In camera downconvert vs Vegas downconvet

Laurence wrote on 11/24/2005, 7:14 AM
Well I videotaped my daughters outdoor 13th birthday in HDV with my HVR-A1. I captured the footage both with ConnectHD and with the in-camera downconvert into two separate folders. I rendered to DVD from both subdirectories and compared the DVDs. Maybe my TV isn't big enough (37" HD Toshiba), but I just don't see any difference. I can see a huge difference between the SD and HDV footage on my computer monitor, but not in the SD DVD's I am planning to distribute for the next couple of years.

No difference in ending widescreen DVD quality, but a huge difference in file sizes and rendering times. For the next year or so while the HD delivery format is being sorted out, I believe I'm going to stick with DV resolution on my projects. If I could see a difference in the final DVD's I'd put in the extra effort, but I just can't.

On the flip side, I see a pretty big difference between native dv and in camera downrezzed HDV (with the HDV originated stuff looking markedly better). For the next couple of years at least, I think that HDV footage downrezzed on capture, edited as DV, and output to 16 SD DVD is the most practical way to go.

Comments

mbryant wrote on 11/24/2005, 11:55 AM
Laurence,

This is interesting... I'm considering the A1 so I've been reading posts about the HDV editing flow, and was wondering if for output to SD if the downconvert would do the job (and be easier).

Although, one thing I was thinking... if I edit (probably via proxy) the HDV, I could do a PTT of the edited HDV project, so later when I have a HD delivery method I could simply reauthor with that. I don't see how to do this if I would use the downconvert in the camera. Unless there is some clever way to apply the edits done on the DV to the HDV source later? Is this something you have considered?

Mark
farss wrote on 11/24/2005, 12:40 PM
The in-camera or VCR downconvert is quite good however if you're working in NTSC then the downconvert is to 4:1:1 which is then converted to 4:2:0 for the DVDs. Doing the downconvert in Vegas by encoding straight from the m2t files takes you from 4:2:0 to 4:2:0!
You might need to have a component interconnect between the DVD player and the TV to really see the difference but you certainly should be able to see it.
Things might be a little different if you're working in PAL projects, I've noticed very little difference going from DigiBeta at 4:2:2 to DVDs 4:2:0 compared to going from a DV25 file from the same tape.
Of course these are straight conversions, no edits or composites etc.
Bob.
Laurence wrote on 11/24/2005, 3:49 PM
I understand about the difference in colorspace, and was expecting to see a real difference in the color especially. I just can't though.

For me, after running this test, I will continue to shoot HDV, but I will only capture HDV if I think I am going to output the project in HDV. There is just not enough (or any that I can see) advantage in terms of extra quality to justify the extra hard disc space and rendering time of capturing and editing in HDV if the output is SD DVD.
PeterWright wrote on 11/24/2005, 4:45 PM
One advantage of using HDV captures for SD output is the ability to move around within the frame by zooming, panning or tilting, without losing quality.

Not necessary if everything's perfectly shot of course, but particularly with live events, this allows you to move from one point of interest to the next as if you knew it was going to happen!
Laurence wrote on 11/24/2005, 5:08 PM
Yeah that's true. especially to add a little movement to tripod shots. Another thing is deshaking with VirtualDub, though it takes so long you really need to be sure the shot is worth the rendering time. Another reason is to grab stills for the cover artwork. Another is to render full resolution PAL and NTSC in both 4:3 and 16:9 formats from the same footage. Rendering true 30p progressive for web distribution is another. All these are good reasons to edit HDV. Getting better SD output quality seems to be the least valuable of these.
Laurence wrote on 11/24/2005, 5:29 PM
I just ran a test of some HDV footage zoomed in about half way then rendered as SD. It looks just great...like it was shot that way! I imported the HDV footage onto a regular DV timeline and zoomed in. I'm amazed at how good it looked. The framing options for an SD project would be pretty outrageous! I guess I'll capture as HDV after all!
PeterWright wrote on 11/24/2005, 7:40 PM
Yes, there are all sorts of applications.
In Gary Kleiners V6 DVD trainer he demos a lacrosse match and using this method he was able to uncannily predict where each pass was going - looked very impressive.

Another simple example - videoing a choir or dance performance, when one performer suddenly sang or danced a solo part, you could "be there" in close up a split second before - looks like very careful pre-production planning! You can even cut between two different "inner" frames and make it look like there were two cameras.

-but don't get too attached to relying on this - when we start outputting HD we won't be able to do this without compromising quality.