Split screen technical explanation

[Barry Green]

Originally Posted by Edwin Huang

Numerous people posted that JVC acknowledged that it is a fault and if you find the SSE at something like +12 gain or so they'll replace it? Some say 0 gain? Other people are complaining that JVC does not acknowledge it as a problem. Which is it?

Some people post that only the 100E and 101E have the problem, the 100U is mostly SSE free because of newer firmware. This is also contradicted. In fact there are posts that claim no amount of firmware will help. Then ones that claim no new firmware has been released. Which is it?

Part of the confusion is that we are a worldwide community, so you're hearing reports from people in Australia, Italy, Canada, the US, England, etc. Yet the camera is marketed by different divisions of JVC, each which can set its own policies. So it's possible that JVC Canada may declare "any split-screen that's visible at less than 12dB is a defect worthy of replacement", yet that doesn't bind JVC USA or JVC Australia to the same conclusion. There is no blanket statement of policy coming from JVC Japan, the parent company of all the JVC divisions.

So what a JVC Canada customer gets told may only have bearing on customers in Canada, and may have no applicability to customers elsewhere.

Now, regarding US units -- the camera was released overseas a full month before it was released in the US, some people had it overseas for two months before some US customers started getting theirs. It is our (perhaps unfounded?) understanding that the US units were being delayed while an extensive quality-control check was implemented. So Steve takes the position that reports from non-US cameras are not necessarily relevant to US purchasers, and I think he does have a point. However it goes forward, though, it should be reasonable to say that what US customers are experiencing is the most up-to-date, most thoroughly quality-checked versions.

My opinion. If they are replacing the cameras officially or unofficially, I don't care as long as they are replacing them consistantly.

Well, that's part of the problem -- consistency only carries as far as the borders of the particular marketing division's territory. JVC Canada can implement a policy of "we'll replace everything", and JVC New Zealand could decide "we'll only replace it if we can recreate the situation in our lab" and JVC USA may decide to say "we don't consider it a defect". Each country is different. Each marketing division is different. And the policies and standards that they decide to adhere to are not necessarily binding on any other territory.

The only thing you can do (and the only answer that is actually relevant to you, the individual customer) is to ask your dealer what their policy is on the split-screen issue. Your particular dealer will be able to converse with their JVC sales representative and will get the straight facts on what options are available to you, the customer. Don't come to the web looking for the answer, as there are too many different answers (and each of them right for each territory, but not for others!) Ask your dealer. Tell 'em you've heard about the SSE and you want to know what the policy is regarding that. That's the only way you'll know what matters to you in your country.

[Barry Green]

Originally Posted by Stephen L. Noe

How did the HVX make it into this thread?

Guilty -- sorry, didn't mean to get things off-track. Since I'm getting both, I was pointing out that they're very different and separating the two cams into what they're individually best at, and Steve made the point that "why would event shooters get this cam? It's made for filmmakers", so I was trying to say that the reasons I got it had nothing to do with filmmaking, it's not necessarily the optimal choice for that environment; instead I was saying that its unique selling points had more to do with the appeal it has (warranted or not) for event shooting. Tied in with, and taken from, Douglas' post above.

So no, no reason to go off on that tangent, and I apologize if it drags the conversation even further off tangent!

[Greg Boston]

Barry and DSE are very correct. In certain market segments, your camera's formfactor will either put the client at ease, or, make them think you are just as amateur as the guy who offered to do it for free before they decided to hire a professional.

I know that's not the be-all end-all of creating good video, but it is a reality. The most common question I get with the XL2 rig is, "What tv station are you shooting for?".

As for 24P wedding videos, the better quality videos often resemble a short film.

=gb=

[Stephen L. Noe]

James Daniels (aka J.D.) District Sales Manager JVC Corporation says he'll be at Resfest Chicago with camera. Here is a nice opportunity to check out the HD-100 and talk to "the horses mouth". So bring your interest in HD as well as your questions.

This was confirmed to me via email so it is not on the Resfest agenda. If any of you would like to link up with me over at the show let me know via email.

@Barry and Douglas, Just for fun take a look at this clip Click here to see what level weddings are cut at these days. Aaron Osborne shot and edited this. I'm not an event cutter but I admire how far the wedding crowd is taking creative shooting and editing. The time warping is very good in NLE's these days.

[Barry Green]

Originally Posted by Stephen L. Noe

I'm not an event cutter but I admire how far the wedding crowd is taking creative shooting and editing. The time warping is very good in NLE's these days.

That's exactly what I'm talking about -- when I was at the WEVA expo to check out the cameras, I saw this kind of video everywhere. And I figured the wedding crowd would just eat up the HD100 -- I mean, not only for the look of it, but also if you notice, probably 90% of that video was all slow-motion... so shooting native 480/60p, and playing that back at DVD speeds of 480/24p -- I thought JVC would have a huge hit on their hands.

The available-light question is the only thing. I'm going to go try the split-screen test in SD and DV modes.

[Soroush Shahrokni]

Im sorry to report that we noticed SSE in one of our footage...sad but expected!

[Douglas Spotted Eagle]

Originally Posted by Stephen L. Noe

James Daniels (aka J.D.)

@Barry and Douglas, Just for fun take a look at this clip Click here to see what level weddings are cut at these days. Aaron Osborne shot and edited this.

Exactly what I was referring to. The scary thing is, some of this stuff is done as SDE (same day editing) although the really sweet stuff isn't done quite so fast. This is a great editing job, although the color could have stood some sweetening. These guys know how to shoot for the edit, they are assembling in their minds before they ever turn off the cam.
Impressive industry these days, it's no longer uncle ralphie.

One thing about the 24p and 30p of the cam is it also gets the footage closer to what bride/groom see themselves as, IMO. I think they wanna see themselves in images that are similar to what they're seeing from at least mid-budget films. Not just the cadence, but the treatments.

[Barry Green]

I'm going to go try the split-screen test in SD and DV modes.

SSE is just as evident in DV and SD60P modes... and under really low-light conditions, it can behave quite oddly. The two sides of the image can flash darker/lighter independent of each other, like the processors are trying to find some common ground and can't.

Adding a little light makes that all go away. But there's no doubt, this isn't a low-light camera. It's extremely noisy and split-screeny under dark conditions. Buy a Frezzi or a Pag and bolt it to the top of the camera and never take it off.

[Steve Mullen]

Originally Posted by Barry Green

Adding a little light makes that all go away. But there's no doubt, this isn't a low-light camera. It's extremely noisy and split-screeny under dark conditions. Buy a Frezzi or a Pag and bolt it to the top of the camera and never take it off.

I'm thinking LED LitePanels would be perfect!

If my theory is roughly right and a small IRE Black Level correction factor is computed for, one or both, A/D converters -- and this value is wrong for either, or both, I think it's easy to see why light level is so critical.

Imagine the factor is 5IRE. It's a perfect match for the Left A/D, but high by 3IRE for the Right A/D.

Now imagine a low-light image that ranges from 0 to 25IRE. Suddenly that image ranges from 3 to 28IRE. You can see the difference between it and 0 to 25IRE clearly

But, once the captured image is 0 to 100IRE, the Right image is 3 to 103IRE. Thus, it should be harder to see the difference between it and 0 to 100.

When the difference it really tiny we only may see a line.

And, since we can have an error on any 1, 2, or 3 colors -- its easy to see how the big differences that some see could be produced. And, easy to see why any kind of White Balance error could increase SSE.

Sorry to go on -- but as you raise gain it's clear that color levels do NOT increase proportionately to increasing gain which causes both a loss of color saturation and increased color error.

Moreover, I've noted that the computed color temp changes by 300 degrees when you go from insufficent light (OPEN) to sufficient light (about F4). So clearly low light hursts three ways -- plus the noise that comes from adding gain. How do you compute a valid correction factor for noise???


I once wrote software that computed a heart beat from the differences in body weight that occur on each pump. It took months to fine tune the algorithm to get rid of errors. It's a very hard task!

I suspect that: (a) JVC had no idea their twin A/D would fail so badly. Either a design error or the parts being delivered are not what they requested. You have no idea how often this can occur.

And, (b) in the last few months they have been scrambling to improve the circuit and tune the firmware. The first batch were clearly lemons/betas that didn't come into the USA. Obviously, they can be "fixed."

Chris -- ideally this thread should be split because WE KNOW the PAL units a bad!

I suspect ours -- except for a few that slip through USA QC -- are "fixed" as much as they can be for now. But, I'll bet work continues -- and folks are right to keep the pressure on. Especially for a definition of failure.

This type of problem is not unique to JVC. A Sony V110 Video8 camcorder had an iris hunt that occurred whenever the sky was bright and the ground was dark. Unuseable in India! I sold it.

A Panasonic editing deck, muted the sound at every edit point. When in Japan, I made a series of phone calls and finally got to THE engineer who said -- roughly, "you are correct that it it does this -- it is a 'characteristic' of the design." Then he said, "that means it was our error, but I can't officially say that. I'm sorry it is causing you so much trouble."

[Taylor Wigton]

JVC is segmented globally, and thus every region exists without always knowing what is happening in Japan. That said, be nice to your JVC reps and JVC dealers. Taking a defensive tone, depite your frustration, will only exxagerate the situation. Have some tact and understand the politics involved and I predict that those who purchased the camera will eventually be taken care of if and when the split screen scenerio is sorted out. Everyone who bought the HD-100 has a warranty, so be patient and again, be nice here.

I'm testing an HD-100 right now so I understand why this thread has continued. I think that this thread should continue until there is an official word from JVC Japan, rather then word from JVC subsidearies on various continents.

At the end of the day, remind yourselves that Matsushita Electric is not a 'mom 'n pop shop' based in Albany, NY. If I actually owned the HD-100, I would remind myself that Panasonic and JVC might be working on this problem together, or not, but most likely.

So confirm that your warantee is valid and just deal with the shooting parameters according to what they are NOW, and just think about Matsushita when you go to bed at night. Sweet dreams.


LA based DP

[Mikael Widerberg]

Okay, I wasent so worried about this split screen effect ar the begining.
But today I made some new testshots indoor in my apartment. I have big windows at home so the light was enough to give me a f8 at my withe wall (no direkt sunlight). And the split screen was there.

Seems like this SSE can apear in good lightning condisions as well.

I taped it, and I will give you a link to a sample later.


Seems like it has problems with withe surfaces?
I have a GY-HD100E

[Michael Maier]

That's three now reporting the split even under enough light. It seems light is not the cause as some believed.

[Chris Hurd]

Originally Posted by Steve Mullen

Chris -- ideally this thread should be split because WE KNOW the PAL units a bad!

Steve, you're right, it should probably be split but I wouldn't know *where* to split it.

[Michael Maier]

Originally Posted by Steve Mullen

Chris -- ideally this thread should be split because WE KNOW the PAL units a bad!

It seems to me, judging by the reports, that the NTSC units are just as bad!

[Robert Castiglione]

I'm afraid that there is no question that the split screen effect will appear under normal lighting conditions as well. I can replicate this and similarly have recorded it to tape.

It is just that you dont notice it unless you pick a flat surface with a single colour. Then you see it so clearly. It then is always lurking there undermining your confidence in the camera's performance.

You just so want to love and trust this camera. You vacillate wildly between wanting desperately to trust JVC and just not believing that it could do this to you.

Meanwhile the split screen effect is always there, waiting, just waiting to emerge ...

Rob