In-camera versus post with HDV

[James Tauber]

I understand that leaving things until post keeps one's options open.

 

But given the amount of compression in HDV, I imagine that one would get a higher quality result doing as much image manipulation *before* compression as possible---in other words, with filters and in-camera processing.

 

When working with HDV, do people recommend trying to do more in-camera? Is there a significant qualitative difference in doing things like gamma and matrix manipulation in camera versus post due to the fact you're doing it *before* things like chroma sub-sampling and interframe and intraframe compression have taken place?

[David Cox]

Actually the thoughts about when to apply filtering apply to all cameras - the HDV compression adds only small extra issues.

 

When you place a filter on a camera, you have an infinite latitude. For example, if you put a nice graduated filter to darken down the sky as a filter on a camera, you would probably reveal some interesting cloud information. If you tried the same trick in post, your sky would have burnt out completely on the video image and a post graduated filter would just darken your white sky to grey, because the cloud detail is lost in the video burn out.

 

The counter argument is that if you apply a filter to your camera, you are pretty well stuck with the results. Remember that colour filters particularly are subtractive processes not additive ones. In other words, if you place an orange filter on your camera you are actually getting rid of blue light. This means that if you later decide you need to get the blue back, well you can't, because it was filtered out at the point of recording.

 

With the high compression of HDV, both these arguments are even more the case, since burnout (loss of information) happens that much sooner and the ability to reverse a heavy colour grade is more compromised.

 

So there isn't really a definitive answer. It depends on the shot and what you want to acheive. I would suggest as a starter, if the effect of your filter brings a greater amount of latitude into the capture range of your recording media, then its probably good to use a camera filter. If however the effect you are going for is going to remove a lot of information, such as a strong colour cast or even heavy diffusion, then post will give you a finer control that is reversable and viewable within the context of your cut.

 

David Cox

Baraka Post Production

[Phil Rhodes]

Hi,

 

I tend to take the position that with heavily compressed video formats, it's often best to take a top down approach. I used to keep getting asked how to make it look "Green, like the Matrix." The obvious advice was, of course, "shoot green objects." I feel that every stage should get you closer to what you want - production design, optical filtration, camera setup, postproduction. And yes, on heavily-compressed tape format, you might choose to try and get closer than usual to the final result in camera, at risk that it might not be recoverable, because that's just the tradeoff you have to deal with.

 

Phil

[Robert Sanders]

With the Canon XL-H1 I try to set a safe exposure; slightly under-exposed. I stretch the blacks and set the knee to protect as much of my shadow and highlight detail as possible. Even if I'm going for a high-contrast look. I prefer to crunch my blacks and set the gamma in post. However, I do try to achieve as much of the "look" as possible through art direction/production design, gelling lights, etc.

 

I use a very mild pro-mist most of the time and I always use a polarized filter when shooting exterior day shots. While I haven't used very many graduated filters, I believe David Cox is right that if you can pull detail out of your skys you should.

[Michael Collier]

Color correction in HDV is pretty finniky, and doesnt allow for much wiggle room. If you are shooting right you know what look you are going for anyway, so yes, I would say there is a huge quality advantage in doing the bulk of color correction in camera, when its still in the 12bit full colorspace DSP phase. Once it goes to compression, you loose any ability to change the data without compromising quality. (especially as low bitrate as HDV is. Block noise is exagerated when color correction is used)

[Guest peter orland]

I shot a short (7min) drama just before Christmas last year and we used two cameras, a Sony Z1 and a Sony VX-1000. The Z1 was for the shoot and the VX-1000 was for behind the scenes. The only scene that was colour manipulated in post was a scene shot in coffee lounge, it was lit with Kino?s and shot flat as we new we were going to adjust it in post. Since the VX-1000 was just for the behind the scenes we played around with the white balance and created what we thought was not a bad look in camera. When it came to the edit (Final Cut 5), and the colour manipulation (Shake) the degradation of the picture quality using this level of software (and a very experienced colourist) was so great that it ended up looking worse than the untouched DV footage off the VX-1000. The difference was so great that we even played around with trying to reframe the DV footage, but then the shots became the compromise. If I could go back I would of shot everything as close a possible in camera.

 

Should of tested. My own stupid fault.

 

Thanks.