DVX100 clips

This is my first post and I hope it's okay for me to post links. But, I found a discussion thread on another forum where some guys have received their DVX100 and are posting uncompressed clips. Man, I'm totally blown away.

http://www.rewindvideo.com/cgi-bin/forum/YaBB.pl?board=news;action=display;num=1035995812

They're posting the clips so that you can run them through your editor and see them on an actual monitor. So don't be put off by the interlacing you see on your computer screen. Also, the guys that posted these clips are serious about their work. They know what they're talking about when it comes to cameras and they seem to be very impressed with this one.

I can't wait to get my hands on one.

Chris

[Chris Hurd]

Hey that's a cool site, thanks for sharing it!

Yeah, REwind is a great resource for any aspiring video/filmmakers out there. It started out as an information site and has quickly developed into a thriving community... kinda like this place.

Chris Sharpe

[David Lach]

I looked at some clips posted on the link above and more precisely at :

http://linnproductions.com/michael/

It looked good but then for the first time I could actually see the purple fringing some people were debating about earlier.

>>> in the uncompressed movie : baby.mov [184MB] <<<

I'm not talking about the fringing on the left side of the image but the pinkish transition artifact that appears between the mother's head and the highly contrasting window in the background (can be clearly seen at precisely 39 sec.). It looks like a very cheaply done blue screen shot...

I find this very ugly and would like some experienced DV techno buffs to explain if this would more likely be an improper setting on the DVX100 or a real design flaw (if anyone has a high speed connection or the patience to dowload it of course).

Thanks for any comments on this

[David Nussbaum]

are those all from the DVX 100?!?!?!

if so, im sold. sorry PD-150..

[David Lach]

It's just the "baby.mov" and the other clips that have "di" or "diana" in the title that came from the DVX100. Did you see those amazing colors in "uncompressed di.mov" ? It looks quite spectacular and this was done only using the light from the building. That's why it's quite frustrating to see that ugly purple fringe (I don't know if it's the proper term) over the head of the mother playing with the baby because the rest is just plain amazing. It seems to happen only when something is overexposed in the shot... I really wonder if this is just a matter of fine tuning a little bit more...

I hope.

[David Nussbaum]

i just DLed the uncompressed.mov, so im gonna watch now.

I watched the starwars clips there, and they looked pretty freakin good too. what camera were those on?

k, watching the uncompressed.mov... i wont post so i dont have to edit.


..............OMFG. just watched. Im sold. the color fringing doesnt bother me at all...even though it probably should...

IM SOLD!!!!!!!! OMG.


I have a Question... does it shoot that sexy out of the box? or are there some mods you need to do...

[David Lach]

Now that I look more carefully, and frame by frame, I can also see in "uncompressed di.mov" at precisely 33 sec. a color stairway effect when a rack focus is done from the building to the girl's face (which is quite nice too *** winking heavily ***). You can see this problem (if it's one) on her right cheek in the last 2 or 3 seconds of the clip. Although I'm not "DV expert" enough to know if this is due to color compression in the DV format, in camera problem or decompression problem...

Again I'm experienced with 16mm but relatively new to the DV scene so if ANYone that is using DV on a regular basis (or with different cams) could throw an opinion on all this, I'd be glad to read it.

[David Nussbaum]

i see what you are talking about. i didnt even realize it untill you pointed it out.

im no expert, but it might just be from the heavy interlacing which is present in the clips...

[David Lach]

That's what I thought at first but this clip (and the others as well) were shot using 24p. The interlacing comes from the pulldown the DVX100 does. If you watch frame by frame you can see a pattern of interlacing : 3 frames progressive and then 2 frames interlaced and so on... So i guess this would mean it was shot using 24p standard (not advanced) which uses a 3:2 pulldown (I might be wrong on this).

Anyhow even if you look at a progressive frame in the clip (meaning 2 interlaced fields taken at the same time) you can see that stairway clearly.

Maybe I'm just too picky... The thing is I'm starting a project that is meant to be blown to film so you can imagine how bad a thing like this would look on screen.

I know Stephen van Vuuren had 2 DVX100 so maybe he could comment on this and say if he was able to get rid of both the fringing and the stairway effect with some tweaking.

Thanks

[David Nussbaum]

ah, you are right. i see the pattern...

hmm...no clue. I know someone who directs and produces a TV show, and he has great knowledge of this type of stuff. I'll ask him when i have the chance.

[Ted Jan]

There currently is a discussion about this on the 2 pop's forum.

http://2-pop.5thavehosting.com/cgi-bin/discussion/forums/panasonicAGDVX100camera.cgi

Look for the thread from skiphunt titled "possible problem"

[Jeff Donald]

If you really want to evaluate the footage you need to import it into your NLE, then output to tape and view on an NTSC monitor. The interlacing hides some defects and makes others more pronounced. I've viewed the footage on a Sony broadcast monitor and Sony Wega consumer TV. The purple artifacts are present in the NTSC video. I would hazard to guess that it is related to the backlighting of the subject. I do not believe it is due to any camera settings or mis-adjustments.

However, I do not get the same effect when I shoot either of my XL1's in severe backlit conditions. I do wildlife documentaries and shoot birds of prey in flight. They can be severely backlit (open sky) and wings produce rapid motion. Never have I gotten that type of artifact.

Possible causes could be a cheap filter, lens, and signal processing. I viewed sample clips earlier this week and I did not see the artifact either. So it could be in just some of the cameras. However, we were looking more at the detail, sharpness, contrast and color accuracy. But if something as apparent as the purple artifact had been present I think one of the 4 sets of eyes would have spotted it.

Jeff

[David Lach]

<<<-- Originally posted by Jeff Donald :

If you really want to evaluate the footage you need to import it into your NLE, then output to tape and view on an NTSC monitor. The interlacing hides some defects and makes others more pronounced. Jeff -->>>


I try as much as possible to avoid watching the footage I'm editing on an NTSC monitor (I know it sounds strange) because it is meant for film transfer so I don't want to see what the footage looks like once interlaced on a tv screen but in true progressive images at 24fps.

I too have seen similar shots from the XL1 and there wasn't such problem. This color blur is really annoying and everybody is saying that the Leica lens on the DVX100 is the next best thing after the 24p recording so I would be very surprised (and I have to add disappointed) if this was caused solely by the lens. You might be right about the filter though...

I might reconsider buying this cam if I cannot shoot subjects with highly contrasting backgrounds. However I'm not trowing the towel yet because the rest of the footage is far too impressive to give up that easy.

[David Lach]

thanks for the link daremo1971...

I did find something interesting there and I'm just going to post the answer given by Michael from 24p.com here so that people in this forum can read it.

here it goes :

<<< I posted a response on this several weeks ago but here it is again in a nutshell:

Video cameras are tungsten balanced at the chip level. When shooting outside and the highlights are going to blow out, there is a pick fringe on the edges. This is because the blue spectrum of light is twice as strong outside (3200 -> 6500). In order to even out the response curve of the RGB in terh highlights, an 85b filter should be used in daylight situations just like when shooting film. When I look at the RGB channels when color correcting, all three now fall off at the same rate,eliminating the pink fringe. The white balance in the camera does not behave the same way the 85b filter does over the full spectrum of light hitting the lens.

I just completed a full day of testing the camera with a variety if charts while hooked up to a waveform/vectorscope and I will be writing these up on my web site one of these days (some intersting discoveries), but suffice it to say that I am still very impressed with a camera in this price range.

Just a quick comment on my camera ( I don't know if all XDV100's behave this way since I only have one) but in order to hit "perfect" on the scopes for black. white, and color - my F5 settings were corrected as follows:

1. Pedestal set to -1
2. Chroma phase set to -4

Detail,Chroma level adjustments are creative choices. More to follow.

Michael >>>