Converting HDV to DV - Getting Best Quality

[Earl Thurston]

I was running some tests tonight to convert HDV to DV. It's been said that doing so is not as good as just shooting DV in the first place; the DV codec and down-scaling required softens the image too much, whereas shooting native DV directly in the camera creates a sharper image right from the start.

Initially, this appeared to be true. My first test was opening some HDV files taken with my JVC HD100U (and converted to Cineform AVI) in Premiere Pro 1.5.1 and scaling them down 68% in an NTSC DV widescreen project. Upon rendering, the images definitely looked soft.

Then I tried the same thing again in After Effects 6.0. Noticeable improvement. It did a much better job of retaining detail from the HDV source.

In both tests, the footage was rendered to DV AVI files and then reopened in Premiere Pro for capturing stills, which are posted here in TIFF format:

HDV to DV via Premiere Pro
HDV to DV via After Effects

Now that I've determined the best method for down-converting, my next test will be to shoot some identical footage in both DV and HDV to see if the native DV footage is indeed better than the down-converted HDV footage from After Effects. Stay tuned.

[Peter Robert]

Your links are dead.

[Earl Thurston]

Originally Posted by Peter Robert

Your links are dead.

They're working for me. If you're still having trouble, try these instead:

HDV to DV via Premiere Pro (alt)
HDV to DV via After Effects (alt)

[Kevin Shaw]

Originally Posted by Earl Thurston

Now that I've determined the best method for down-converting, my next test will be to shoot some identical footage in both DV and HDV to see if the native DV footage is indeed better than the down-converted HDV footage from After Effects. Stay tuned.

How about comparing to the quality of HDV downsampled to DV directly out of the camera? I've been using this approach for editing some of my HDV footage from a Sony FX1, and that seems to work fine.

It's also educational to compare HDV encoded directly to SD MPEG2 with DV encoded to SD MPEG2: when I tried this a while back the HDV solution produced visibly better results. So if you want the best quality from HDV, capture and edit at full resolution and downsample to SD at the end of a project. If you don't want to fuss with full-resolution editing for a project, just downsample to DV out of the camera and save your HDV tapes in case you ever want to refer back to them.

One thing for sure: in terms of image clarity, HDV is way ahead of DV when viewed at full resolution. I'd have to have a good reason to shoot DV instead of HDV on an HDV camera, regardless of how I think I'm going to edit it.

[Earl Thurston]

Originally Posted by Kevin Shaw

How about comparing to the quality of HDV downsampled to DV directly out of the camera?

From what I understand, the HD100 can't do this through the FireWire port. I thought it could at first but then someone (in another topic) mentioned it wasn't possible, so I haven't tried it. (Anyone know for sure?) You can do it through analog if you have a capture card; however, I don't and JVC even says in their FAQ that the quality isn't as good.

Originally Posted by Kevin Shaw

One thing for sure: in terms of image clarity, HDV is way ahead of DV when viewed at full resolution. I'd have to have a good reason to shoot DV instead of HDV on an HDV camera, regardless of how I think I'm going to edit it.

Oh definitely. If you're interested (and haven't seen it yet) I posted some comparisons from my old GL1 and the HD100 taking the same footage. Huge difference. They're at:

http://www.neopics.com/gl1-hd100/

[Craig Terott]

Originally Posted by Earl Thurston

I was running some tests tonight to convert HDV to DV. It's been said that doing so is not as good as just shooting DV in the first place; the DV codec and down-scaling required softens the image too much, whereas shooting native DV directly in the camera creates a sharper image right from the start.

Initially, this appeared to be true. My first test was opening some HDV files taken with my JVC HD100U (and converted to Cineform AVI) in Premiere Pro 1.5.1 and scaling them down 68% in an NTSC DV widescreen project. Upon rendering, the images definitely looked soft.

Then I tried the same thing again in After Effects 6.0. Noticeable improvement. It did a much better job of retaining detail from the HDV source.

In both tests, the footage was rendered to DV AVI files and then reopened in Premiere Pro for capturing stills, which are posted here in TIFF format:

HDV to DV via Premiere Pro
HDV to DV via After Effects

Now that I've determined the best method for down-converting, my next test will be to shoot some identical footage in both DV and HDV to see if the native DV footage is indeed better than the down-converted HDV footage from After Effects. Stay tuned.

You should know that Premier Pro rated worst/poor in EventDVs downsample tests (and this is from a publication with a reputation for being careful not to step on anyone's toes).

Here's the link:
http://www.eventdv.net/Articles/Read...0446&PageNum=2

[Chris Barcellos]

That test was of Premiere Pro 1.51. 2.0 is newly revamped.

[Robert M Wright]

If you want some control over what method of resizing is used, I'd suggest using VirtualDub and it's internal resize filter for the downsize conversion (you have a few choices - one of the Precise bicubic modes or Lanczos3 mode will work best for this type of resizing).

[Chris Barcellos]

I'm using Premiere Pro 2.0, and I've edited some 27 minutes of air show footage as a first full blown test of Premiere Pro 2.0. I rendered the file to a new HDV .m2t file. I then down rezzed to Standard DV widescreen, and then rendered that to a DVD using the Sony DVD Architect. Came out pretty nice.

But my question is: Do I need to take the extra step to DV Wide first, or would it be a better product if I just rendered to DVD format in first place ?

[Miguel Lombana]

Originally Posted by Chris Barcellos

I'm using Premiere Pro 2.0, and I've edited some 27 minutes of air show footage as a first full blown test of Premiere Pro 2.0. I rendered the file to a new HDV .m2t file. I then down rezzed to Standard DV widescreen, and then rendered that to a DVD using the Sony DVD Architect. Came out pretty nice.

But my question is: Do I need to take the extra step to DV Wide first, or would it be a better product if I just rendered to DVD format in first place ?

Chris what preset are you using in PP2.0 to render out the m2t? I'm miffed by 2.0 as I've still not been able to get a quality HDV export. Any help would be greatly appreciated.