S16 To Video Workflow
14th November 2004
I've my first film project coming up shortly. More specifically, we're shooting a demo commercial on S16 with the intent to distribute on a video format such as DigiBeta, or DV.
Not being very experienced in the postproduction workflows involved, I had envisioned to telecine (one light) the negative to some DV format for editing with FCP. After completion of the editing, I was hoping to continue on with the DV material, i.e. I was hoping to perform the grading of the footage with after effects right on my desktop. the final result (DV) could then be converted to whatever format was required by this or that TV station (beta, DVD, etc.).
When I described this idea to the guy at the local post processing
facility, I was laughed at. He told me that that was definitely not the way to go, as I would be severely limited in what I could achieve in after effects. Furthermore, I would be taking a serious quality hit when compared to working in DigiBeta. In short, the guy told me, I'd need to have the negative conformed to my FCP edit, so that a subsequent telecine to a high end video format could be performed (DigiBeta). this second telecine would then be graded in a session with the facility's colorist (using their expensive grading bay).
My question to people more knowledgeable than I in these matters is whether what this guy tells me is really true. I know that DV compresses at a higher ratio than DigiBeta, and Iâ€™m also aware of the fact that there's 4:1:1 versus 4:2:2 sampling involved, but, do these format differences really make such a great difference considering the acquisition and output formats? How on earth am I supposed to be limited in what I can do in after effects?
Any and all comments appreciated,
It all depends on how careful you are shooting and telecine'ing.
If you can get something clean and VERY close to what you want in DV, then you should be OK. You face two issues. One is that the limited bandwidth will cause banding in gradients if you move things around too much. The other issue is re-compression. Any grading will involve a re-render - that's twice through the codec - not good. You WILL see it. BUT - if you're delivering on DV, it's not a bad way to go because you will have to run through the codec at the end anyway. It's much better to go to DV at the end, but if you're really careful to avoid too much pushing the colors around and no more than one re-render you may get away with it - depends who's looking at the final product.
Don't underestimate the power of an experienced colorist though. Apple would like everyone to think that even a monkey can do color correction, but it's a very specialized skill.
BUT - sometimes you're just trying to get something done and you can't afford this and that. Don't let people tell you you're crazy - have them give you a test clip and try it for yourself.
Red Sands Production Co.
> I would be taking a serious quality hit
Your post house is right. If you are originating on film, then it seems foolish to throw away the quality by finishing on your compressed desktop editing format.
You could either conform negative after your FCP edit - or, more likely, extract the takes you have used, transfer those to Digibeta and conform and grade on-line - or, depending on the shooting ratio, have everything transferred in the first telecine session, to DigiBeta for the online AND to DV for the FCP edit.
Naturally you'd expect the post house guy to be protecting his company's investment by talking it up, and talking your desktop down, but if you want the quality that your S16 shoot is obviously going to offer, you need to use it. If no-one needed it, the post houses wouldn't be in business.
>The final result (DV) could then be converted to whatever format was >required
"Converting" is easy to do, but you only get out what you put in.
"You should be OK. You face two issues. One is that the limited bandwidth will cause banding in gradients The other issue is re-compression. Any grading will involve a re-render -that's twice through the codec - not good. You WILL see it."
"Don't let people tell you you're crazy - have them give you a test clip and try it for yourself."
Yea - everything Greg said....and....and be aware that PAL Dv has 4.2.0, not 4.1.1 that NTSC DV has, so much less room for CC in that colorspace
We did a series that was shot in PAL Dv & 16 that was Xferd to Dv
All cut in a Dv Avid, but cut only, no effects what so ever - no CC - no re-framing - no nutt'n - cuts only, all footage needed for effects was dropped at the end of the timeline
We played the DvAvid out through a Miranda Dv -> SDI converter, it is a very high quality piece of kit.
We recorded the entire show in SDI into a DS finishing system uncompressed.
We then went back into the Dv Avid and added all effects reframing, dissolves etc etc, and saved as an OMF
We did all effects and CC in the DS from information derived from the Dv Avid's OMF
We used "Anvil" software for the color correct in the DS.
All broadcasters were very happy with our shows, USA, UK, Germany included.. three hard markets to deal with for QC issues, it went to 60 markets around the world, and not a single rejection ever.
I would stress that you should stay away from any hint of re-compression - even a dissolve or a fade to black, these will be noticed on a grade A monitor, and cause a rejection from many broadcasters - all in the UK will reject for this, Discovery in USA & Canada will as well, as will all major's in the USA, ZDF & Channel+ both are very demanding as well and will reject these shows - and so they should - it is appallingly obvious at home.
I do think that software CC has come of age, esp tools like Luster, DS with Anvil, Discreet with Colorfront, Cyborg, and all recent Quantel tools.. but there is nothing even close in the FCP land to match these toolsets,
AE is competent software from a technical point, it is far from my fav tool, but there is nothing wrong with it's ability to complete work if you are careful - many plug-ins are limited to 8bit color space, negating the value of the ability to work in 16bit, as the weakest link in the chain is the limiter of quality - be very careful, you cannot use the tools without finding out if you are hooping the bit depth, and may find that many popular plug-in's are so limited and then you must never import the final comp back into FCP - at least not as a final, Xfer the uncompressed render to your finishing system
We finished shows that are examples of how to do Dv well, but they were budgeted well, and shot well, and were very successful as a result
You can Subtuite a on-line FCP set-up for the DS in this scenario, but still cannot re-compress at any stage - after you have finished your test shots, send them to your broadcaster's QC department fro a quick look, and do this before committing to a post work flow, or budget for the more expensive option, then you will be able to finish the show fro sure.
Good luck - your mileage may vary
Vancouver, Canada & Shanghai, China
I realise that you may need to do this for budgetary reasons but, contrary to previous post, I think it's an awful way to go.
If your intention is to grade tape to tape then you need as much colour information as you can get to start with.
I think this approach is severely limited with DigiBeta, my opinion as to how limited it is with DV is unprintable.
If at all possible I would use the FCP for off-line and go back and grade from the neg for an on-line.
I do think you can get away with using DV as the final TK record format if you conform this way.
Geoff Boyle FBKS
Director of Photography
Generally what the folks at the post facility was correct, but it seems they could have explained it better.
You went to the expense of shooting film. To telecine it as a one light would do you and your project a great disservice. You may as well have shot on video if you plan to work from a telecine one light. In one light transfers the colorist doesn't have the time to tweak the image very much and there may be blacks crushed and whites clipped. Shadow and highlight detail gone forever. One light implies no color correction, so the balance will not be all that good. Even best light transfers are prone to clips and crushes. So you would in effect be limited in after-effects. But shame on them for laughing at you.
Grading on the telecine offers the most color correction bang for the buck because when you are coming right off the negative, with its huge contrast ratio, you have total control over the image. You can't apply a power window to a clipped video sky to bring out the clouds, but from a film negative you can. As this process can be expensive over the cost of a feature, often simple telecine dailies are used for editing, then the selects can be retransferred fully color corrected right off the film negative. Nothing else can look as good, but there are other ways to do this, less expensive and with only a modest quality loss.
One possibility would be to have a "flat grade" film transfer. Unlike a one-light, which is a rough approximate grading, a flat grade transfer requires scene to scene attention, but the colourist's job will be to preserve the highlights and shadow detail as well as reasonably balance the color. Then, if you wish to, you can grade it yourself with access to most of the negative's information. Of course, an experienced colorist would IMHO be a better choice for grading as we are generally more aware of what can be done to make scenes look better quickly, so bringing your edited flat grade to a (nicer) facility for correction would be good if it can be afforded.
As to formats, DBeta is far superior to DV, and considering the acquisition format is film, why would you want less than the best? Transfer to DBeta, dub to DV for editing, and regrade either from the film or a DBeta flat grade.
You shot film. Don't be in such a hurry to turn it into lower quality
video. Keep it real!
Shooters Post and Transfer
Dominic Case wrote :
> "Converting" is easy to do, but you only get out what you put in.
I keep telling people that if you put a quart in a gallon jar, its still a quart.
Why doesn't anyone believe me?
Is it because digital quarts are a different size from analog quarts?
Oh Seven Films
143 Grand St
Jersey City, NJ 07302
Francesco Krattiger wrote :
>my question to people more knowledgeable than I in these matters is >whether what this guy tells me is really true.
For your video destination, the quality difference is present but is not fatal. Consider doing your editorial on the desktop in DV but do not attempt color correction on a desktop. Bring the assembled spot back to any *competent* post house for tape to tape color correction. Record the color corrected result to DigiBeta. Dub to other formats as required.
The extended precision and 4:2:2 bandwidth in the color corrector while manipulating the colors will be a great improvement over desktop color correction tools and the resulting re-compression back to DV.
I've seen some very impressive results with this workflow.
Thanks for the many informative responses.
As someone pointed out, budgetary reasons compelled me to consider the route I had described initially...from reading your posts I've come to realize that I'll be shooting myself in the foot in going the DV route...as anybody, I am a friend of quality, which motivates my originating on film...so if I'd choose a post workflow that resulted in a less-than-perfect final result, I'd probably suffer a 2 week depression...even though some people were kind enough to offer explanations as to how I might proceed along the DV path with acceptable quality, I think I'll go the safest route...i.e.
I'll do what Geoff suggested : off-line edit in FCP, followed by a grading off the telecine for an online... it's simply a case of cold feet...screw the budget.
Thanks again... you guys are great!
Robert Lovejoy wrote :
>Grading on the telecine offers the most color correction bang for the >buck because when you are coming right off the negative, with its huge >contrast ratio, you have total control over the image.
Bob speaks the truth !
Hi Francesco Krattiger (and the rest of CML-Post!)
An option to consider which hasn't been floated here is telecine to DVCPro50 and going in via firewire from the new decks. Its something I have considered but yet to try...
I don't think you can push Pro50 around as much as uncompressed 10bit Digibeta -- but on the upside, you have far more information than you do with DV (ie its a 50megabit format with 4:2:2 colour space) and its just as easy to edit (its data rate is only 7.7megs a sec... and you can work in an uncompressed 10bit YUV colour space in FCP4 and/or After Effects).
From what I've played with stuff originated on Pro50, its a good compromise between the convenience + cost of DV and the quality of Digibeta...
Otherwise you have to go the trad route of timecode locked DigiBeta for your offline etc. etc.
You can still do it on the desktop with some dollar outlay (worth considering if you expect to get the work)
You buy yourself a DeckLink card, 2x 200gig ATA drives and RAID stripe em (off ATA 133 controllers), and then hire a DigiBeta deck so you can go in SDI and work 10bit uncompressed. You can produce good results... vis a vis:
(A music video I produced -- which was shot 16mm, telecined to DigiBeta, offlined on a powerbook and onlined on the same powerbook -- colour correction was done in shake. Total hard drive space available was 60 gig. The footage was captured on another system but transferred to the powerbook for the online)
Stuart m. Willis
>From what I've played with stuff originated on Pro50, its a good >compromise between the convenience + cost of DV and the quality of >Digibeta...
How is the performance editing with DVCPRO50 coded files? What about multiple video streams? And real time effects?
We are using DVCPRO25 with Media100lx and considering moving to DVCPRO50 and just software based (FCP) editing.
Thanks for your advice!
Audio & Video
Ass. Torre de Vigia
Sao Paulo - Brazil
>I don't think you can push Pro50 around as much as uncompressed >10bit Digibeta
I know it's been pointed out here before, but in the interest of accuracy...
DigiBeta is not an uncompressed format. The only current uncompressed standard def format is the Panasonic D5. Unless you've got some D1 machines sitting around (8 bit, of course).
Of course, this brings up a different discussion - what is acceptable compression. I believe this discussion is very relevant, as many in the production community want to record from the "new generation" 4:4:4 cameras in RGB on disk recorders, and yet there is no real practical solution for this that has an established post path, and no simple way of "compressing" the physical space required to house all of that data from, say, a 60 day shoot. Many decry the use of compressed video formats, and yet the complaints about DigiBeta ceased long ago, to the point that many, like the previous poster, are convinced that DigiBeta is itself an uncompressed format - simply because the results are so superb.
IATSE Local 600
Mike Most Doth Quoth:
>like the previous poster, are convinced that DigiBeta is itself an >uncompressed format - simply because the results are so superb.
FWIW, I'm well aware that DigiBeta is a compressed format. Roughly half as much compressed than DVCPro50 -- thou, I believe the argument is that as DVCPro50 is a newer format, its compression is more comparable in quality.
I was talking about post-production workflows, where the norm is to capture DigiBeta in uncompressed 10bit. The original footage may be slightly compressed but it never gets compressed again till tape out and then only slightly.
Compare this to what I was suggesting: working with Pro50 captured in its *native* codec. While it doesn't have to be compressed again (and all the rendering will occur in 10bit YUV space) chances are you will finish back to the Pro50 codec for output via firewire. Recompression of a 3.3:1 compression format. I believe that Pro50 is still an 8bit codec, even tho the SDX900 shoots 12bit...but I could be wrong.
I've only just began playing with Pro50 in its native codec and it seems, at this early stage, that I can't push it around as easily as I could DigiBeta which has be captured uncompressed 10bit But its a good compromise between the quality of DigiBeta and the convenience of DV.
Being a post tweaker, that's important to me. I'm not really that impressed by the new DVHD format (or whatever JVC/Canon/Sony are going to call it) because I know once I bring it into an NLE I can't do shit to it without the compression breaking up.
Stuart M. Willis