Home of Professional Cinematography since 1996

style="margin-bottom: 0"> 

class="style8" HDCAM-SR For DI

>Published – 7th January 2005

>I'm wondering where the cheaper DI places are now in Los Angeles. I've got a project coming up that is planned for 35mm anamorphic but we have some shots that were done in 24P HDCAM last year by the director that have to be incorporated into the last reel. That and the fact that the editor is keen to do a DI as well as the director means that I'm exploring how to make it affordable.

>3-Perf is one option but it's not enough of a cost savings, so I'm wondering what the current state of DI's are right now in Los Angeles. I read that Post Logic did a DI for "Undertow" and I'm wondering what it cost them.

David Mullen, ASC
Los Angeles


> David Mullen wrote:

class="style9">>I read that Post Logic did a DI for "Undertow" and I'm wondering what it >cost them.

>First let me say that I really admire Tim Orr's work. But when I saw a projected extract from 'Undertow' I was not convinced by the DI at all.

>There was one scene in particular with the father sitting in the kitchen lit apparently by a bare light bulb giving out a warm light. That scene looked very videoish. Especially the skin tones didn't look pleasing in this light at all, which is a dead giveaway for a bad DI.

>I have seen other DI’s deal much better with warm light on skin, but unfortunately they were all done in the UK.

>Max Jacoby
Lux/UK
www.imdb.com


>Probably the two best DI's I've seen lately have been "After the Sunset" (Company 3/E-Film) and "The Life Aquatic" (E-Film).

>David Mullen, ASC
Los Angeles



>I really admire Tim Orr's work. But when I saw a projected extract from >'Undertow' I was not convinced by the DI at all.

>I may be wrong about this, but the last time I was at Post Logic they were working in a Linear TV Gamma space (video), not LOG. This is likely to generate the issues you mention, as the dynamic range is very restricted, with the probability of crushed shadow detail and clipped and washed out highlights.

>I was a far while back that I was there, so they may have moved into a LOG workflow now.

>Companies such as E-Film, CO3, MPC, One Post, MotionFX, Lipsync, etc. all work LOG, and the results tend to be much more filmic in feel and response.

>It is obviously possible to use HDCAM-SR (or any other tape deck) to save LOG images, the 'video' format is simply a wrapper for image data, but with the specific attribute of the vtr being used (compression, YUV, pixel reduction, etc.) AHC in London do a lot of commercials this way, with the film being transferred to HDCAM or D5 in LOG form from a self calibrating
telecine/scanner (DSX as it happens as St.Annes Post).

>The commercial is then posted in a DI workflow, with the colour correction being done after the on-line editorial. The results are very good, better than with a Linear TV Gamma workflow, even though the final is for TV use, not the cinema.

>As always, there is a fair bit of detail on these issues on the Digital Praxis website.

>Regards,

>Steve Shaw
Digital Praxis Ltd
www.digitalpraxis.net


>>Companies such as E-Film, CO3, MPC, One Post, MotionFX, Lipsync, >etc. all work LOG, and the results tend to be much more filmic in feel >and response.

We also finish in log, though in all cases, one is looking through a print predictor LUT, so it looks like normal print density on the projector.

>There's nothing really wrong with working in float, but I believe it tends to foster black-point color and density errors and odd shifts in over-values, but that's more a workflow/QA thing than anything else. We think it's a bad idea to work in 601 or 709 for film output.

>There's a (lazily written) description of our work on Lions of the Kalahari in last month's American Cinematographer.

>Tim Sassoon
SFD Vfx & creative post
Santa Monica, CA


class="style9">> We think it's a bad idea to work in 601 or 709 for film output.

>On the other hand it would be a good thing to work in____________ for film output?

>Curious minds would like to know.

>Mark Smith
Oh Seven Films
143 Grand St
Jersey City, NJ 07302



Mark Smith wrote :

class="style9">>On the other hand it would be a good thing to work in____________ for >film output?

>10 bit log. And eventually maybe 16 bit Open EXR.

>Dave Stump ASC
VFX Supervisor/DP
on location in Vancouver



> 10 bit log. Â And eventually maybe 16 bit Open EXR.

>Store 10-bit. Calculate in 16-bit or float. Calculation bit depth always needs to be higher than storage depth, since one is trying to avoid rounding errors. There's nothing wrong with 10bpc log storage of _film_ original material.

>CGI is a different beast, and one wants to store Z-depth, rainbow lighting, ambient occlusion, layers, and lots of other things into a composite.

>Tim Sassoon
SFD Vfx & creative post
Santa Monica, CA



>10 bit log. And eventually maybe 16 bit Open EXR.

>My main problem with OpenEXR is that it is a compressed format. It's great for metadata transfer, archiving, color management, etc., but not really for playback.

>Without an awful lot of expensive heavy iron, current systems can't really handle the overhead of a decompression pass when trying to stream 2K in realtime.

>Lucas Wilson
Assimilate, Inc.
Los Angeles


class="style9">>My main problem with OpenEXR is that it is a compressed format. It's >great for metadata transfer, archiving, color management, etc., but not >really for playback.

>Compression is optional with OpenEXR. Your software's plug-ins may only support compressed output, but you could process the files and create uncompressed versions for playback.

class="style9">>Without an awful lot of expensive heavy iron, current systems can't really >handle the overhead of a decompression pass when trying to stream >2K in realtime.

>Which is why everyone uses uncompressed OpenEXR for playback.

>Brian Willoughby
Sound Consulting


>Brian Willoughby wrote :

class="style9">>Without an awful lot of expensive heavy iron, current systems can't really >handle the overhead of a decompression pass when trying to stream >2K in realtime.

>The other option would be to use a 16 bit DPX file with an increased density range, as can be output by the Northlight scanners.

>We haven't come across many uncompressed OpenEXR 2K playback systems in the wild (so to speak), is there one you can recommend?

>| Kevin Wheatley, Cinesite (Europe)Ltd | Nobody thinks this |
| Senior Technology | My employer for certain |
| And Network Systems Architect | Not even myself |