Vegas Pro & HDV.

Vegascat40 wrote on 8/25/2008, 3:07 PM
I am a novice editor. I am trying to figure out the best worflow to use for an upcoming project. I plan on shooting 1080/24p HDV anamorphic using my Canon XH-A1. I eventually plan on delivering this project on standard def DVD video. My goal is to minimze transcoding and recompression by bringing the m2t files into Vegas for use in a 1440x1080/24p HDV project using that template. I will edit the m2t files natively without intermediates or proxies. I will render the project out using the 720x480 widescreen progressive scan template and then finish everything off in DVD Archictect.

I want some opinions on wheter you guys think this is the best workflow to preserve the color space by minimzing transcoding and recompression. Here are some things that bother me:

1. Can all DVD players recognize the final product as progressive scan DVD video? All the set top DVD players I own are progressive scan. I want the interlace conversion to take place in the DVD player at playback.

2. Would I loose color space resolution if I bring the m2t files into Vegas for editing using the 720x480 widescreen template?. This appeals to me because all the exhibits and pictures I want to use are not hi-def and would need to be upscaled only to be downscaled at rendering. I have tried bringing pictures into Vegas using the HDV template and with the pan & crop tool I selected match output aspect ratio and the results looked ok on the preview? What should I do here?

Any input would be appreciated. I don't want to spend long hours pursuing a workflow that will be problematic later. Thank you for your time.

Rgds/Steve

Comments

S35 wrote on 8/26/2008, 11:29 AM
Hi, I'm not a pro, so forgive me if I say something you know already.

OPTION A:

1. If you're shooting 1080-24p, don't forget your 24p stream will be encapsulated in a 60i stream. This means you'll need Cineform Neo HDV or something like it to capture and remove the 3:2 pulldown and get the 24p back.

2. Once you've done this, import your Cineform files into 1080-24p project in Vegas, right click each file (or select them all at once if you want), and press "h" (for switches) and then "d" (for disable resample). This step is probably not necessary for a 100% 24p workflow but is IMPERATIVE for OPTION B below.

3. Edit your files and render out to the DVD Architect NTSC 24p widescreen template with the following optional modifications:

(a) Make sure your video rendering quality is at Best.

(b) On the audio tab, check "include audio stream" so you can have the option of previewing your render in Windows Media Player with audio.

(c) Only you can determine whether or not to alter the bit rate setting, but I've found for complex encoding (such as dense moving shadows from leaves on a person's face) works best (no pixelation) at a CONSTANT bit rate of 8 Mbps. (Not tested with 24p yet, but I assume this is still true.)

4. Render out a separate 48,000 Hz .wav file and use it to replace the audio in DVD architect.

NOTE: Don't try to take your 24p video when it's still 60i and put it on the timeline. It will be a disaster! The 3:2 pulldown MUST be removed first, or you'll have "combing" (horizontal artifacts) in your rendered video. (Smart resample will fix this, but at the expense of EXTREMELY blurry video.)

OPTION B:

1: Okay, if you don't want to deal with Cineform, and to be honest I don't have it so I don't know how well it works, shoot in 60i with a 1/30 or 1/60 shutter speed. (Film yourself walking by your camera in both shutter speeds so you can determine what you think looks best when converted to 24p.)

2: Capture your files, and place in a 1080-24p project.

3: Right click each file (or select them all at once if you want), and press "h" (for switches) and then "d" (for disable resample). You must do this or your video will look blurry!!!

4: Same as #3 and #4 from OPTION A.


As for your question 1:

I don't think so. However, even a progressive scan dvd player can only show progressive if it connected to a television capable of producing progressive scan (such as an HDTV or EDTV) via a cable capable of progressive scan (such as component video or HDMI).
You need all 3 things (player, cable, television) to get progressive scan viewing.

As for question 2:

Not sure about that. Do some tests with HD and SD material mixed: one in an HD project, one in a SD project. Keep in mind that NTSC SD video is slightly wider than your HD source, so you will get slight pillarboxing left and right unless you select "no" in the "Maintain aspect ratio" option (in the pan/crop dialog) or select "stretch video to fill output frame size" (during rendering, [the easier of the two]) [which both result in slight horizontal stretching],or use the "match output aspect" option in the Pan & Crop dialog (for no stretching). (BTW, thanks for informing me of that option :-)

Hope this helps in some way, let me know if it does.

-Copperplate
Jessariah67 wrote on 8/26/2008, 12:45 PM
I'm currently editing a feature I shot on the XH-A1. Created proxies with VASST's Gearshift, and am editing on a wide 24P timelineusing the SD files, then will shift gears (one button switches everything out with the M2T files and render out to Wide 24P.
corug7 wrote on 8/26/2008, 5:46 PM
DO NOT shoot with a 1/30 shutter speed, as it will most likely result in a 29.97PsF file (resolution may be halved as well, I don't know how the Canons handle this). Trying to convert what is effectively 29.97p to 24p will not look acceptable.
Vegascat40 wrote on 8/26/2008, 11:11 PM
Lots of cool information here thanks everyone, especially you copperplate. One thing I want to comment on is that Canon has devised a very ingenious method of putting a 24p image in a 60i wrapper. My understanding is they are very hush hush as to how they accomplish this so efficiently. My understanding is that it can be treated as a fully progressive image and a 3:2 pull-down is not necessary. Neo HD and other intermediates are only necessary if your computer's processor is not fast enough to handle an m2t file or you plan on doing a lot of processing on the footage as in color correction. Thanks again.

Rgds/Steve.
farss wrote on 8/26/2008, 11:38 PM
I'd love to know where you got that information from. There's nothing new about putting 24p into 60i. Telecines have been doing that since day one of television.
You don't have to remove the pulldown in any 24p/60i video, it's not uncommon to not do so. The only issue is that dissolves etc get rendered as pure 60i and then the pulldown cannot be removed.

As far as I know the only thing Canon have done is to not flag the video correctly so that the pulldown cannot be easily removed by programs such as Vegas. Cineform use optical analysis to determine the cadence so pulldown can be removed.

Bob.
Vegascat40 wrote on 8/27/2008, 2:59 PM
Bob:

That was a topic that was covered in HDVideo Pro Magazine. You'll have to search the archives on that one I do not recall what issue it was in. But from what I recall it is the way Canon achieves this that makes it proprietary. I have rendered projects in their 24F mode without a 3:2 pull down and have encountered absolutely no problems. Why should I transcode everything in NEO and spend all those hours rendering those intermediate files for naught!

Rgds/Steve.
farss wrote on 8/27/2008, 3:47 PM
If you're talking about Canon's 24F mode there's no secret at all.
Canon call it 24F instead of 24p because they scan the CCDs in interlace and then de-interlace in their Digic processor. For this reason they decided not to call it 24p as they're not doing a true progressive scan of the imager. The only 'secret' is the algorithm used to do the de-interlacing in the camera.

On the T/L Canon's 24F should edit exactly the same as 24p from any of the XDCAMs or any of the other cameras that add pulldown, once you've removed pulldown.

Why it all is a bit of a kludge is the HDV spec doesn't cover 24p without pulldown. This means that 24F tapes cannot be captured using any available HDV VCR, only Canon cameras.

Bob.
Vegascat40 wrote on 8/27/2008, 4:50 PM
Interesting. Thanks for your input.

Rgds/Steve.
Laurence wrote on 8/27/2008, 8:39 PM
You may think there is no problem when there actually is. The problem with editing video with a 3:2 pulldown is that the 3:2 sequence gets messed up on the edits. About half the time it will mistakenly go to 2 when it should be at 3 and vice versa. It will look the same at 60i but can't properly be decoded to 24p on Bluray/HD TV combos that remove pulldown to give you proper 24p playback.
Vegascat40 wrote on 8/28/2008, 2:22 PM
Laurence I think you need to re-read, I am not editing with a 3:2 pull down. The whole idea is to edit natively and if the project gets played back on an NTSC SD analog television, to let the DVD player do the math.

Rgds/Steve.
farss wrote on 8/28/2008, 2:54 PM
The confusion occurs because you're perhaps not explaining yourself very clearly. The XH-A1 doesn't ADD pulldown into 60i. This is different to some of the other Canon cameras (HV20 etc) which do add pulldown that idealy should be removed. Unfortunately it would seem that on those cameras Canon has omitted to add the flags which would enable normal pulldown removal. The only option with footage from those cameras is to use Cineform's optical cadence detection.

Bob.
Laurence wrote on 8/28/2008, 7:05 PM
OK, I think I understand now. This particular camcorder must shoot real 24p mpeg2 instead of 60i with added pulldown. If that is what is going on, your approach should be fine.

An aside. My Sony HVR Z7 has two separate 24p settings which I believe give you both options (real 24p and 24p via 60i with added 3:2 pulldown). Can anyone confirm that that is indeed what is going on in the Z7 as well?
farss wrote on 8/29/2008, 4:01 AM
Don't have a Z7 so I'd need to read the manual if I could find one so I'm taking a bit of a stab at this.
I suspect when the Z7 records to the CF card you could use 24p with no pulldown as there's no real reason to add pulldown. When recording to tape it'd probably add pulldown to make the tapes compatible with VCRs etc.

Bob.
Vegascat40 wrote on 8/31/2008, 10:57 AM
Your input is welcome:

I was tweeking an LCD Hi-def 720p monitor using SMPTE color bars
generated Vegas. I noticed that there is anextreme difference in the default picture and settings at thecomposite input versus the component input to the monitor. At thecomponent input I can see all 3 IRE bars, and at the composite input
no matter where the brightness control is I can only Ssee the 11.5 IRE
bar? I thought SMPTE bars were designed to accomodate a composite
signal, but the adjustments I made using the bars generated through
the component input are far more accurate. I dont understand why there
is such a difference? I wonder if I should be using a Hi-def test
pattern as well? Does anybody have any thoughts on this subject?

Rgds/Steve...