the sync will be done in MC (Mix only). It should not be too much of a hassle creating offlines of all the shots. The LUT will be general – maybe something more sophisticated than a Rec709 LUT.

About the sound, I don’t have a precise idea yet, I still have to talk with the sound people.

This is an independent film, there is some money (not run and gun style), but it’s not Hollywood, so I consider with attention the most expensive options.

I’m used to S3D production, with huge amount of datas and any kind of sync issues to solve ahead. I am trying to get rid of some the reflexes I acquired – like those related to the huge time that data transfer + transcoding take.

Resolve transcode, after what you told me, seems a pretty good option.

[Carlo Sirtori]

Hello Michael,

First off, thanks for your remarks, those really help!

Things just got clearer today: there was a meeting and this is what came up.

The workflow would be: record in Prores 4444 Log C; go through DaVinci Resolve to create offlines in DNxHD with a LUT rec709; edit in Avid.

My (bad) idea of recording in DNxHD 444 was to avoid the transcoding step before editing – and edit natively the rushes. But we would lose too much giving up the Prores (12 bits + 2K format) and we would “gain weight” (as you point out 440Mb/s in DNxHD vs 330Mb/s in Prores). Plus, I’m not sure that you can always edit real time in Avid when editing with those big rushes.

Now, I’m still not sure that the director, that might like to see the rushes in Avid after each day of shooting, will like that transcoding phase in DaVinci.

There is the possibility of using an external recorder and, with the Alexa Rec Out, record a proxy in DNxHD 36 with a LUT Rec709.
If we do it, it depends on the budget and the confort of the operator, I guess.

I would really appreciate your advice on this all.

Thanks again,

Carlo

[Carlo Sirtori]

Thanks for your quick answers!

Actually, the 3D-rig is more complex than what I said. I didn’t want to bore you with the details. I based the 3D-rig on a tutorial my company bought online, readapting it to my needs (the 3D-rig in the tutorial had interaxial and convergence related, and that’s a real nonsense in S3D shooting). My 3D-rig is a four comps thing, where I have a “central” camera that I move around for animations, I place layers etc. Then I have a left and a right eye comps (with two cameras linked to the “central” one, and the central comp embedded in them). And finally there’s a stereo control comp, where I see the left and right comps fused in anaglyphs and where I can move the interaxial and the convergence, ie change the depth of the scene and its global placement back/forth. I said that my 3D-rig was “basically two cameras linked to a null”, it’s somewhat true… except the two cameras don’t share the same comp, nor does the null. To do a 3D-rig “the easy way”, but probably not the best way, one could create the animation in one comp, and call it “left eye”. Then duplicate that comp and change the camera’s X values to “+something” and call it “right eye”. Render the two eyes and set the convergence in a third comp or in any editing software (it’s a simple shift of one eye as regards to the other – they call it HIT, “horizontal image translation”).

Anyway, for my noobish issue, I’ll try the “Scale Composition script” for sure (I just have to find what it is first, but I’m sure I’ll find the info in the forum… or in the manual!).

I hope I can adapt Z positions without touching X and Y (that are automatically adapted when resizing the comp).

Thanks!

Carlo