Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Data workflow for high-end online (broadcast/TV)?

G

GlennChan

Banned
Joined
Dec 30, 2006
Messages
933
Reaction score
0
Points
0
Website
www.glennchan.info
I'm curious:
How will the data workflow work for high-end online systems? i.e. Discreet FFI (SGI and Linux), Quantel iQ, Nitris DS, etc.

Presumably...
A- Finish offline edit. Output EDL, cutting copy.

B- Bring EDL into Redcine. Redcine copies section of Redcode media onto an external drive (i.e. RAIDed FW800 so the transfer goes faster).

C- Drive is couriered to online facility. This is unless the offline editor is in the same building as the online, which does happen.

D- An online assistant takes the footage and dumps it into a computer with Redcine.

E- Redcine converts media into a file format that the online system will accept (still TIFF sequences?). Is it safe to assume that performance of these (possibly uncompressed 4:4:4) files will be the same as 4:2:2 Y'CbCr? And what about clipping of out-of-range values?
*It may be that the online system accepts some 4:2:2 Y'CbCr data format. I don't know.

F- These files are written directly onto a SAN or the online system's storage. Is it safe to assume that the network protocols and/or file systems are compatible, and that network speed is fast?

G- Things like field order and pixel aspect ratio are guessed correctly. If not, it is not a big PITA to manually change these settings.

H- If you had shot 24p (NDF) and you need to edit in 60i (NDF), and you go through this process, with that complicate things? Can the Red pull list work backwards from a 60i EDL and reference that to the 24p Red clips?

I- Can the online system take in an EDL, and conform that from data files? The online system needs to be able to read timecode from the data files.

J- If you need to change the reel names to no more than 5, 6, or 8 characters, can Redcine handle that?

K- Can Redcine warn you if you put a "B" on the end of the reel name? This would be to avoid conflicts with B-reel naming convention; I'm not sure if this is necessary.

My inquiring mind would like to know! (in the context of HD and SD production for broadcast)

Some previous discussion:
http://www.dvxuser.com/V6/showthread.php?t=76081&page=2
 
Have it your way

Have it your way

Based on what we have been told about RedCine: if you know what flavor you want to have go into your online pipeline then you should be able to set RedCine to output that particular format and everything after that should be your usual program. Issues of file naming, EDL management and the rest of the details may require a bit of testing but IMHO it shouldn't be a big PITA.
 
Glenn, when these high end suites take still image files, it's usually dps, which is 10bit, 4:4:4 RGB. I don't know any that would deal with Y'CbCr at this point.

G - what field order? We're talking progressive here, surely. RED PAR =1.0;

H = datacentric, so why would you be dealing with pulldown and interlace and all that nonsense?
 
  • Thread starter
  • Banned
  • #4
1- You mean DPX?

G- Field order: You might end up in a situation with interlaced output, if you need to add 3:2 pulldown.

As far as PAR, what if you need to output 720x486, or 720x576 ?

In any case, field order and PAR shouldn't be that big a deal if they are misinterpreted since you can grind through and tell your program to interpret everything correctly. It's just an annoyance when importing material... i.e. IME when you import footage into AE or Combustion, it may not always be interpreted correctly.

H- You may need your offline edit to be 60i if you are editing a documentary with a lot of 60i footage. 60i also has an advantage when it comes to overall workflow + packaging, since packaging almost always needs to be done in 60i (broadcasters require a 60i master). Yes interlacing is evil but... sometimes you have to deal with it.
 
But all of the above are NLE issues. REDCINE makes the files you ask for. After that, the NLE has it's job to play.

Graeme
 
  • Thread starter
  • Banned
  • #6
I think it's a Redcine issue.

If Redcine is to add pulldown (Stuart seemed to agree that it would be useful; the original plan was no), then when you send an EDL back to Redcine it should know to pull the right timecodes. It should convert from 59.97NDF timecodes to 23.98NDF, then pull, and then add pulldown to the output with (the same) 59.97NDF timecode.
 
Okay, I'm out of my area of expertise here, but I'm gonna try to say something useful:

1) The 59.97 framerates are only important for TV. So that's SD or HD. My suggestion: shoot at 720 or 1080 in either interlaced or progressive at 59.97 or 29.97. This should mix well with other clips you might pull in.

2) If you're working in 23.98 fps you're probably shooting a movie all on this camera, so you don't need to mix & match, so no problem there either.

So you only need pulldown for #2. This can be done when you export your 24 fps (or 23.98) timeline to whatever format needs pulldown. No need to do this during editing.

Am I missing something, or?
 
  • Thread starter
  • Banned
  • #8
Rob, that would be a very sensible thing to do. I agree with you on #1.

However... some people will want to shoot 24p for the ""film look"". And because of the marketing by companies selling 24p cameras, a lot of people think about shooting in 24p. And in practice, some productions are shot 24p and 60i on the same camera. Because, supposedly, 24p=film look and 60i=documentary/news look. It's silly, but that's what happens. HDCAM decks can deal with adding pulldown (and TC issues), so this route is doable (although it's still a PITA, and complicated and confusing and slows you down when you set the system up wrong).

2- People might assume that it's ok to shoot 24p (i.e. with huge workflow hassles).

You even see that happen now, where people shoot 24pA and don't realize that the online system they will be using can't remove the advanced pulldown.
 
I am going to use the RED like it is a film camera running at 24p.
Feature films are shot and edited 24p. You don't see Spielberg
saying "This is going to TV lets shoot 29.97". There are established
workflows to go from film back to TV.
 
"2) If you're working in 23.98 fps "

is it really 23.98 or are you just rounding up from 23.976 ...
 
23.98fps really refers to 24 * 1000 / 1001 fps (which is close to 23.976fps). In common use, it doesn't refer to anything else.

The 1000 / 1001 part is important since it affects timecode and you get into potential timecode issues.

Feature films are shot and edited 24p. You don't see Spielberg
saying "This is going to TV lets shoot 29.97". There are established
workflows to go from film back to TV.
Click to expand...
Hi Red Guy, I don't believe I understand what you're trying to say. I am talking about projects targeted primarily towards broadcast. Some of them mix 24p and 60i for the different motion aesthetics, in 60i timelines. So those projects need to deal with these issues. Obviously you can avoid these issues by shooting 24p, but some people want to mix 24p and 60i.
 
Sure you can add pulldown to 24p material to make it 29.97. I just hate editing 60i and have not shot 60i since I retired my BetaCam almost three years ago. I guess if you were doing a sports show and wanted to shoot the live part in 60i and the feature segments as 24p that makes sense. But I doubt you will see much RED shooting in 60i.
 
Glenn is right about 1000 / 1001. 23.98 was just convenient.

Obviously for my option 1 you could also shoot in 2K or 4K at those frame rates and down-sample in post. If you want to maintain a higher resolution negative.

The other question should already have an answer outside RED. A lot of people shoot 24 fps film. No doubt a lot of them edit in 24 fps. But what if they do need to add in 60i originated footage or add the film footage into a 60i timeline, what happens? These questions should already have answers before RED arrived on the scene.

And I think that's what Graeme was sort of implying (correct me if I'm wrong here Graeme!). Shoot the way you want to shoot it and use the established methods to get it into your proper editing workflow. Whether that is adding pulldown or shooting at a proper framerate.
 
Rob, you read me right, but also there are too many workflows (like tape to tape colour correction) that are archaic and not data centric. I think that's the issue that Glenn is talking about. I guess such people could just put their RED footage on a timeline, output to HDCAM and then "treat as normal".

Graeme
 
GlennChan said:
23.98fps really refers to 24 * 1000 / 1001 fps (which is close to 23.976fps). In common use, it doesn't refer to anything else.

The 1000 / 1001 part is important since it affects timecode and you get into potential timecode issues.
Click to expand...

Glenn, sorry to contradict - but this is incorrect. There is no timecode difference between 23.976 and 24fps. Take this example: two cameras - one shooting at 23.976 and one at 24. They both shoot for (by a stopwatch) an hour. The end timecode will be different on both tapes, because they count time-by-the-clock differently. But the timecode on the two tapes will count identically. The frames go from :00 - :23 on both tapes.

The #1 reason for deciding whether to shoot 23.976 or 24 in a digital world is keeping your deliverable in mind. What will the final result be -theatrical or broadcast? If the answer is US (or another NTSC-based country) broadcast release, then it's a no-brainer. Shoot 23.976. If it is going for theatrical release, then the number one consideration is audio.

In digital acquisition, there are 24 frames of imagery in the camera regardless of 23.976 or 24. They are just played back at a different speed. In audio, that's not the case. Assuming a 48K sample rate, in digitally acquired audio, there are 48,000 samples in one second of 24fps synced audio. There are 47,952 samples (48 * (1000/1001)) in one second of 23.976 audio. On a dual-system show, if you now attempt to playback that audio at 24fps, it will not be in sync with a true 24fps-acquired material. Happens all the time. And then the fun starts...

There are great ways of dealing with 23.976 vs. 24fps audio in the Avid world. Because of the history of telecine - which *always* slowed down film to 23.976fps in order to properly telecine to NTSC - Avid came up with what at the time was a groundbreaking way of dealing with this and editing at true 24fps for theatrical. Thus was born the Film Composer... and a Sci/Tech Academy Award to go along with it...

But I digress. :)

GlennChan said:
Hi Red Guy, I don't believe I understand what you're trying to say. I am talking about projects targeted primarily towards broadcast. Some of them mix 24p and 60i for the different motion aesthetics, in 60i timelines. So those projects need to deal with these issues. Obviously you can avoid these issues by shooting 24p, but some people want to mix 24p and 60i.
Click to expand...

Yes... and then the question becomes - what is the dominant format? If *most* of the footage is 24p, and there is just some 60i material to deal with, then the standard in most online bays is to Teranex the 60i material to 24fps, and edit it in with the 24fps material, then re-introduce a consistent cadence on output.

Follow the workflow here. Let's assume (for an hour-long doc) you have 50 minutes of 24p material, and 10 minutes of 60i material. In an editorial timeline - in the actual application - how will you mix these? This is not a rhetorical question... :)

Lucas Wilson
------------
ASSIMILATE, Inc.
jet-lagged in Asia
 
Luki: My comment about 1000 / 1001 was in the context of dropframe versus non-dropframe; not 23.976 versus 24.000. (I believe DF is never used since it gets into TC weirdness.)

Yes... and then the question becomes - what is the dominant format? If *most* of the footage is 24p, and there is just some 60i material to deal with, then the standard in most online bays is to Teranex the 60i material to 24fps, and edit it in with the 24fps material, then re-introduce a consistent cadence on output.
Click to expand...
Yes, that would work. But a few oddball productions want that 60i look. The Canadian show "I do let's eat" uses 60i and 24p footage (it's all shot on HDCAM); incidentally, the 24p motion segment disturb me but that I suppose was part of the creative intent (interviews = 24p/"film", documentary segments = 60i).

I guess such people could just put their RED footage on a timeline, output to HDCAM and then "treat as normal".
Click to expand...
That does work, but defeats the point of a data-centric workflow (no expensive VTRs; ideally, more efficient).
 
But if the editing software doesn't support a data centric workflow (and if it did, we wouldn't behaving this disucssion) you wouldn't need to dub to HDCAm and treat as normal! It's about time that NLEs stopped treating "time" as a fixed parameter and gave it the flexibility it deserves.

Graeme
 
I guess it would make life (for you North Americans at least) a lot easier if eg Final Cut would let you drop a 24p clip on a 60i time-line, and add pull-down on playback. No real reason this shouldn't be possible. Even if the computer couldn't do it at broadcast quality in real time, it could give you a preview while working, and render it before mastering.

Problem solved!

Nick
 
Indeed, but FCP was designed with a very rigidly defined concept of time - and conversion of one time base to another has always been screwy. It would be lovely to see them move to more of a timebase agnostic solution, but I don't see it happening myself.

Graeme
 
Just realised after I wrote that, your 'G Add Pulldown' plugin does pretty much exactly what I was describing! However as you say, Final Cut's rigid time-base structure makes it a less than ideal workflow.

Nick
 
You must log in or register to reply here.
Back
Top