Dear Mel,

Please let us know how the pro division responds.

I am certainly looking to swap in the near future unless a solution emerges.

Rob

One of our producers took out our first HD100U over the weekend to shoot some magic hour stuff along the coast. At the time he was unaware of the split screen issue and only noticed the issue upon reviewing tape back at our studio.

The shot that showed the serious split was a sunset shot of the beach, pier and water. The sun was about 20 degrees off the horizon. Long lens shot of pier and sun. I first noticed some serious vertical smear. From the top of the frame to the bottom about the same diameter as the sun, but the scene starts out with no split. Then it appears. The right 1/2 look almost like it has a different gamma. Much lighter blacks. I'll post a still when I get it. This was 0db, 720 24p, ND and stopped down with the stock lens. Not sure about shutter setting. We're having trouble digitizing the 24p stuff with FCP on our Macs.

Just got our 2nd unit today. Still haven't seen it myself inside. It seems like it appears under specific conditions of lighting and or camera warm up. I could see how QA could miss it in a quick test.

Dave
Ft. Myers, FL

We're having trouble digitizing the 24p stuff with FCP on our Macs.


That's because FCP isn't compatible with HDV 24P mode yet. Update is coming.

This has probably been asked and answered elsewhere, but does the split-screen appear in SD?

Well, I guess Ken is really right. Everybody reads what he wants to read.

I guess the guy must be some kind of psychic genius. Who else could have predicted a backlash while making a post which tells users that the defect is a limitation based on how cheap the camera is? What Ken did was a pre-emptive strike, a common practice used to deflect critism that's sure to follow. I've covered enough politicians to recognize spin when I read it:

"I know some nay-sayers will criticise my plan to use the education budget to buy a really big yacht..."

I want the HD100 to work as much as anyone here. But this post, official or not, is the only comment I've heard coming from JVC, and it felt disingenuous to me. It had more disclaimers than substance. If you feel that I've twisted Ken's words, please point out my mistakes. Maybe I could have left out the White House joke, but reading the "not a defect" line reminded me so much of the position Canon took with the XL1 banding issue. Everyone remember when Canon said banding on the XL1 was a 'limitation' of the DV format? It took months of complaining on the the Internet, but a solution was found and the XL1 is a success story. Maybe the HD100 will be home run in the long run, but if I bought a $500 camcorder that had a split screen, I'd send it back. So, I don't take kindly to hearing a rep. saying that this effect is acceptable in a professional camcorder because it only costs $6000.

If we all had the same feelings on the subject, there would be no need for a forum. I personally don't think its ethical to sell a product before its complete, unless you tell customers upfront that the product will have limitations until a fix is found. If you and Nate disagree with my first post, please tell everyone what part was wrong. I don't think of debates as minefields.

Steve

Yes, the same problem of split screen in SD as well.

Also, although I very much appreciated having contact from Ken (particularly the bit about the fact that a solution is probably emerging) and the general intent of his email , I am bound to confess that it was not helpful to be told that an obvious and irritating defect apparently in the design of the camera is something to be accepted as a "limitation" in its performance. That wont stand up to scrutiny.

Cant shoot in low light with any confidence, cant shoot magic hour shots? Would you start a shoot with the camera in its current condition?

I await the fix for this problem as soon as possible.

Rob

JVC have to fixe this, no doubt about that.

I would not be worried at all, you can just bring it in to the rep I they will fix it for you.

For me the case is a little different. I bought my cam from Globalmediapro (they are wery good). Living in Sweden I have to pay for the transport to the other side of the world (270 USD).

I am also in the midle of a project, and I have to kepit untill I am done. Lyckely most of the shots are outdoor in good light.

I have also E-mailed JVC-pro, but didnt get any reply.

I am hoping a firmwher can fix this splitscreen effect

Mikael.... you are a brave man to undertake a project with new gear on the fly.

Let us know how it turns out.

Here is a link to the shots from the beach. Stills. The first show the split and as the sun set it went away.

http://www.dreamtimeentertainment.com/hdv/HD%20test%206-7.jpg

http://www.dreamtimeentertainment.com/hdv/HD%20test%207-1.jpg

Dave Beaty

Was it taken at 0db? I'm sure it was, but I just wanted to make sure.

Here is a link to the shots from the beach. Stills. The first show the split and as the sun set it went away.
Holly s%@t! That's even worse than I though...

Our producer Matt, reported that his shot on the beach was with stock settings. It was his first shots with the camera. He told me there was no gain. Looks kind of grainy.

It's odd that the split seemed to appear and then just as quickly disappears as the scene evolved. Again, a tough condition for the electronics to address with the full sun and dark forground. Yet the clear image looks good.

Speculating here, but it seems like the right half is in a seperate memory buffer that is being processed ahead or behind the left half. The camera is trying to apply seperate processing to each half. A challenge almost like trying to match two cameras perfectly.

Dave B

here are all the still image captures....

www.dreamtimeentertainment.com/hdv/

Looks like the black stretch is turned on for the right half of the screen, and off for the left.

Dave,

I imported your shot into FCP5 on V1. I then copy and pasted it into V2 and LEFT CROPPED it 50% with an edge feather of 1.
I then added the 3-way colour corrector to the cropped right-side (V2) and quickly created a simple adjustment to make the two sides match.

Here's your original:

http://homepage.mac.com/timdashwood/.Pictures/JVC-split-screen/HD_test_6-7before.jpg


Here's the corrected version:

http://homepage.mac.com/timdashwood/.Pictures/JVC-split-screen/HD_test_6-7after.jpg


And here's the 3-way CC settings:

http://homepage.mac.com/timdashwood/.Pictures/JVC-split-screen/split-screen-fix.jpg


See how this works on the moving footage.

Tim

Tim,

Well, that fix is pretty close. At least it's not a total loss.

We'll be sending in our camera to our dealer for, hopefully, a swap. It still bothers me the issue was so hard to produce. I tried to duplicate the split yesterday in the studio with no luck. I had the camera on DC for several hours and tried various lighting conditions. I never saw the severe splitting like in the beach. But it's there somewhere. Also, any comments on the vertical smear. I know it's a $5500 cam, but should it be so bad?

Dave Beaty

I just wanted to re-iterate a thanks to Ken Freed from JVC--he's a VERY nice guy and I've met him personally a couple of times. He doesn't B.S. anyone and he knows his stuff. The fact that he isn't here to sell but rather help is a testament to his commitment with JVC.

Thanks Ken! Next time we're at a camera show, I'll buy you a drink!

Heath McKnight

Hi
I havn't been here in awhile. I bought a HD10U from JVC and i'm going to upgrade to a HD100 in the next year. I read alot at times on this thread and I was looking at the Pic's of the split screen shots. Could the problem be in the lens instead of the HD100. Has anyone tryed it on a different lens

Would like to thank everone for there info. It makes for some good reading


LHR

I really don't think it's in the lens. Rumors say it's because 2 different processors handle the left and right side of the image.

The problem is absolutely not in the lens; there was one shot we did with the mini35 where the splitscreen effect was visible.

The split screen phenomenon is indeed a function of two separate processing functions; this is confirmed by JVC.

The problem is absolutely not in the lens; there was one shot we did with the mini35 where the splitscreen effect was visible.

Was it a shot which didn't make to the final samples?
By the way, there was no split screen in that low light night shot, where the camera dollies in on a passing by woman. Was it corrected in post? Since it was a low light shot, and the split is most visable under those conditions, I wonder if it was fixed.

Michael:

It was indeed a shot that I didn't post. The split screen phenomenon has been documented elsewhere so I didn't feel it necessary to include the shot in a review that was oriented specifically to the use of the Mini35 and the HD100. The clips as seen in the article were direct from the camera with no correction or processing.

My best guess of why the one shot exhibited the split screen phenomenon so visibly is that it had lower contrast and more overall black in the frame, as opposed to the posted shot which had a fair amount of highlights present to help the contrast.

My best guess of why the one shot exhibited the split screen phenomenon so visibly is that it had lower contrast and more overall black in the frame, as opposed to the posted shot which had a fair amount of highlights present to help the contrast.

How long had the camera been on (up and running) before you saw the split screen? Late in the day? At first turn on? Was the camera cold? hot? anything environmentally that you could point your finger to?

I shot some footage in a very dark reception hall this weekend - just test shots to see what would happen.

I only notice the 'split' when the camera isn't getting enough light (it's like it's fighting to process what is/isn't coming in) - Then I did some tests using it at 1/30, with 6 or 9db gain, and I maxed the gamma and used a color gain of 5. I didn't stretch the blacks much - but that would help in a pinch. This was supplemented by an on cam varilight that had a small chimera box on the front of it.

It was more light than I'm used to using in SD, but it wasn't completely obnoxious. Once I got enough light in there (via light & various gain/shutter), I saw no splits. The color in the viewfinder seems flat, but I was able to see the split if I was looking for it, and correct the situation.

The more light the better, but I feel more comfortable knowing you can still get some footage in tougher venues - just make sure you've got some light. Looking forward to their update - I'm sure it'll be cool.

Could you post some screen grabs with the slpit and without (after you tweaked the camra added light)?

Hello Everyone,
I was shooting test clips today using 1/4 ND, no gain, f4-8, shutter 1/60...I had the dreaded split screen show up in some of the test clips. I had ample light..Is it just me? Has anyone had the same problem?

Okay Guys,
JVC's Pro Division referred me to a 1-800 number for service on the split screen & verticle mid screen red line, got alot of automated messages...very frustrating. I finally gave in and delivered the camera myself to JVC's repair center in Pine Brook, NJ. Strangely, the employee said that although he was told to expect "a batch" of defective cameras, I was the first one to send it in (to their location ??) He said he'd have the techs look at it with an est. wait of 7-10 days.. I'd be very surprised if I get the same unit back....we'll see.

Hmm, interesting. Defective...

heath

Mel, we're all very interested, so please keep us advised on how this goes. Thanks,

I'm glad that folks now are confirming that the amount of light seems critical in this issue.

In my examples, I noted that on a low-light pix (say 0 to 30IRE) verses an adequately lit (0 to 100IRE) pix suffer very different fates if, for example, a tiny 3IRE Black Level error between the L and R sidees is present.

In first case, it is a 10% error -- in the second, it is only a 1% error.

What can create the such a tiny Black Level error?

1) Not correctly white balancing. If you simply aim the camera at a white wall you may indeed see the split. But upon Manual WB, it should disssappear. And, WB is clearly inaccurate at low light levels which compounds the issue.

2) Gross levels of smear on one side. This is very clear in the sunset shot. No one would use a shot with that much smear. It never should have been published as an example of anything but gross over-exposure.

3) Too much noise -- which is not simply a matter of absolute gain. Some tones more clearly suffer from noise. But, the rule of +12dB, or less, works for me. I can evoke the same amount of SSE at 0dB as at +18dB. I don't think one MUST avoid high gain.

4) Defective units. I keep wondering HOW bad units are getting past QC in Japan? If things can go wrong at the factory -- they can go wrong at the incoming inspection. So it's possible that some percentage of the units will always be defective. (A vertical red line sure sounds like a very defective unit.) Given this, what really counts is that a company will -- without a hassle -- make things right. And, while one can wish things were different, if JVC steps-up the way Canon did, the negative PR will go away as bad units are quickly replaced.

2) Gross levels of smear on one side. This is very clear in the sunset shot. No one would use a shot with that much smear. It never should have been published as an example of anything but gross over-exposure.
You are not talking about this, are you?
http://www.time-code-media.de/split.jpg
Which sunset shot is it?

I don't see that as a sunset. It looks like the SSE you get when you have not correctly color balanced.

There is a series of shots of a setting sun over a pier where the sun is a ball of super bright light with a mear the width of the sun going downward.

So Mel, you have a NTSC model and have the split screen even when there's enough light? F4-f8 sure seems like enough light.

For those experiencing SSE you should definitely try using only manual white balance (as recommended above) - I have been doing so and it looks very promising at the moment for sharply reducing incidence of the problem. I will keep testing.

Rob

There is a series of shots of a setting sun over a pier where the sun is a ball of super bright light with a mear the width of the sun going downward.
I remember those, they didn't seem overexposed at all to me. Yes, the sun itself was of course hot but a sunset is a pretty common shot, especially for nature documentaries and we need the flexibility to be able to shoot such shots.

So Mel, you have a NTSC model and have the split screen even when there's enough light? F4-f8 sure seems like enough light.
Yep, I believe I had adequate lighting..What concerned me more was the faint but consistently present mid screen verticle red band. Once I keyed on to it, I saw it everywhere in varying degrees...different shots/settings & tape. Is this a free sefx from JVC?...There's no button on the camera for it. I prefer adding all effects on my NLE. I returned the camera in hopes that the company will remove the "special" feature.

I remember those, they didn't seem overexposed at all to me. Yes, the sun itself was of course hot but a sunset is a pretty common shot, especially for nature documentaries and we need the flexibility to be able to shoot such shots.

When have you seen a pro shoot a sunset with vertical smear? Never!

Any time you get smear that wide from the sun -- the shot is over exposed by definition.

To do a sunset correctly one must bias the exposure for the sun which means the beach would be much darker. This guy likely did an AVERAGE exposure.

In any case, the shot could never have been used because of the smear -- the SSE was irrelevant.

I remember the shot. The white balance was off as well.

As many of us were (or still are) former JVC HD10 users - I've read this entire thread and thought about the times we all talked about new firmware for the HD10. We really wanted it and asked for it, but never got it. Granted, the "limitations" weren't hidden when we bought the camera.

It's strange, but my HD10 had that horrible streak too...except it was green. It showed up when shooting anything bright like chrome or the sun.

I don't think it's to much to ask for them to stop by here more than once a year and talk about their defective or limited products.

The only people who deserve beers are the people trying to make a living with defective and limited gear!!!!!!!!!

It's strange, but my HD10 had that horrible streak too...except it was green. It showed up when shooting anything bright like chrome or the sun.!

This streak is called "vertical smear" and is present on EVERY "IT" CCD camcorder. It is inherent it IT CCDs which are used in almost every camcorder.

Naturally, the higher the pixel count and the smaller the CCD -- the greater probability of smear. It can be any color.

Over time, CCD smear rating is getting better and in some more expensive camcorders is now about -128dB.

To imply it has anything to do with JVC is, frankly, to reveal a lack of experience.

This streak is called "vertical smear" and is present on EVERY "IT" CCD camcorder. It is inherent it IT CCDs which are used in almost every camcorder.
Naturally, the higher the pixel count and the smaller the CCD -- the greater probability of smear. It can be any color

The difference between 1/3" and 1/2" must be like night and day when it comes to smear then. Because I have shoot with Sony DSR300 and Panasonic DVC200 pointed directly to car head lights and stadium spots and never got smear.

Or perhaps you exposed the footage somewhat less...

Hey Werner, if you are referring to my post, I don't think I could have underexposed at all, since I wanted detail in the background, and if I had underexposed, the background would have gone dark.

When have you seen a pro shoot a sunset with vertical smear? Never!

Any time you get smear that wide from the sun -- the shot is over exposed by definition.

To do a sunset correctly one must bias the exposure for the sun which means the beach would be much darker. This guy likely did an AVERAGE exposure.

In any case, the shot could never have been used because of the smear -- the SSE was irrelevant.
It seems that you never shot a sunset, Steve. The sun will be ALWAYS overexposed. No matter what camera, what size chip, to expose for the sun, you would get nothing in the foreground, all pitch black and the sun would be still overexposed. Vertical smear is an unfortunate characteristic of CCD sensors and sometimes cannot be avoided, especially with smaller chip cameras. When I shot sunsets with my DSR500, the sun itself was overexposed but there was no smear and that's thanks to the bigger size chip. With the HD100 there will likely be some smear in sunset shots, as well as, in night-time shots with car headlights but there should not be SSE.

I am pretty tired and don't have time for your on-going defence of this particular issue and I am just gonna leave the battle to someone else.

However, I want to mention one more thing, the suggestion of white balancing in low light. One technique used by electronic cinematography is NOT to white balance for every setup and to use a pre-set WB instead. The reason is that the process of white balancing removes hues that are dominant in the shot (i.e. WB with blue gel minimizes blues and 'warms up' the image). If one manually WB in the evening (sunset) the camera electronics will attempt to 'compensate' and remove the warm hues that make a sunset what it is, giving the image a non-descript flatness. Also, in order to create a colour continuity it is important not to change WB between setups, among other settings (of course that doesn't apply when switching between daylight and incandescent lighting). So, even rental houses set up high end cameras (like CineAlta and others) to the liking of the DP/client and then the settings are not typically changed and certainly one doesn't manually white balance.

"The difference between 1/3" and 1/2" must be like night and day when it comes to smear then. Because I have shoot with Sony DSR300 and Panasonic DVC200 pointed directly to car head lights and stadium spots and never got smear."

One thing to remember is the smear is related to the size of the pixels right?
THese cameras have many more pixels in the same size CCD so when it comes to things like smear and sensitivity they will perform like a much smaller SD block. So a 1/3" HD CCD will proabably have pixels about the size of a 1/6" or smaller SD CCD. Your probably going to need a 2/3" HD block to get the performance of a 1/3" SD block with 1280x720.

Still, I havn't heard of the Sony 1/3" block having such bad smear and it has about the same size pixels.

Anyone know who actually makes the CCD block in the JVC?

Guy, I haven't had any smearing with the Z1U. I haven't shot into the sun yet, but I will try this weekend.

FWIW- The XL-H1 smears in the sun too. If you download the attempted CA shot over on the H1 community you can look at a nice smear from the sun all the way to the bottom of the frame.

We're getting way off topic here, but one shouldn't shoot into the sun to begin with, because of all the reasons we've discussed and more.

heath

"Your probably going to need a 2/3" HD block to get the performance of a 1/3" SD block with 1280x720.

Still, I havn't heard of the Sony 1/3" block having such bad smear and it has about the same size pixels.

Anyone know who actually makes the CCD block in the JVC?

Sony's Hyper HAD CCDs have really, really low vertical smear. Nevertheless, trying to shoot a direct sun and keep the forground visable is exceeding the exposure latitude one would expect from a 1/3-inch CCD. It's simply not good practice as was bourne out by the results.

Guy's right -- for those tough situations there's a reason pros buy camcorders with 1/2-inch and 2/3-inch CCDs!

And, it is rumored that Sony makes the CCDs.

I've heard that Sony has made the Canon XL H1's chips.

heath