Things about the menus I dont understand

[David Slingerland]

There are some things about the menu I dont understand, who can help me?
I am still finding it difficult to understand how the camera writes to tape...
page 62 the video format menu....


In the menu there is mention of U model (USA) and E(europe) The E model is missing the DV-60I setting and HDV settings hdv-sd60p/hdv-sd30p.
Nothing wrong there...Both models have HDV-HD24P, why? The E model already has HDV-HD25P Film runs in europe at 25 frames and not at 24.....

I did a test with a sony 790 digibeta and tried to compare it to the JVC.
But I am afraid its not worth much, we looked at resolution and color...
The thing was that the testspace( a testbal, I dont now how you call it in english) was SD only so that was not fair to the JVC. The digibeta does not have component out so connecting it to a jvc-hd monitor. we could only see pal out of the digibeta. We did have component out JVC. The results then were that the JVC had a way better resolution, was much sharper on the charts... when we looked at the tape recording there was very little difference between the two. Color was almost the same (hdv25p)

But I am afraid the test results are worthless we needed a DVI card for the beta but did not have one last night....

[David Slingerland]

I have a few other questions regarding the manual. Maybe they were answered somewhere else but I have not found it...
How come the data stream of HDV is actually lower then that of DV?

HDV 19,7 Mbps
DV 25 Mbps

I think it is because the mpeg2 compression is more advanced?

Also does component output allow you the capture the raw feed if you had it connected to the proper stuff(what?)

[David Slingerland]

I found the answer to my last question about the raw feed:

"The ccd block already outputs 720p 50 /60 so live uncompressed HD can be obtained from the component output sockets. This could be sent directly to a monitor or projector or via an external component to SDI converter to produce an uncompressed HDSDI feed."

That also explains why the test with the digibeta was not fair....I hope someone else does it....

[Chris Hurd]

Hi David,

I split your posts out of Guy Barwood's "Missing Features" thread and created a new topic for you. Hopefully this will make your questions more visible so you'll get a better response.

[David Slingerland]

thanks chris,

It is a great little camera so far....

will you be going to IBC this year?

[Chris Hurd]

Hi David,

And I have a thread about IBC too! Please check in here:

http://www.dvinfo.net/conf/showthrea...ewpost&t=48128

Thanks,

[Barry Green]

Originally Posted by David Slingerland

I have a few other questions regarding the manual. Maybe they were answered somewhere else but I have not found it...
How come the data stream of HDV is actually lower then that of DV?

You are correct, the JVC HDV data rate is actually lower than that of DV.

MPEG-2 is a different type of compression than DV. Both are based on the same fundamental principals of macroblocks and DCT, but DV compression treats each and every frame individually -- you can almost think of it as a series of JPEG-style pictures. Whereas with MPEG-2 as implemented by JVC HDV, only one out of every six frames can stand by itself. The other five are all based on other frames, and they encode the differences between the frames. So, if you were shooting a static scene where nothing changes, HDV would encode its main reference frame (called the I-frame), and then the subsequent five frames would pretty much just point back to the I-frame and say "what he said". Those other five frames would take up very little data, so the vast majority of the bitstream could be devoted to encoding the I-frame, resulting in the highest quality HDV can deliver.

This can give MPEG-2 incredible efficiency, and allow it to encode 3x as many pixels as DV but in a lower data rate.

The downside is -- what happens when there's a lot of change between frames? MPEG-2 can run out of bits (a condition called "bit starved"). There may not be enough bandwidth to encode all the differences. When that happens, the image can start to break down, resolution drops, blocky noise artifacts can appear. And it'll continue for the entire Group Of Pictures. This is something that never happens with DV or any other frame-based compression. DV and its variants (DV50 and DV100), and MJPG, and DigiBeta, and Sony's IMX system, they're all frame-based with no GOP. You get consistent, predictable results -- but you usually need a higher constant bitrate to get adequate performance. With MPEG-2 delta/GOP compression you can get amazing results from low bitrates, but sometimes it gets overwhelmed and breaks. Watch an HDTV broadcast of a basketball game, or a diving competition, or a horse race, and you'll see how MPEG-2 breaks down when it can't handle all the detail in the rapidly-changing frame.

So as long as it's not pushed too far, MPEG-2 can be quite efficient, and the bitstream does seem adequate for most circumstances. JVC's implementation seems a bit more robust regarding avoiding a bit-starved situation than Sony's does, since even though JVC's data rate is lower (19mbps vs. Sony's 25mbps), JVC is attempting to encode far fewer pixels per second than Sony is (27.7 million pixels per second, vs. Sony's 46.7 million pixels per second) so the JVC actually has a higher data rate per pixel than the Sony does. Even so it can be pushed too far -- some of JVC's demo footage at WEVA showed dramatic macroblocking and a "quilting" appearance -- but most of it looked very clean. 19mbps doesn't seem like a lot, but it looks like it's adequate for most tasks the JVC asks of it.

[David Slingerland]

Barry,

do you als know why Both models have HDV-HD24P, The E(europe) model already has HDV-HD25P. Film runs in europe at 25 frames and not at 24 so why is there a 24p on an european model?

[Barry Green]

Well, the answer is: "why not?"

Actually all models of the HD100 have all three HDV shooting modes: 24P, 25P, and 30P. There's not a whole lot we in the US can do with 25P, but I'm glad it's there vs. not having it -- if I wanted to shoot a project for European delivery, I could (and not have to hassle with the 4% speed change).

I applaud JVC for including all three modes. I mean, more options = better, and if there's no good reason for not including it, then I'm glad they did include it. As a DVXUser member, I can't begin to tell you how many times people have complained that the European DVX only does 25p, and that it doesn't have 24p. I've tried to explain to them that 25p is the right choice for Europe, but they seem to think they're getting "gypped". Well, JVC doesn't have to worry about such claims, as they've given all of us all the modes. However, I do think that most European users would be much better served by shooting 25p rather than 24p.

IINM, the HD100E/U also both include 480/60p and 576/50p. That's not PAL and NTSC, those are HDV modes, but obviously 480/60p could easily be downconverted to NTSC and 576/50p can easily be downconverted to PAL. So while it's not a true universal standard camera, it does give the ability to shoot footage that can be used in both standards. You can't shoot PAL footage on the 100U, and you can't shoot NTSC footage on the 100E, but you can shoot footage on the 100U that you could easily convert to PAL, and you can shoot footage on the 100E that you could easily convert to NTSC.

Good move, JVC.

[David Slingerland]

Okay one more for you: when changing the frame mode in the camera it has three options 24, 50/25 and 60/30. A friend of mine said: he why is it called 50/25 and not just 25P maybe we are dealing here with 25 segmented frames and 30/60 30 segmented frames. And only the 24 is really true 24?

About the camera I am also really impressed. The only letdown for me so far is the viewfinder but I understand they could not make a better one for that price!! The so called focus assist also does not work to wel under lowlight conditions and if there are a lot of lines the complete shot lits up like a christmas tree! And it tells me that some parts are sharp were I know they are not...so it is easy fooled...

One more thing about the mpeg2 did you notice any mpeg2 artifact with fast motion or very 'busy' shots?

[Barry Green]

Originally Posted by David Slingerland

Okay one more for you: when changing the frame mode in the camera it has three options 24, 50/25 and 60/30. A friend of mine said: he why is it called 50/25 and not just 25P maybe we are dealing here with 25 segmented frames and 30/60 30 segmented frames. And only the 24 is really true 24?

It's called 50/25 and 60/30 because the live output is 50 (or 60) frames per second, and the in-tape recording is 25 or 30. There's nothing about segmented frames going on.

The so called focus assist also does not work to wel under lowlight conditions

Ah, interesting point. I didn't try it under low light. I would expect it to not be as good as under bright light conditions, because the basic underlying technology is similar to what an autofocus system would use (just without the motors on the focus ring) and autofocus also performs poorly under low light. But I didn't try it...

One more thing about the mpeg2 did you notice any mpeg2 artifact with fast motion or very 'busy' shots?

In what we shot, no. In the demo footage that JVC had running at the WEVA booth, yes, and in some cases it was extreme (a thick "quilting" effect happened on one scene where the camera was watching a duck or penguin or something, and the camera was looking through the rippling surface of the water to the mossy rocks underneath -- that part totally macroblocked and quilted).

But in what we shot, there's not a hint of it. And I think that's something that the mini35 *really* helps HDV with. When HDV chokes, it's because there's too much detail and too much changing between frames. The mini35 really assists because it lowers the detail level in the background, thus letting HDV devote its available bandwidth to where it needs it. I think the JVC HDV implementation (which already allocates more bits per pixel than Sony's does) combined with the mini35 (which softens backgrounds, letting HDV spread its bits to where they're most needed) is likely going to provide the most robust, artifact-free footage you can get from an HDV-format shoot.

[Scott Webster]

Originally Posted by David Slingerland

About the camera I am also really impressed. The only letdown for me so far is the viewfinder but I understand they could not make a better one for that price!! The so called focus assist also does not work to wel under lowlight conditions and if there are a lot of lines the complete shot lits up like a christmas tree! And it tells me that some parts are sharp were I know they are not...so it is easy fooled...

I would have to agree with your comments on the VF and focus assist in low light.

I'm hoping that Optex will develop an alternative hi-res VF for the camera much like they provided for the Canon XL series.