JVC HD100 FAQ -- includes addressing split-screen

Barry,
I just attended a tech session presented by a JVC tech guy...He claims that the JVC has a tape that records HD and it has the required resolution of 1200X1440 or so....Is this right?
If one uses their tape in Hvx200 will it work? Curious as to why Panasonic cannot do what JVC claims can do.
 
Padmapriya said:
If one uses their tape in Hvx200 will it work? Curious as to why Panasonic cannot do what JVC claims can do.
Click to expand...
The tape itself is only a minor part of the equation. The recording format is primarily determined by the tape transport -- its speed and what information the heads lay on the tape. So, simply putting a JVC tape in the HVX200 will not "enable" recording in JVC's format.

You'll find many posts on this issue/concern throughout the site. In a nutshell, Panasonic has deliberately chosen not to put an HD tape transport in the HVX to help keep the cost down. This is one of the most expensive components in their higher-end offerings. (So, at least the option is available in their other products, if you can afford it.)
 
I thank you for the link, Barry.
Can you really say, that HD100 ratio is 2.69 (adjusted for 5:1 MPEG2 efficiency), while HVX200 ratio will be 6.7? Is such a comparison legitimate or just a "promotion gag"?
 
Let's say that I do not at all accept their interpretation of those particular facts, and Graeme Nattress isn't exactly buying their presentation either. He's involved in a fairly spirited debate over those charts at DVInfo right now.

The thing is, people are arguing numbers, when they should be looking at the footage. People have been saying that 1080i is so superior to 720p because of the fact that it uses twice as many pixels per frame -- but resolution charts are showing that both the HD100 and the Z1 are resolving about the same amount of detail! That's not something you'd expect if you just looked at the numbers, but it's what's actually happening.

JVC is of course presenting the information in such a way as it paints their product in the best possible light. Unfortunately we can't do a direct comparison between an HD100 and an HVX and see the footage yet, because the HVX isn't on the market yet. We can compare an HD100 against a Z1 and we can see that they both resolve about the same amount of detail, and the HD100 does have fewer compression artifacts. But (and this is a BIG "but"): the Z1 is resolving twice as much temporal resolution as the JVC! That's something you don't see in JVC's chart, but the Z1 is updating 60 times per second, the HD100 can't. So how much do you trust the numbers?

Arguing numbers won't work in JVC's favor anyway: if they want to compare numbers against an HVX, the HVX shoots 24p with 50% more luma resolution and three times as much chroma resolution, and no variable-frame interframe compression artifacting at all. If they want to stick to 720/24p or 720/30p (the only place where their argument holds any water), then they have to face that the HVX can deliver 60 full frames per second, and they can't.

As far as the actual comparison of "effective 2.69" vs. actual 6.7, all I can offer is actual real-world codec tests that show how well 720p HDV performs in its very, very, very, very best case (which is what they're basing their numbers off of), as well as in its very, very, very, very worst case (something you'll never hear them talk about, but a circumstance that's just as likely as the "best case" scenario they outline in their chart). Again, for comparison, DVCPRO-HD is always the "same" case, because it's frame-discrete -- there is no such thing as best-case or worst-case, it's always same-case, so I show only one picture for that. And in the best-best-best-best case, the HD100 is cleaner in mosquito noise, but still offers only half the color resolution. In the worst case, well...

Click here for a comparison picture: http://www.icexpo.com/dvx100/720vs720.png


That test was generated using the Apple FCP DVCPRO-HD codec and the Sony Vegas HDV codec. No arguing numbers, just looking at what the actual codec actually does.
 
The fact is, they get full raster HD 720P at that chip size. Everyone else interpolates, so there is quite an accomplishment from JVC. Getting full 1280 x 720 out of 1/3" is innovation. The compound chip idea may have tremendous possibilities.


We'll see...
 
Their 5:1 MPEG2 efficiency, is, as far as I can tell an absolute maximum on a still image. Last night the JVC page said "about 5 times as efficient". This morning it said "up to 5 times as efficient". How's that for revisionist history for you?

As far as I can see, from my own MPEG2 experience, you can get about anywhere from 1.5 to 2.5 image efficiency on real world images. You can get more with 2 pass VBR, but you can't do that in camera.

From my weekend with the JVC, using it in quite a few situations, it looks much more compressed than the Varicam footage I have, although DVCproHD720p24 is still too compressed for my liking.

Here's a real world image from the JVC I took in my front garden. It is from a medium pan over my front garden trees:
http://www.nattress.com/JVC_Artifacts.jpg
This is about as bad as it gets. Most footage looked an awful lot better.
Here's a much nicer image I took:
http://www.nattress.com/lego.jpg
I was pleasently surprised at how it picked up the dust on the model. It was shot in macro mode.

Graeme
 
I don't think the lens is sharp enough to justify full raster. They'd have got better results compressing a lower resolution image less, imho, or using the extra data to go up to 60p. When I get the camera back again I want to play with the 60p SD mode as I think that might be better still....

That said, full raster is nice to have, and to have the pixels keep square. That's certainly a "good thing". Although as I've said before - "I want high definition" which is not necessarily the same as "high resolution".

Graeme
 
Graeme, My reference is before the codec straight off the chip. Pixel for Pixel the codec is starting with a full 1280 x 720 scan to begin with. I've worked with the Z1/FX1 stuff for a year and the Sony implementation and JVC implementation have very different characteristics. Sony's motion estimation needs to be revised a great deal for their next camera. Higher bitrate is not going to cut it. They need to work on their motion estimation math for their GOP. Still the Z1 is an incredible value for the $$$$ in 1080i

JVC's motion estimation is extremely tight. No macroblocks unless in extreme cases. The JVC is quite a technical feat. With it's outstanding motion estimation, it is feature worthy (in my book). It wouldn't surprise me if some great feature motion pictures were made using it, or the next generation of it.

The workflow with the JVC ProHD is MUCH smoother than HDV2. To each his own and there are limitation to all sub HD(V) cameras in workflow and image quality.
 
Graeme_Nattress said:
Here's a real world image from the JVC I took in my front garden. It is from a medium pan over my front garden trees:
http://www.nattress.com/JVC_Artifacts.jpg
This is about as bad as it gets. Most footage looked an awful lot better.
Click to expand...

Is that what you're worrying about? With all the motion blur it is very hard to see the compression artefacts. As a matter of fact, I think it is not bad to have artifacts that are just that faint in such a high movement shot (judging from the image)...
 
Sorry, correction, I was viewind that jpg downsized and couldn't see the artifacts very well. In full res there a lot more obvious...
Still, doesn't motion mask this?
 
Nope, the motion doesn't mask it. I showed the tape to the JVC dealer, and they were quite horrified by it, almost as much as they were pleased by the quality of the rest of the stuff I shot. I like the way that MPEG allows you to get a very high quality image in a small data-rate, but I don't like how it makes you modify how you shoot to avoid it's problems.

As far as I could see, and I wasn't deliberately trying to stress the MPEG in tha shot, is that slow motion is ok, because it's slow, and fast motion is also ok, because motion blur creates a blurry image that's easier to compress, and there's a degree of masking due to the speed of it all. It's the "medium" level of motion, especially over high detail that's the problem. It's what you see in digital broadcasts of soccer matches, where the player on the field is ok, but the crowd is all macroblocky as the camera pans over it.

I'm trying to figure out a way to do a more standardized set of tests so that we can get a handle on exaclty what's happening, and how to avoid it, so that we can maximize the quality from MPEG based cameras.

Graeme
 
I did see some questionable motion in a shot of trees in the distance provided by Tim Dashwood. But overall, at a thrifty 19Mb the JVC codec is worthy. Graeme is right in that you'd have to be aware of shot composition in certain cases. In the 20 to 30 different m2t's provided on the web of HD-100 snippets, there was 1 that I saw questionable motion estimation.

With the Z1 we shot a goose taking off from a pond and the nobody could miss the macroblocks display on the front of his wings (viewed at 1080i). With the JVC codec I haven't seen this in similar situations.

Graeme, I'd be obliged if you'd post the m2t that your referencing for your tree/leave screenshot.
 
I'll see what I can do, as I captured it as native in FCP5, so it's no longer in m2t format. I've not played enough with the Sony to make a really valid comparison, but I thought the picture quality was very similar to that of the JVC, in both levels of artifacting and definition.

Graeme
 
This is not a defect. If present, these differences, a characteristic of HSTR, are visible on the LCD monitor and are easily eliminated by the user adjusting operating parameters (iris, shutter, picture content, illumination, etc.)
Click to expand...

I guess adjusting picture content means that it's the shot stupid, not the user right? :)
 
It means you have to adjust the shutter on the eyes in your head - yes, just shut those eyes and the split screen amazingly vanishes :-)

Graeme
 
Dear All,

I just bought a JVC HD 100. I didn´t have any dead pixel or splitscreen. I just think that the picture-quality is a little bit noisy. When you watch it on the monitor it is perfect. If you record it on DV-tape and then watch it it is kind of noisy. Reminds me of movies in the 70s.
Does anybody have a solution? I shot with no Gain, outside, no ND-Filter.
Thanx
Markus
 
The image is a touch noisey, but you don't really see that on the HD CRT monitor, but you do see it on the LCD HD monitor quite easily. I made a "feature" of it, and it helped with the film look I was going for. Without seeing your shots, it's hard to know whether what you're seeing is normal or not.

Graeme
 
Hi Graeme,

thanx for your reply. I watched the pictures on the computer TFT 17". It looks like Standart definition. To compare with my sony z1 makes a really sharp picture. Therefore I wonder if it is better to shoot with sony and the micro redrock adapter to get a sharp and filmlook style.
Cause most of my clients just watch it on PC-Screens.
Why do you think the picture is not that sharp on JVC? Because of the 720p instead of 1080p?
Or is there something in the menu that can make a sharper picture.

best

markus
 
it is 720p versus 1080i, actually, no 1080p... That's why most to all reports say that the JVC is SHARPER than the Sony... I don't know how your focus was and how the detail setting... and did you actually play it on full res?
 
You must log in or register to reply here.
Back
Top