recap DV through SDI
by Julia Barreto
on
Aug 4, 2008 at 4:43:58 am
I'm managing the postproduction of a low-buget documentary all shooted in dv 29,97fps.
I have a doubt that my online searching cold not solve. I hope someone here can help.
I'm considering to recapture the film trough SDI, because the film will have a lot of color effects and I think If I can capture through SDI I'll have less lost when I render the final arquive.
Re: recap DV through SDI by Dan Dennis on Aug 4, 2008 at 12:14:57 pm
You can recapture DV through SDI, as long as your deck supports it. Capture using the Uncompressed 10-bit setting. Then create an Uncompressed 10-bit sequence. Uncompressed 10-bit is essentially lossless. Your color effects should also be lossless. Then you can output whatever format you want.
"Like the fella once said, ain't that a kick in the head?" - Dean Martin
Re: recap DV through SDI by David Roth Weiss on Aug 4, 2008 at 3:41:20 pm
[Julia Barreto]"I know that I can recapture it through SDI, but is it worth?"
Ah ha!!! That's the question I was thnking when you posted initially.
The answer is, not really. You'd do best editing DV captured via firewire, and once you're finished with the edit, simply change the QT video compressor in the Sequence Settings to ProRes and re-render. That will preserve the best quality for all your text and graphics -- capturing to another codec will not make your DV any better.
David Roth Weiss
Director/Editor
David Weiss Productions, Inc.
Los Angeles
POST-PRODUCTION WITHOUT THE USUAL INSANITY ™
A forum host of Creative COW's Apple Final Cut Pro, Business & Marketing, and Indie Film & Documentary forums.
Re: recap DV through SDI by Bill Dewald on Aug 4, 2008 at 5:47:40 pm
[David Roth Weiss]"The answer is, not really. You'd do best editing DV captured via firewire, and once you're finished with the edit, simply change the QT video compressor in the Sequence Settings to ProRes and re-render."
I'm glad this came up again - I've found conflicting information on the COW regarding the DV firewire vs. SDI issue. Here's what Gary Adcock had to say about it last time - he insists that if you're going to a non-camera codec, that you should let the hardware in your deck do the work, rather than the FCP software:
I've tested reingesting DV with my AJA-IO and my SONY DSR-40, and there was a huge difference in the image between the firewire signal and the compotent signal into the AJA.
I'm not trying to discredit David Roth Weiss' advice - I was trained to treat DV footage in the way he suggested, and I've gotten great results doing it that way. Just wanted to put a different point of view out here...
Gary can you wade in? by David Roth Weiss on Aug 4, 2008 at 7:47:17 pm
[Bill Dewald]"I've tested reingesting DV with my AJA-IO and my SONY DSR-40, and there was a huge difference in the image between the firewire signal and the compotent signal into the AJA."
Bill,
I love to be discredited any time I can learn something new or find a better way of doing something. But, in this case I'm not convinced that the difference is really "huge," or at least huge enough to discount the convenience of capturing via firewire.
I wanna know, what qualifies as "a huge difference" your opinion?
Gary, if you're paying attention, is there really a "huge" difference?
David
David Roth Weiss
Director/Editor
David Weiss Productions, Inc.
Los Angeles
POST-PRODUCTION WITHOUT THE USUAL INSANITY ™
A forum host of Creative COW's Apple Final Cut Pro, Business & Marketing, and Indie Film & Documentary forums.
The first was ingested via firewire from a DSR-40. The second was captured thru a AJA-IO via component ins.
Note the pixelation in the DV image (esp. around the man's nose and forehead), and the difference in saturation.
I've been reading up on this issue because of the problems that I had with the pixelation that you can see. Its not so apparent on a CRT monitor, but we took the image thru a Teranex to HDCAM-SR, and the upres made the problem much worse.
I tried to use the component in footage to fix the problem, but the difference in colors gave me pause.
I ended up using the 4:1:1 chroma smoother filter on the 'DV-firewire' footage, and that got rid of the jaggies.
Hope this helps... - Bill
ps - what's the best way to embed images in posts? I tried for about 5 mins before I gave up....
Re: Gary can you wade in? by David Roth Weiss on Aug 4, 2008 at 8:50:23 pm
Bill,
All I can say is, if I were having noticeable pixelation issues such as the ones you posted I'd be looking for a solution too. Completely understandable...
Thankfully, I haven't ever seen anything like that in the DV footage I've cut. I guess the bottom line is, if it ain't broke, don't fix it, if it is broke, looks like you found a good fix.
David
David Roth Weiss
Director/Editor
David Weiss Productions, Inc.
Los Angeles
POST-PRODUCTION WITHOUT THE USUAL INSANITY ™
A forum host of Creative COW's Apple Final Cut Pro, Business & Marketing, and Indie Film & Documentary forums.
Re: Gary can you wade in? by gary adcock on Aug 4, 2008 at 10:58:01 pm
[David Roth Weiss]"Gary, if you're paying attention, is there really a "huge" difference? "
Always lurking....
OK to clarify capturing content via baseband (component or HD/SDI) is allowing each manufacturers decompressions to be done at the very best quality- proprietary hardware will always do this better than software, which can choose to playback the image at lesser quality. RT Extreme will soften and desaturate an image to allow it to playback in realtime, hardware either does it or fails horribly.
Secondly - capture from the 8bit camera codecs into say ProRes or UC offers the users access to the full 10bit color space - and since it was acquired from the "optimally decompressed image" you are going to have better picture quality.
Unlike a Firewire Transfer - which is very highly compressed in the Mfg's customized version of compression. Working in these compressed codecs can and do limit what can be done to the image.
Camera codecs are limited in colorspace and converting from one 8bit codec to another 8bit codec can create artifacts, compression errors so the results are really, really nasty < do a search for DV to Mpeg conversion>
I am a hardware geek, I found out a long time ago that the right hardware can solve any number of the minor issues that we run into in day to day production, yet there are price vs need, cost vs. payback.
I always look at the bottom line, if it is faster in hardware and me save time and frustration- I am there.
gary adcock
Studio37
HD & Film Consultation
Post and Production Workflows
Inside look at the IoHD
Re: recap DV through SDI by Dylan Reeve on Aug 4, 2008 at 11:26:28 pm
My approach to this is slightly different as I've come at it from Avid generally. I've not used DV in FCP since 2002 or something (when it was actually pretty iffy I reckon).
If capturing from DV and laying back to DV it's been my experience that staying DV/Firewire for ingest and output is the best. For straight cuts only, up to 8 generations that I've tested there was absolutely no degradation.
The problem with capturing baseband (especially analogue component) is that it risks making minor spatial shifts in the footage, which if going back to DV means that DCT blocks will no-longer align which causes massive degradation.
I have also compared video captured into Avid at 1:1 (uncompressed) both via SDI from DSR-1800 and Firewire. There was no visual difference and on a very very minor technical difference.
I am not as familiar with FCP's inner workings, so would defer to Gary on that one, but in essence the video from the SDI source isn't any better, it still has been through DV compression and suffers all the resulting issues, however depending on the software's methodology for DV to Baseband conversion that aspect of the capture or conversion from DV to Uncompressed may be less than ideal.
I would expect, as an online editor, that if I had DV in a time and was rending to a higher quality format that my result renders would be as good as they could be. Any potential gains from a conversion to a higher quality earlier in the chain, I would assume, are going to offer very minimal, if any, improvement in overall results.
Perhaps in a case like this the best thing to do is test? Process only a small selection of shots and see if there is an appreciable difference that would make the extra effort worthwhile.
