The limitations of HDV!
 

I have an HDR-FX1 and wanted to share some of my issues with it. It helps me to know the limitations, so please chime in.
I've noticed some bad interlace looking lines during a shoot. I've found that it was an issue from panning the camera extremely fast. It hasn't been a problem with regular pans except when there is anything red in the frame. I have a pan of a temple and everything is fine except for a red column which streaks strange horizontal (interlace looking) lines. On another shoot I was whipping some red board to create wind on a green screen subject. The red board had awful artifacts when moving. Literally big red blocky artifacts trailed behind the movement and even stayed there momentarily.
Besides that the picture was really amazing. I'd be interested to know why people generally don't think HDV is very good. I know for example, that PBS doesn't accept HDV footage at all. Is it because the picture is just a bit softer?
By the way, I love my camera.

Glad you like your FX1, Aric; I love mine too.

It's the conventional wisdom that you can't do fast pans with HDV. That's wrong; you can. I do crazy whip pans for the sports videos I do all the time and it looks great.

Where are you noticing these artifacts? On direct playback from cam to HDTV via component? Or on your PC? Remember, PC monitors can't handle interlaced footage very well and don't do hardware deinterlacing the way real HDTVs do.

And if you dubbed your HDV to HDCAM tape, the pinheads at PBS would never know *how* you shot it.

This was informative. The playback however was to HDTV, I'm afraid the unusual horizontal lines and red artifacts are there to stay.
Two things that should be noted: The bad interlacing during the extreme pans was a shot of dark tree branches over light sky (very contrasty). I shoot with the shutter speed on 30, cause I kinda like the jittery-ness. It is wierd that on my normal pans that everything is smooth accept reds, they emit these interlace looking lines!?
But the blocky artifact trails where shot with the default auto shutter speed and they are freakin bad! I can upload something if you want, I just generally don't take the time to do that.

Have you considered running the footage through a deinterlacer like JES deinterlacer to get a 30p or 60p shot?

Interlacing artifacts aren't a result of HDV - they'd be there on any interlaced format. The other artifacts however are most likely a consequence of the heavy interframe compression required to squeeze HDV into the same bandwidth as DV. But - the crucial question - can you see these when watching the video at normal speed? Looking at individual frozen frames is inappropriate because our eyes are fooled when watching moving video and MPEG2 compression relies on that.

Auto shutter can go as low as 1/4 second, I believe. I just checked, and I can manually set shutter to 1/4 second on my FX1 (makes some pretty trippy images in the viewfinder, panning anywhere near fast). That could certainly explain "trails." I can imagine, that "trail" effect might be causing the codec to look at each frame as having no similar blocks, and make the recording really blocky.

Yes the problem did go away when I de-interlaced.
On closer inspection the OTHER block artifacts I was seeing are the same interlace trails, but they just stay in the air and float a while rather than disappear like this example:

Note: Link removed by moderator due to a flood of popups

This is a grab from final cut, see how red is bad and everything else is fine. I know not to trust FCP display, but since it was just one part of the image with a problem I thought it strange.
I don't really see any draw back to HDV.

The way both HDV and DV 'do' and 'deal with' colour is, IMHO, causing the problem.

The colour information is about a quarter of the resolution of the frame, washed over like a watercolour. Usually, you get away with this and the very low resolution is hidden in the detail of the picture.

Now, as soon as you get really strong bright saturated colours, the blockiness of the colour starts to define the shape rather than its luminance. Add movement across every pixel in the frame, stressing the bitrate budget further with interlacing, you get compression artifacts. Add lots of detail to THAT and you get a bit of a mess.

Your image contains everything: every pixel's changing because you're doing a fast pan. Your image contains lots of detail (foliage, rain, coarse textures). And of course there's the big red object.

Now here's where I get a bit wooly: traditionally, red has been a difficult colour for video, because as it's at the low end of the visible light spectrum and therefore long floppy waves don't make for sharp detail (unlike blue at the other end of the spectrum). In an RGB colour space it shouldn't make a difference, but in YUV (yes, everyone, imagine the little Log 1s in there) it may do.

However, I'm guessing not. you still get big blocks if pulling a standard chromakey, but not the 'damage' in your image. I also shoot and edit lots of corporate events and presentations - lots of blue washes on big sets, and in DV and HDV, I have to desaturate slightly to avoid big ugly boiling pixels spoiling the picture, so my guess it's the strength of the colour rather than the hue its self.

Just have to watch very strong colours, and be aware that scenes encompassing lots of movement (water, foliage, etc) can stress the codec.

Aric: I removed the link you posted to "imagebananna" because I got a whole flood of popup ads along with the picture when I clicked. If you want to share images please attach them to your post or use another site which doesn't subject us to so many unwanted ads.

Thanks!

I've been capturing with the apple intermediate codec lately, and I've been working with it enough to really see the difference between that and HDV. I had thought that the awful artifacting, compression, and horrible interlacing which was exaggerated on reds, would just be visible in Final Cut and QT playback. They weren't on the actual tape so I tried to ignore it. I see now, by comparing the footage in after effects that all that compression is apparent in "HDV", and not very apparent in "apple intermediate codec".
Capturing in HDV is so compressed I can't even look at it. I'll never capture in HDV again, but I do like the apple intermediate codec, and hope to someday be able to capture in pro res.
I do a lot of fast action, and landscape video, so maybe this is not a problem with documentary or corporate videos. But I just wanted to report that I would never dream of capturing HDV again, and I hope that when quality is most important -others can use this advice.