Working with HDV and DVCPROHD

adkimery

adkimery

Veteran
Okay, so I'm on a new gig and we use DV, HDV, and DVCPro HD. This is my first experience cutting w/these two HD formats and so far my experience w/HDV has been sucktacular. I'm on a dual 2ghz G5 and even a simple cross-dissolve takes forever to render in HDV. Is this normal? Working w/the DVCPro HD is a breeze (not much worse than working w/DV), but the HDV is so painful that I swear something is wrong. The other editors I'm working w/are like, "Yeah, it's that slow", but I just can't believe it. I'm using the easy setups but is there something else I'm missing? Or is HDV just that slow to render?

-A
 
Just to be certain, you're not working with HDV footage on DVCPRO HD sequences, are you? I don't mean to insult your intelligence or anything, it's just that you're talking about using so many different formats on one project and if you were in fact doing so, that will easily explain the especially long render times for the HDV footage.
 
Geoff_R said:
Just to be certain, you're not working with HDV footage on DVCPRO HD sequences, are you? I don't mean to insult your intelligence or anything, it's just that you're talking about using so many different formats on one project and if you were in fact doing so, that will easily explain the especially long render times for the HDV footage.
Click to expand...
No insult taken as we've all made the "duh" mistakes before. The seq settings do match the format. Although I'm was recently handed a project that has mixed HDV and DVCProHD footage and I was like, "Oh, hell no" and transcoded the HDV footage I needed into DVCProHD.

When I'm working in HDV and go to render an effect the dialog box that pops up says "conform" instead of "render". Is that normal? I thought I remembered reading that FCP needs to conform the seq before outputing to HDV, but I thought that was a final step before outputing to tape. So while working in HDV am I always "conforming" or is there a setting where I can make it just "render" and then I'll conform at the very end?


-A
 
Im a little confused in your first post you say your editing in HDV and its causing you a problem. The context of your post is mixing the formats.

If you convert all your HDV to DVCpro and insert into a dvcpro timeline with stuff that is originally dvcpro there should be no difference in render time at all.

Am I missing something?
 
Sorry for the confusing OP. I tossed in too much superfluous info. My question is this,
working w/HDV from a Z1U in an HDV sequence (easy setup 1080i60) is it normal to have long render times (compared to working w/DV or DVCPro HD)? For example, a 30 frame cross dissolve took about 30 seconds to render in my HDV sequence.


-A
 
I think this might be due to the nature of HDV. Think DVD encoding. (mpeg2 GOP) It's pretty tolling. Why don't you just transcode all media to one common format for editing?

DVCPRO HD or DVCPRO 50?

It would make it much easier to assemble a final edit.

Just my 2c

Gunleik
 
30 seconds seems way to long for a dual 2.0 g5. I'm thinking you're on track with the whole conform/render thing. I've worked with HDV once a while back and I don't remember it saying conform, it'd just 'render' dissolves or text overlays, ect. It was slower than DVCPRO HD but not as slow as you're getting. '

I thought the same, that the only time you had to re-confrom the GOPs was during export to tape. But maybe when you're throwing dissolves in, it has to go in and recalculate all the GOPs as well.
 
gunleik said:
I think this might be due to the nature of HDV. Think DVD encoding. (mpeg2 GOP) It's pretty tolling.
Click to expand...
That's what I assumed, but I never thought it would take as long as it does to render.

Why don't you just transcode all media to one common format for editing?
Click to expand...
We have pretty tight turn arounds so time constraint is a problem. I'd love it if we did all of our HD in DVCPro HD, but logistically I don't think it's possible. We do a lot of long records (live events, recording video game game play, etc.,) which means HDV gives us more bang for our buck.

I wish Canopus made a DVCPro HD ADC product so we could go Analog HD->ADC->Firestore and not even mess w/tape when we record stuff in the office.


-A
 
OK. But the bang/buck/time thing moves down the line, as you are explaining.

What about buying some hardware transcoding, like decklink or AJA and record directly to the format you'd like to edit in? You don't need Canopus to do this (even though I like Canopus)

Gunleik
 
gunleik said:
OK. But the bang/buck/time thing moves down the line, as you are explaining.

What about buying some hardware transcoding, like decklink or AJA and record directly to the format you'd like to edit in? You don't need Canopus to do this (even though I like Canopus)

Gunleik
Click to expand...
I definitely agree the bang/buck up front can have a cost further down the line. There is room for improvement in the workflow, I just don't know enough about the "big picture" to understand why we do things the way we do and what it would take the change it (I've only been here for a bit over a month). The cards are a good suggestion, I had completely spaced a solution down that path.


-A
 
Virtual-one said:
With my G4 this render 5 sec
Click to expand...
Hm. What version of FCP are you using and when it does the render does the pop up dialog box say "conforming" or "writing video"? Are you using the 1080i60 Easy Setup or a custom setup?

-A
 
First of all, you should have used the Apple Intermediate Codec instead of native HDV editing. I just finished an article for Event DV magazine about these issues. I got to speak to product managers for Vegas, Edius, Avid, I'm an FCP trainer, and Adobe would only answer questions, and with very incomplete, vauge answers, via email.

No one says their systems edit native HDV "as easily as other formats", period! Even Edius said you need top of the line horse power to do REAL HDV editing as easily as regular DV editing.

There is no image quality loss when using AIC in FCP. There's no conforming, and it plays well in Sequences using other formats. Trust me, anyone doing native HDV editing is wasting time, no matter what NLE you use. The Intermediate Codecs are still very important, and are really nice to use.
 
I'll look into using the AIC. I thought there was a quality hit w/AIC, but considering I'm working on video for the web I think that's a moot point anyway.

I think I might have narrowed down part of the problem. I think my machine at work is always trying to "conform" the timeline instead of just "render" the effect when I hit "render." I did a quick test at home (on my ancient dual 1ghz G4 tower) and it rendered much faster than my work machine (a dual 2ghz G5). On my home machine when I hit render the pop-up dialog box said "writing video" but on my work machine it said "Conforming HDV Video". Both machines are running FCP 5.0.4

-A
 
There is not image degridation by any NLE using HDV Intermediate Codecs. Technically on paper you can show some very tiny bit of theoretical degridation, but no one's shown for sure on monitors that there is any.

When you edit native HDV on anyone's system, it can only place cuts on "I-Frames", not the P or B frames. So when you make a cut, the NLE has to work it's butt off in the background to recalculate the MPEG-2 GOP structure and "conform" the frames to be I, B, and P frames in the correct order, and in the correct places. If your cut is on an I or B frame, and the NLE can't make it fit within the native GOP, it will force it to the nearst frame that can "legally" be an I frame. It is a tremendous amount of calculating for the computer to do. And if you start to composite more than one track of video, use filters, etc, most NLEs, according to the product managers I had the privledge of interviewing recently (Avid, Sony, Apple, Edius, Adobe would only email me answer and refused a telephone interview, go figure), it all bogs down pretty quickly in everyone's NLE.

Thus, the AIC makes everything I frames, so it edits just like regular DV, no more conforming. So use the AIC for sure, it'll make your life so much better.
 
BenB said:
There is not image degridation by any NLE using HDV Intermediate Codecs. Technically on paper you can show some very tiny bit of theoretical degridation, but no one's shown for sure on monitors that there is any.

When you edit native HDV on anyone's system, it can only place cuts on "I-Frames", not the P or B frames. So when you make a cut, the NLE has to work it's butt off in the background to recalculate the MPEG-2 GOP structure and "conform" the frames to be I, B, and P frames in the correct order, and in the correct places. If your cut is on an I or B frame, and the NLE can't make it fit within the native GOP, it will force it to the nearst frame that can "legally" be an I frame. It is a tremendous amount of calculating for the computer to do. And if you start to composite more than one track of video, use filters, etc, most NLEs, according to the product managers I had the privledge of interviewing recently (Avid, Sony, Apple, Edius, Adobe would only email me answer and refused a telephone interview, go figure), it all bogs down pretty quickly in everyone's NLE.

Thus, the AIC makes everything I frames, so it edits just like regular DV, no more conforming. So use the AIC for sure, it'll make your life so much better.
Click to expand...

Cool, thanks for all the good info Ben:beer:


-A
 
So, I've been reading up on using AIC (haven't had a chance to do my own tests yet) and it looks like a couple of short comings are that I'll lose my TC from the original tape and that while progressive footage looks good captured to AIC, interlaced footage captured to AIC doesn't look as good (and of course we are using Z1Us). Are these issues still there, or is my info dated?


-A
 
AIC is an HDV Intermediate Codec. The few folks I know using it regularly tell me it works great. Never saw anyone lose timecode or have problems.
 
You must log in or register to reply here.
Back
Top