HDV versus DV as rec mode at DVinfo.net

Chris Westerstrom · March 20th, 2007, 09:31 AM

Hi!

Ever since I bought my HDR FX1e, I have been recording in HDV and capturing it as DV in my editing program. A few people I have talked to don't recommend this and I've hear of the news cameramen recording in DV on their full HD cameras.

Why is this? Is there any particular reason why I should do the same?

Thanks

Boyd Ostroff · March 20th, 2007, 10:21 AM

My experience with the Z1 is that you're better off continuing to record HDV and capture DV as you're doing. I think that gives a little cleaner results than recording in DV mode, and as a bonus you have the HDV source tapes if you ever want to use some of the material in an HD project.

Some people prefer capturing as HDV and downconverting with software. I guess it depends on your software and your methodology. I just put a small project together and captured HDV which I downconverted in FCP to make a DVD. I wasn't all that impressed with the results. If I didn't need the HD version then it would have been much easier and faster to capture the tape as regular DV. Rendering one hour standard definition from HDV on my dual 2.5ghz G5 took almost 4 hours....

Dylan Pank · March 20th, 2007, 11:25 AM

As they're news cameramen, it might be they're worried that they'll encounter situations (rapid movement, sudden image changes) that stress the codec, or worried that there could be a drop out that takes out a full 12 frame GOP. In such cases DV might be the better option, though in my opionion, HDV artefects are pretty negligable when shrunk down to DV res and I never really got the panic about the GOP thing because to me a drop out is a drop out. I don't see why one frame is hugely less of a problem than 12.

Chris Barcellos · March 20th, 2007, 12:18 PM

I think it makes a big difference in what you are converting with. Premiere seems to have an issue with encoding from HDV to DV. To get best results, to an ultimate DVD format, I do convert to DV first, but I make sure it is deinterlaced int that conversion. An interlaced conversion seems to look lousy. My unscientific impression is that I get better DV from an in camera convert that I shoot in HDV. I also now have Vegas 7, and while I haven't compared the two converted outputs there, I suspect that HDV conversion to DV might be cleaner in Vegas based on other encoding I have done out of Vegas. Frankly, I think Adobe has left Premiere hard to manage output in order to sell AE and Encore. (I love the editing format of Premiere, but have issues with the encoding process. I find myself editing in HDV in Premiere, outputting a finished .m2t file, and then converting to chosen encoder in Vegas7.)

Boyd Ostroff · March 20th, 2007, 12:30 PM

Quote:

Originally Posted by Dylan Pank

As they're news cameramen, it might be they're worried that they'll encounter situations (rapid movement, sudden image changes) that stress the codec

That's a good point Dylan, and now that I think about it they might also want to use DVCAM standard definition mode if they have Z1's (although the FX1 doesn't have that option).

If downconverting in software, consider the rendering time needed and also the fact that you need twice the disk space since you'll be storing both the HDV and DV versions.

Chris Barcellos · March 20th, 2007, 01:50 PM

Followup. One thing I sometimes forget to do is if I am going to keep the resulting DV file interlaced, you need to select "lower field first" in the encoding options.

Steven White · March 20th, 2007, 04:24 PM

A while back I did some HDV stress tests to try and demonstrate situations where DV would produce better results than HDV. Basically, there are none, because the data rate is the same.

At its very worst, long GOP MPEG-2 downsampled to DV looks no worse than DV. In almost all situations, long GOP MPEG-2 looks better.

I wouldn't downsample to the DV codec though. Downsampling to something like DVCPRO-50 would be a lot better.

-Steve

Gareth Watkins · March 21st, 2007, 04:14 AM

Hi there

I've been comparing several ways of getting decent SD footage from the HDV original... but my major issue with all the software solutions is the "time" factor...

To import the HDV, save out to SD by which ever method I tried took a long time... And this before I started editing... and all this for an SD DVD.

There was a very marginal, to my eyes, improvement in the best SD I could get with my software over the camera down convertion. But took a while to get..
This has lead me to the conclusion that while we can't supply HD to the punter, I'm better off using the camera to down convert, it is so much faster, as are the render times for the footage in SD.

When I get my new Dual core with loads of gigs of ram, I'll probably review this situation.

cheers
Gareth

John Miller · March 21st, 2007, 12:46 PM

Here's an experiment you can try:

With the camcorder hooked up to your computer and with a tape in, record to HDV on tape but capture the down-converted DV output to disk (1).

Now, play the tape back and capture to disk - again as down-converted DV from the camcorder (2).

Finally, capture the tape in HDV to disk (3).

Now, on the PC, convert the HDV to SD DV and compare to the two DV captures (1) and (2).

I would expect that (1) would be the best - since it should have been generated from the equivalent of I-frame only video. (2) and (3) have to be created from GOP sequences.

March 20th, 2007, 09:31 AM	#1
Chris Westerstrom Major Player Join Date: Jan 2006 Location: Gothenburg, Sweden Posts: 259	HDV versus DV as rec mode Hi! Ever since I bought my HDR FX1e, I have been recording in HDV and capturing it as DV in my editing program. A few people I have talked to don't recommend this and I've hear of the news cameramen recording in DV on their full HD cameras. Why is this? Is there any particular reason why I should do the same? Thanks

March 20th, 2007, 12:18 PM	#4
Chris Barcellos Inner Circle Join Date: Nov 2005 Location: Elk Grove CA Posts: 6,838	I think it makes a big difference in what you are converting with. Premiere seems to have an issue with encoding from HDV to DV. To get best results, to an ultimate DVD format, I do convert to DV first, but I make sure it is deinterlaced int that conversion. An interlaced conversion seems to look lousy. My unscientific impression is that I get better DV from an in camera convert that I shoot in HDV. I also now have Vegas 7, and while I haven't compared the two converted outputs there, I suspect that HDV conversion to DV might be cleaner in Vegas based on other encoding I have done out of Vegas. Frankly, I think Adobe has left Premiere hard to manage output in order to sell AE and Encore. (I love the editing format of Premiere, but have issues with the encoding process. I find myself editing in HDV in Premiere, outputting a finished .m2t file, and then converting to chosen encoder in Vegas7.) __________________ Chris J. Barcellos

March 20th, 2007, 01:50 PM	#6
Chris Barcellos Inner Circle Join Date: Nov 2005 Location: Elk Grove CA Posts: 6,838	Followup. One thing I sometimes forget to do is if I am going to keep the resulting DV file interlaced, you need to select "lower field first" in the encoding options. __________________ Chris J. Barcellos

March 20th, 2007, 10:21 AM	#2
Boyd Ostroff Wrangler Join Date: Dec 2002 Location: Mays Landing, NJ Posts: 11,787	My experience with the Z1 is that you're better off continuing to record HDV and capture DV as you're doing. I think that gives a little cleaner results than recording in DV mode, and as a bonus you have the HDV source tapes if you ever want to use some of the material in an HD project. Some people prefer capturing as HDV and downconverting with software. I guess it depends on your software and your methodology. I just put a small project together and captured HDV which I downconverted in FCP to make a DVD. I wasn't all that impressed with the results. If I didn't need the HD version then it would have been much easier and faster to capture the tape as regular DV. Rendering one hour standard definition from HDV on my dual 2.5ghz G5 took almost 4 hours....

March 20th, 2007, 11:25 AM	#3
Dylan Pank Major Player Join Date: Nov 2004 Location: Portsmouth, UK Posts: 611	As they're news cameramen, it might be they're worried that they'll encounter situations (rapid movement, sudden image changes) that stress the codec, or worried that there could be a drop out that takes out a full 12 frame GOP. In such cases DV might be the better option, though in my opionion, HDV artefects are pretty negligable when shrunk down to DV res and I never really got the panic about the GOP thing because to me a drop out is a drop out. I don't see why one frame is hugely less of a problem than 12.

March 20th, 2007, 04:24 PM	#7
Steven White Major Player Join Date: Nov 2004 Location: Canada Posts: 547	A while back I did some HDV stress tests to try and demonstrate situations where DV would produce better results than HDV. Basically, there are none, because the data rate is the same. At its very worst, long GOP MPEG-2 downsampled to DV looks no worse than DV. In almost all situations, long GOP MPEG-2 looks better. I wouldn't downsample to the DV codec though. Downsampling to something like DVCPRO-50 would be a lot better. -Steve

March 21st, 2007, 04:14 AM	#8
Gareth Watkins Major Player Join Date: Feb 2004 Location: France Posts: 578	Hi there I've been comparing several ways of getting decent SD footage from the HDV original... but my major issue with all the software solutions is the "time" factor... To import the HDV, save out to SD by which ever method I tried took a long time... And this before I started editing... and all this for an SD DVD. There was a very marginal, to my eyes, improvement in the best SD I could get with my software over the camera down convertion. But took a while to get.. This has lead me to the conclusion that while we can't supply HD to the punter, I'm better off using the camera to down convert, it is so much faster, as are the render times for the footage in SD. When I get my new Dual core with loads of gigs of ram, I'll probably review this situation. cheers Gareth

March 21st, 2007, 12:46 PM	#9
John Miller Major Player Join Date: Oct 2005 Location: Hillsborough, NC, USA Posts: 968	Here's an experiment you can try: With the camcorder hooked up to your computer and with a tape in, record to HDV on tape but capture the down-converted DV output to disk (1). Now, play the tape back and capture to disk - again as down-converted DV from the camcorder (2). Finally, capture the tape in HDV to disk (3). Now, on the PC, convert the HDV to SD DV and compare to the two DV captures (1) and (2). I would expect that (1) would be the best - since it should have been generated from the equivalent of I-frame only video. (2) and (3) have to be created from GOP sequences.