DSLR Audio Time Code/Workflow

[Brent Lestage]

Sorry guys... More DSLR time code questions (ad nauseam)...

Working with a Nikon DSLR using 24p frame rate... Planning double-system sound with smart slate... Post workflow is structured so that the project will be edited in 24p timeline... files will be converted first with ProRes then imported into FCP...

Should audio time code be set at 29.97ND? Production states that their last mixer used 24 (23.976?) and time code began drifting 2 frames four-minutes into a scene...

Your thoughts?

Brent

[Philip Perkins]

Sorry guys... More DSLR time code questions (ad nauseam)...

Working with a Nikon DSLR using 24p frame rate... Planning double-system sound with smart slate... Post workflow is structured so that the project will be edited in 24p timeline... files will be converted first with ProRes then imported into FCP...

Should audio time code be set at 29.97ND? Production states that their last mixer used 24 (23.976?) and time code began drifting 2 frames four-minutes into a scene...

Your thoughts?

Brent

If he used 23.976 and it drifted then the Nikon is a true 24, and that is probably the fps you should use?  With FCP the TC of the file is mostly important for FCP interpreting what sample rate to play the file back at--for picture syncing it is much faster to ignore TC and use a hard mark (like a slate).  Also helpful, esp if they are trying to use an app like PluralEyes, is having a clean scratch audio track on the cam files.  I should say that I haven't done a job w/ a Nikon yet, so don't know for sure that their recording method does hold sync.  Have they done any double system testing with it?  Might be a good idea--I'd love to hear the results.  Suffice to say that how the session in the FCP is initially set up is very important....

Philip Perkins

[Marc Wielage]

Always, always test it first before going into the battlefield. One five minute sync test, with camera and sound running together, and a clap at the head and tail, will confirm for sure if the set up is right. A timecode slate is pretty much mandatory, because none of these still cameras can accept external code (to my knowledge, anyway).

My guess is that Phil is right and the previous crew had a classic .1% sync drift error. 24.00 TC (or 30.00NDF) should sync up fine with 24fps video. I checked the Nikon site here:

http://imaging.nikon.com/products/imaging/lineup/brochure/index.htm

and it seems to me that their cameras are shooting at true 24.00fps. There's no big deal with this, except they'll have to let post fix it for final delivery for cable, network, or Blu-ray, which all require 23.98 for HD. If it's just a demo or for the web, it can stay at 24.00.

--Marc W.

[Philip Perkins]

Always, always test it first before going into the battlefield. One five minute sync test, with camera and sound running together, and a clap at the head and tail, will confirm for sure if the set up is right. A timecode slate is pretty much mandatory, because none of these still cameras can accept external code (to my knowledge, anyway).

My guess is that Phil is right and the previous crew had a classic .1% sync drift error. 24.00 TC (or 30.00NDF) should sync up fine with 24fps video. I checked the Nikon site here:

http://imaging.nikon.com/products/imaging/lineup/brochure/index.htm

and it seems to me that their cameras are shooting at true 24.00fps. There's no big deal with this, except they'll have to let post fix it for final delivery for cable, network, or Blu-ray, which all require 23.98 for HD. If it's just a demo or for the web, it can stay at 24.00.

--Marc W.

All the local 5D stuff has ended up doing all post at 30 and then converting to whatever is needed for delivery--they could do the same re 24p.  I like the idea of everything staying at the original fps all the

way through post a lot.

Philip Perkins

[Brent Lestage]

Thanks, guys...

I'll definitely be sending the camera a reference audio track as well -- at least, for the interviews. Unfortunately, I won't have the luxury of a pre-prod. sync test (they are coming in from out-of-town).

Interesting that the Nikons shoot at true 24.00 fps. What I don't know, however, is whether their last mixer recorded at 24.00 or 23.976. If there was a mismatch, that would certainly explain the drift.

So, as with film, if the Nikon is shooting a true 24.00 fps, there should be no sync issues with me recording at 30ND?

--

Brent

[Eric Toline]

Thanks, guys...

I'll definitely be sending the camera a reference audio track as well -- at least, for the interviews. Unfortunately, I won't have the luxury of a pre-prod. sync test (they are coming in from out-of-town).

Interesting that the Nikons shoot at true 24.00 fps. What I don't know, however, is whether their last mixer recorded at 24.00 or 23.976. If there was a mismatch, that would certainly explain the drift.

So, as with film, if the Nikon is shooting a true 24.00 fps, there should be no sync issues with me recording at 30ND?

--

Brent

24/30 has always worked before no reason it shouldn't work now.

Eric

[Philip Perkins]

Thanks, guys...

I'll definitely be sending the camera a reference audio track as well -- at least, for the interviews. Unfortunately, I won't have the luxury of a pre-prod. sync test (they are coming in from out-of-town).

Interesting that the Nikons shoot at true 24.00 fps. What I don't know, however, is whether their last mixer recorded at 24.00 or 23.976. If there was a mismatch, that would certainly explain the drift.

So, as with film, if the Nikon is shooting a true 24.00 fps, there should be no sync issues with me recording at 30ND?

--

Brent

Whoa-- I didn't say the Nikon definitely is rolling at true 24--that's just a plausible explanation for what happened to this client's last shoot re sync.  What I said is that I haven't worked with or tested one yet.

You need to find out for sure what the framerate of the camera is from someone (Nikon, the owner etc),

and a test would be a very good idea, otherwise you can't be certain your MO will work.

Philip Perkins

[studiomprd]

" Production states that their last mixer used 24 (23.976?) and time code began drifting 2 frames four-minutes into a scene.. "

summarizing/repeating:

24.000 and 30.000 Frame rates are "integer" frame rates, and divide the same  whole, exact seconds into fractions of either 1/24 or 1/30. the :00 frames should match up perfectly, and remain in sync...

the 23.976 and 29.97 frame rates are the non-integer rates, that have been slowed down by .1% so that the seconds to get the last frame finished completely, are thus a little longer. there are still complete frames!!  these non-integer frame rates will also line up on the :00 frame, and stay in sync.

so, you need to do a test to determine if the 'camcorder de jour' is using an integer or non-integer frame rate. (this not considering the accuracy of the crystals or stability of the frame rates!!)

[Brent Lestage]

Got it. Production is going to run some tests. The drift may have occurred for a variety of reasons -- exact frame rate discrepancy, conversion in ProRes, etc.

Will wait for them o report back and will post the findings here.

Thanks to all!

--

B

[milos_m]

Please excuse the slight tangent, but

would recording LTC on one of the audio tracks of the DSLR work?

[Jeff Deeth]

DSLR Audio, 30.00 or 24.00 Timecode Slate.

I have always used 30 w/out issue, but 24 should work for you as well.

If in doubt test if possible.

So far I am enjoying DSLR from my vantage (Sound). Like shooting film w/out the Mag Noise.

Jeff Deeth

[studiomprd]

" would recording LTC on one of the audio tracks of the DSLR work? "

maybe, if the rest of the workflow supported it, but the TC on the audio track would probably affect the audio recording of audio on the other track...(remember 'stereo linked' and fixed levels, w/AGC ...)

[Philip Perkins]

Recording TC on one channel of a DSLR might actually work, but you would probably have way to much crosstalk within the TRS mini connectors--the TC would be all over your audio, esp vis a vis the camera's VERY aggressive AGC.  Timecode is nasty stuff to have with in any multichannel cable.   Also, as was said, you'd have to set up the post workflow to handle it.  In truth, TC to the camera isn't all that important anyhow--what the syncer-uppers really need is a good slate mark and some notes about what audio file goes w/ what pictures.

Philip Perkins

[Brent Lestage]

As promised, here are the findings of our Nikon D3s sound sync test:

Camera: Nikon D3s (720p)

Frame Rate Fixed: 24p

ISO: 200

Shutter: 125

Lens: 50mm

Aperture: 1.4

Sound recorder: Sound Device 744t

Test Takes: each are 5 minute takes

Test take 1: 744t Frame Rate Setting: 24 (‘sub-frame’ drift)

Test take 2: 744t Frame Rate Setting: 23.98 ( about one frame drift per minute)

Test take 3: 744t Frame Rate Setting: 29.97 (about 5 frame drift per minute)

Test take 4: 744t Frame Rate Setting: 30 Non Drop (‘sub-frame’ drift)

Our test shows that the sync ‘drifted’ at one frame a minute at the 23.98 setting and 5 frames a minute at the 29.97 settings.

At the 24fps and 30fps Non-Drop settings there Is a ‘sub-frame’ sync drift (1/2 frame).

We set up our Final Cut Pro timeline sequence as follows:

Sequence Settings:                              

Frame Size: 1280x720

Aspect Ratio: HDTV 720p (16:9)

Pixel Aspect Ratio: Square

Field Dominance: None

Editing Timebase: 24              

Audio Settings

Rate: 48 kHz

Depth: 16-bit

Config: Default

QuickTime Video Settings

Compressor: Apple ProRes 422 (HQ)

Quality: 100%

[studiomprd]

" At the 24fps and 30fps Non-Drop settings there Is a ‘sub-frame’ sync drift (1/2 frame). "

as expected for integer frame rates, and non-SMPTE crystal accuracy of the camera...

[Marc Wielage]

Yes, all of that makes sense: 24.00 for camera and sound (or 30.00ND for sound). There's been lots of conversation about workflow for DSLR's over on the LA Final Cut Pro User Group .

I have to kind of chuckle at people using Red cameras and DSLR's. A couple of years ago, the Red guys jumped around and said, "wow, this is the greatest -- we can shoot on digital for $17,000 and it looks as good as film." Now, the Red people are angry that ultra-ultra-low-budget people are shooting on Canons and Nikons for $5000. The quality ain't quite as good, but it's acceptable, assuming good lighting.

--Marc W.

P.S. I know, I know... archers vs. arrows!

[studiomprd]

" assuming good lighting. "

That's a pretty big assume on low budget stuff...

and there was just a comment about lighting on AI in another thread...