Would shooting in HDV and down-converting result in better quality?

[neolinski]

i know that shooting in HDV and down converting to DV in the post is much better than just shooting in DV.

what I am wondering is the hv20's internal mechanism of down converting (when I plugged in firewire to comp, there is the option of downconverting HDV to DV. so i'd get dv footage on the editing software directly.)

does that still give (much?) better quality than shooting in dv??


ps : you might ask me why. i just want to casually tape my friends goofing around and casually edit a little bit. so i don't want to do anything extra. like i don't mind DV quality. but if i can get something significantly better without doing more stuffs, i would want that.

[neocastillo]

HDV have the same about of data they just use different compression. I you are not doing any fancy editing, just putting clips together, I would say edit it the HDV format and then just export the final to DV or to a dvd.

[Atnas]

Agree with neocastillo, because when you edit in HDV and want DV in the end, you can stabilize your clips without loosing resolution.

[24Peter]

iwhat I am wondering is the hv20's internal mechanism of down converting (when I plugged in firewire to comp, there is the option of downconverting HDV to DV. so i'd get dv footage on the editing software directly.)



Tried that in Vegas 7 (capture my HDV footage as DV from the camera using DV lock and then editing DV on the timeline and outputting to SD DVD.) Looked significantly worse than just capturing HDV and outputting/rendering for SD DVD.

You should try it yourself as your mileage may vary.

[Lunchbox]

I think the best way is to record and capture in HDV. Downsize to DV with your NLE in the final step.

With such approach, you can also use your HV20 with digital zoom to 40x shooting and still get nice output in the final DV content.

I also read this tips online... someone mentioned shooting an interview scene. In the old day, you will need 2 cameras one pointing at each person to have multiple angle. With HDV, you can just shoot the entire scene with both the interviewer and interviewee in the frame. Now in the NLE duplicate the clip in roll-A and roll-B. Crop to 720x480 to each person separately. In that case, you have multiple angle of the same interview piece with only 1 camera. What a brilliant idea.

[Volstag]

As someone who has been experimenting with this a lot, I find that down converting the HDV footage, after capture, doesn't yield a very attractive picture. It's very "soft" and anything in the background (trees, clouds, etc) will show a disturbing amount of pixelation, along with a weird (and hard to describe) stop/start motion -- for example, clouds drifting across the sky will move in discrete chunks, instead of "flowing".

I capture HDV footage w/ HDVSplit (m2t), or inside of PrEL 3.0 (mpeg) and then export (to mpg, or avi -- doesn't seem to matter which).

What bothers me the most, is that I've read posts from users who claim that shooting in HDV and down-converting to SD will result in a "much better" picture than if it was originally shot DV. From my perspective, this is hardly the case.

I still feel like there's something I'm doing wrong.

-V

[pc001]

Volstag, It may have to do with the editing software you have and how it downconvert the HDV to DV. Simply, they don't do very good job (many of them, including Adobe Premiere 2.0 w/ MainConcept MPEG encoder). I found TMPGENC a lot better and faster.

Perry

[Volstag]

Perry,

Where can TEMPGENC be obtained? Commercial or freeware?

Thanks in advance,
V

[bluegrass]

As someone who has been experimenting with this a lot, I find that down converting the HDV footage, after capture, doesn't yield a very attractive picture. It's very "soft" and anything in the background (trees, clouds, etc) will show a disturbing amount of pixelation, along with a weird (and hard to describe) stop/start motion -- for example, clouds drifting across the sky will move in discrete chunks, instead of "flowing".

I capture HDV footage w/ HDVSplit (m2t), or inside of PrEL 3.0 (mpeg) and then export (to mpg, or avi -- doesn't seem to matter which).

What bothers me the most, is that I've read posts from users who claim that shooting in HDV and down-converting to SD will result in a "much better" picture than if it was originally shot DV. From my perspective, this is hardly the case.

I still feel like there's something I'm doing wrong.

-V

I fall into the majority group of those who see a very nice improvement in SD quality by shooting HDV and rendering down in post. Also, what spec are we using to mean SD & HDV. Isn't there a range for each of them?

I have to say that I think we're comparing apples to bananas when each of us gives our opinions of our own downrezed files. There are so many different capture, editing, and rendering paths the data goes through for each of us. Who's to say that maybe your data path is where the problem lies.

It might be interesting for a group of us to use the same target to shoot (any suggestions), the same camcorder configuration, and than document in detail the software path the data goes through to the final output file. We would need to try and render the same resolution and file format. We could than have a moderator post links to the various files (in one post) so we could visually make a comparison.

What we could do as a group for a benchmark, is all pull in into our RLE the same HDV file (probably 30 to 60 seconds would be adequate) and render it out into the same file format and resolution by all the various RLE's that this forum group is using than post the results for comaprison. This benchmark would bypass the caputre process but it might reveal something interesting about our RLE's. We wouldn't do any editing other than opening the file into a project and rendering it out with us all choosing the the same resolution/bandwidth & file type.

anyone interestd? if so, let's discuss what content might be the best to use for a benchmark comparison. if we do it, no hanky panky with the benchmark file - just open & render. we could also post some statistics with the rendered files, such as render times. this might be helpful in comparing the various RLE's.

i don't know about the rest of you, but there is starting to be a lot of duplication of information to wade through here. i think it would be nice to come up with some innovative ways to create new useful information to help us all become better video graphers.

i know some of the information here is dynamic such as hot new prices, etc. but wouldn't it also me nice to run all of the information that has been posted here through a funnel and come up with a section that contains the real "meat & potatoes".

examples:
1. 24p - shooting to final product.
2. audio - ways to improve and things to avoid.
3. video - how to get the best results in low light.
4. video - when to shoot in auto and when to shoot in manual.
5. video - when to use Cine, TV, AV, etc.
6. 24p - why?
7. tape - when to clean, what types of tape to use.
8. addons - wide angle, telephoto, filters.
9. 35mm adaptor - why?
10. 35 mm adaptor - comparisons

[Volstag]

Bluegrass,

I'll gladly help w/ any experiments anyone wants to conduct. Incidentally, I just watched some of your bluegrass festival footage -- on the one hand, I'm glad that someone out there is getting good down-conversions, on the other I'm incredibly bummed that I apparently lack the technology and/or know-how to do it myself. I'll keep plugging away at it though.

-V

[bluegrass]

Bluegrass,

I'll gladly help w/ any experiments anyone wants to conduct. Incidentally, I just watched some of your bluegrass festival footage -- on the one hand, I'm glad that someone out there is getting good down-conversions, on the other I'm incredibly bummed that I apparently lack the technology and/or know-how to do it myself. I'll keep plugging away at it though.

-V

I have a lot of on hands experience with computers but I don't feel that I'm very smart with video editing. I use a program from Pinnacle called Studio. My version happens to be 10.7. I've heard others here have had success with their latest product, Studio 11. I don't want to try and promote any product here but I will tell you once you get a solid version of Studio functioning, you probably won't find any RLE application that is easier and much more functional than this. You can get various levels of the program from about $70 to $120. That said their are other apps out there that don't cost much more.

One thing I caught in your previous post was that you output to an AVI file. You didn't mention what RLE you were using but were using the free capture program. I'm not sure why you would ever want to output an AVI file since that is completely uncompressed and would make a very large file. I would suggest that if your editing program will render a wmv or mov file you try that format. Each of us has a different idea of how we want to view or use the video we capture. You always want to consider how the video will be used. I've stated this before, but I'll state it again, the dirty little secret about HDV is there is hardly anyway to view true HD other than playing it via HDMI to a high resolution monitor at the present time. Of course we can all watch it on or computers but that shouldn't really be our goal for final viewing.

Once the Blueray and the HD DVD recorders come down to a consumer price and we have the software to create an HD DVD or Blueray disk we can play on or high def tv's than most of us will be where we want to be. Until that happens we can either watch or unedited tapes or render a DVD that can be played on or standard DVD players. I haven't actually created an SD dvd yet but I could. I'm rendering wmv files for a specific reason - for my website so that my website audience can easily view my videos.

I suggest you might want to invest in an app that supports HD and experiment around some more. You might want to try rendering some SD files in the mov or wmv format and see how they play on your computer. I didn't do anything special in creating the 640 wmv file other than capture it, a start & finish, add some title overlays and a few transitions and than chose the format and resolution to render the file in. Did you notice any artifacts or distortion in my sample video? My eyes aren't the best & I may be missing seeing in my file what you were saying you had with your down converted videos.

Most of us are willing to help each other learn about how to make our camcorders do what we want them to do. If you can give any more details on your project stream and show us a sample of the problem you're having, maybe we can help.

[neolinski]

simple question..
is there any quality advantage of shooting something in hdv and capturing in dv (using the camcorder's internal hdv->dv conversion)??

does that produce better quality than shooting with dv?

[wendxy]

I guess I fall into a minority group.
I shot and captured in hdv using Adobe premiere pro 2.0 and down converted
to sd using premiere pro 2.0 also. But I found that the quality is much worse
than I captured in sd directly. Not sure why.

[bluegrass]

I guess I fall into a minority group.
I shot and captured in hdv using Adobe premiere pro 2.0 and down converted
to sd using premiere pro 2.0 also. But I found that the quality is much worse
than I captured in sd directly. Not sure why.

Is there anyone out there who has experience managing project that could help develop a test plan for a shoot out among different applications. I envision someone posting two 30 second files that are shot within seconds of each other with the same settings except one being HD and the other being SD. Get these two thirty second clips (raw) posted where several users can download them open them with their NLE and render them to the same format and resolution. Have each resultant file rendered with different apps, posted for comparison.

If the project was developed with a stringent set of rules, we might get some revealing answers to several questions that have appeared on this forum about best approaches for shooting videio, including the "Is it best to shoot HD and render in post to SD, Shoot in SD and render to Hd (no just kidding - render to SD), Shoot in HD and force output to SD from camcorder. There is three scenarios, now multiply them by about 4 different NLE's and you have 12 different benchmarks to compare.

[propellerhead]

Here's my thinking:

DV & HDV Bitrate: 25Mbps
DV resolution: 720x480 (345,600 pixels per frame)
HDV resolution: 1440x1080 (1,555,200 pixels per frame)

An HDV picture contains 4.5X more data and subsequently must be more highly compressed than DV resolution in order to fit within the bandwidth available. Clips with motion are going to show more compression artifacts than static shots. It seems to me that for action sequences, shooting in DV resolution would be preferable to shooting in the more highly compressed HDV resolution.

Of course little of this will mean squat if you capture and transcode to some lossy codec as your intermediate.

[Lunchbox]

It's party true. DV is a intra-frame compression while HDV (MPEG2) uses both intra-frame and inter-frame compression.

I would still considering shooting in HDV instead of DV as there're more pixels to work with to begin with.

[wendxy]

Here are the files I used to test for hdv capturing and sd capturing.
the down converting used the same setting as the sd output,
the Adobe premiere pro 2.0 media encoder NTSC DV 16:9 high quality 4 mb vbr 2 pass. The sd captured by pp2.0 is in avi format.

Original footage:
http://hv20.info/yopu/canonhv20.mpg
Captured in hdv and down converted to sd
http://hv20.info/yopu/hdvcapture.mpg
Captured in sd and exported
http://hv20.info/yopu/sdcapture.mpg

The quality captured in sd and output to mpg is better than the one captured in hdv and down
converted.

[wendxy]

Here I uploaded my original test file and hdv capturing and sd capturing test results.

original file: http://hv20.info/yopu/canonhv20.mpg
captured in hdv and down converted: http://hv20.info/yopu/hdvcapture.mpg
captured in sd and converted to mpg: http://hv20.info/yopu/sdcapture.mpg

The output to mpg format for down converting and sd mpg output used the exactly same mpg setting-- Adobe media encoder NTSC DV 16:9 high quality 4Mb VBR 2 pass.
When captured in sd, pp2.0 saved the file in avi format. The hdv capturing was saved in mpg format.

[wendxy]

After some more testing. I found out that the best quality output can be obtained in the following work flow using Adobe PP2.0:

1. create a hdv project and capture the footage in hdv
2. create a sd project and import the hdv footage
3. edit the hdv footage and scale the hdv image to fit sd size.
4. output the project to sd dvd format.

[24Peter]

After some more testing. I found out that the best quality output can be obtained in the following work flow using Adobe PP2.0:

1. create a hdv project and capture the footage in hdv
2. create a sd project and import the hdv footage
3. edit the hdv footage and scale the hdv image to fit sd size.
4. output the project to sd dvd format.

FYI - You can do all that on one timeline/in one project with Vegas.