Practicalities - working with HDV (all cams, all NLEs) - your experiences

filmguy123

Veteran
Ok, I've been using my HVX for a while now, and avoided HDV due to the MPEG-2 long-GOP compression. However, I decided to read up on this and look into things, after reading about the Sony EX1, and Panny's new AVC-HD camera (HMC150).

I've heard many people talk about avoiding HDV or any long-GOP compression and all the problems with it. However, is this *REALLY* that big of a problem for most people anymore?

NLEs handle native HDV editing and modern computers are fast enough where actually EDITING the footage isn't much of an issue in terms of speed.

You can also transcode your footage to a different codec and avoid working with HDV footage alltogether. I learned that in FCP, you can transfer from a tape and encode to Apple ProRes422 ON THE FLY as it is streaming onto the system via firewire. Yes, this takes more space, but no more than working with DVCPRO-HD.

That would also seem to solve any keying or motion tracking problems?

At that point, it seems the only real issue working HDV (or AVCHD, or any long-GOP format) wouldn't even really be in the editing anymore, would it? Is it mostly just the footage captured?

And in terms of the footage captured and recording to long-GOP... I hear about the theoretical "problems" of this with fast motion or action stuff, but yet I know a *LOT* of people are shooting HDV stuff for all sorts of uses... is this really a practical issue? Shooting sports highlights or action films or high action music videos, etc. Is HDV or long-GOP footage (in the camera) really produce a lot of artificating or problems, or is it really a moot point?

And then, if you're ok with the increased HDD space, you can even just edit in ProRes (taken care of direct via capture in FCP), and/or native editing is pretty robust now...


SOOO... practically speaking, are there really any "real" issues with the HDV workflow? What are experiences here? What kind of person really *needs* to avoid HDV/long-GOP footage, or has it become a very strong and effective method, making Intra-frame "nice", "superior", but not really needed for most practical purposes (even speaking of some effects work/keying - I've heard people have good results with HDV/long-GOP keying/effects?)
 
I shoot on HDV and the workflow is just fine for me.

People can analyze specs all day, but I can't possibly imagine any average person watching footage shot on HDV and noticing artifacts.

FWIW, my Canon XHA1 footage is sharper and cleaner on our HDTV than broadcast HD.
 
I've been using HDV since I got my first prosumer cam, and in my experience most of the problems that used to be associated with HDV have been mitigated to the point where they don't matter (for most people). Recent implementation of the codec, such as in the XH-A1, have been very robust.

In theory, you can overload the bandwidth with fast motion. In reality, I haven't seen anything like this happening at all, even when I shoot fast action scenes. We've all seen Barry's side by side grabs from the HD100 vs. the HVX where the HD100 shows massive macroblocking when shooting fireworks. I believe that most of us are unlikely to encounter this scenario.

The single biggest problem with HDV is in the editing stage. While for short projects it won't matter, it is my experience that large amounts of footage cause my NLE (Premiere Pro) to become exceptionally unstable. For my last two films, I was working with about 6-7 hours of raw footage. On an AMD 3.2 Ghz Dual Core processor with 2 GB of RAM, Premiere eventually becomes so unstable that fully a quarter of my time in post was simply restarting the program and loading up the project again. For this reason I would never be able to edit a feature on HDV.

The other end of it is VFX work. It's simply impossible to get real time preview (even at low res) on HDV in After Effect, precluding all but the simplest things. Fortunately I won't have to deal with this anymore since I just purchased Cineform, which I expect will speed up the process considerably.
 
Yes, HDV isn't friend of fast action, zooms, fast pans, but the artifacts are almost unnoticeable by the average audience.
I recommend to edit in a ProRes timeline, and the workflow will be a joke.
Color correction will be great too.
 
Yes, HDV isn't friend of fast action, zooms, fast pans, but the artifacts are almost unnoticeable by the average audience.
I recommend to edit in a ProRes timeline, and the workflow will be a joke.
Color correction will be great too.

You mean the artificats in shakey cam, fast action, snap zooms, quick pans, etc. are almost unnoticeable to the average audience nowadays?

Yeah, I'd be thinking of editing in a ProRes timeline - which means I would be working with Intra-Frame footage in post, correct? (but may still see any artificats that were present from the original image captured in the camera, although from what I am hearing is practically a non-issue nowadays... is that right as well?)
 
For me, using HDV is a rose and a thorn. For me, my biggest complaint would be the low light situations in HDV. I never have a problem with the camera,except night time/ low light situations, I literally cant seem to get anything I see as OK. Its always noisy, or almost unuseable for me.

The other problem I have been running into is editing. I would prefer to not have the Mpeg2 encoding, but I deal with it for now, but my computer hates it. I have about 15 hours of HDV footage on my hard drive, mixed with about 40 hours of DVX footage and my computer (Dual 3.0 G5 etc) hates it. I have to render everything everytime I do anything (my settings, but I need good previews).

Other than that, I love it, although I prefer my dvx still.
 
Yes, HDV isn't friend of fast action, zooms, fast pans, but the artifacts are almost unnoticeable by the average audience.
I recommend to edit in a ProRes timeline, and the workflow will be a joke.
Color correction will be great too.

How long does it take for you to move the HDV from your XH-A1 to ProRes(HQ) 4:2:2 and what does the average Gigabyte of HDV equate to in ProRes(HQ)?
 
the only problem I have with HDV is how sh*tty is looks when you try to color grade it. unlike other formats, i find myself fighting between the look I want and the image falling into a blocky disgusting mess.
 
How long does it take for you to move the HDV from your XH-A1 to ProRes(HQ) 4:2:2 and what does the average Gigabyte of HDV equate to in ProRes(HQ)?

i can't remember exactly, but i think it more than doubles or triples in size.
 
The problem again is this stupid lumping of all HDV cameras and codecs as the same--they're not! Some manufacturer's codecs are better than others.

I think it breaks down like this:

Best HDV codecs: Canon and JVC

--Canon 24F HDV is the most resistant to artifacting. (progressive 24fps encoding) 60i is still fairly robust for most situations, but interlace MPEG2 encoding is less efficient. With that said, very few people complain about artifacting with standard 60i on the Canon's. Not so with Sony Z1, FX1 or V1.

--JVC 720 24p is fairly robust too, particularly with the updated encoder on the higher-end models. It has the same advantage of progressive encoding, along with a shorter GOP structure and fewer pixels to encode. (6 GOP in 24p)


There are lots of ways to go in post, and there seems to be an interesting mix of nearly half the users out there being just fine with native HDV editing and the other half preferring to work in an intermediate.

My own personal opinion is that with Final Cut Pro and a relatively fast Mac, native HDV editing is quite acceptable. Where things get slow is when you have fx-intensive edits with lots filtering/motion fx/compositing, etc, in which case transcoding to ProRes solves a lot of the issues. Standard ProRes is roughly the same data rate as DVCPRO HD (sometimes a little more or less depending on the content of the imagery--it's a lot like PhotoJPEG in that regard) and ProRes HQ will make the data rate jump another 50-65% from there.

I've done extensive tests comparing camera HDV originals to ProRes conversions and I can say that it's essentially visually lossless. (standard ProRes is fine fwiw--no need to use the HQ mode for most things)

Where ProRes really helps is when you do multiple generations of renders on material, and of course it's nice to work in deeper color space when doing greenscreen, CC, etc. It's also less time consuming to export a ProRes timeline, particularly if the edit is, again, fx-intensive. The "conform" process of HDV rendering (rebuilding GOP structure) can be very time consuming, especially with older computers.

I think you'll find that most don't get a "blocky, disgusting mess" with HDV in post if they're careful with the footage. I'm particularly happy with the way I can push Canon 24F HDV around in color correction. I think most A1 owners are fairly happy in this department too.

Lastly, with the latest Macs being so fast, the easiest thing to do is simply buy plenty of hard drive space and convert everything to ProRes upon ingest. Edit and finish in ProRes and you won't have much to worry about.
 
Last edited:
I'm not sure the artifacts are "unnoticeable" by the average person. To some, maybe, especially based on the degree to which they're visible. But I've seen "average people" notice them, indeed.

Now, the average person may not CARE. But that's another issue.
 
Just say no to HDV, sign on to cineform and you'll get discrete frames and you can edit it with pentium 3. It just takes up a bit more space. But so what, hard drives are dirt cheap. Again, intermediary codecs like CINEFORM. Works great.
 
How long does it take for you to move the HDV from your XH-A1 to ProRes(HQ) 4:2:2 and what does the average Gigabyte of HDV equate to in ProRes(HQ)?

After capturing in HDV format, i just drag the HDV clips in the ProRes sequence, and Boom, long gop disappears, 4:4:4 appears and i did a test, the compression becomes better, the color blocks become smaller by 1/10.
HDV is 25Mb/s, ProRes is almost 10 times more.
 
Elton - when ppl have CC problems, is this eliminated by going ProRes, or mostly or partly eliminated...?

It doesn't necessarily eliminate them, but it gives the NLE or Afer Effects, etc, more headroom for corrections and interpolations of color. It's just a container for more information than standard HDV could display.
 
Back
Top