HD100U Split Screen Will Not Be Fixed?

[Brian Duke]

Originally Posted by Soroush Shahrokni

Ditto. Everytime I refresh this site in hope of finding something intresting to read I find a new post in one of the SS topics. I have been shooting with this cam ever since it arrived a week ago and have only come across SS once and that was when I was shooting in 0 light and a M2 adapter. Im so sick of this SSE topics!

Personally I dont think any camera in the market can compare with the JVC...not even the new Canon. Perhaps the HVX will but that remains to be seen...I think JVC has done a tremendous job in its compression!

I have never owned a camera as good as this one and Im basicly sleeping with it. I dont regret for a second the choice that I made getting this camera!

Hopefully these SSE threads will be replaced with intresting ones á la Tim Dashwood!

I agree. And I am really looking forward to recieving mine this week=)

[David Dessel]

This "Community" seems to be filled with an overwhelming amount of negativity regarding this camera, and much of it by folks who don't seem to own the unit.

How many of you have seen the image it produces on a high def monitor?

The JVC is really exciting. The visuals it produces are quite amazing given its price point. Isn't more productive to focus on how to get the very best out this camera?

I'm not learing anything new from most of the posts on this forum, so why stick around and read to all the whining?

If you don't want to buy it, fine. Understood. No one is forcing you to. Why do we need to know over and over again? Are you hoping JVC will read your post? Why not write them a letter?

For those of us who own the camera, I hope that we can share what we learn in a more positive and productive context. This way we can create really great work with this amazing tool while others spend their time complaining about a camera they will never buy.

I would also like to thank those who have shared information that is actually useful for owners who have made the committment to the JVC. It is truly appreciated.

-Dave Dessel

[Chris Hurd]

Hi David,

I said pretty much the same thing in our "other" Big Long Thread about SSE, basically stating that it's a matter of policy that we don't allow anyone to talk trash about the gear they're *not* using. Feel free to look up my posts on that topic. Meanwhile, if you can bring yourself to ignore the noise in the SSE threads, I think you'll find plenty of very positive discussions about all other aspects of this camera from actual, bonafide owners. Either way, I am watching things very closely in here. Thanks for your input,

[Tim Dashwood]

I'm just glad someone added a "?" to the title of this thread.
It has obviously stirred up alot of reaction, but Brian didn't even say who the "JVC representative" was, what department she was from (Pro sales, engineering, maybe a dealer...), or if she represented Victor of Japan, U.S, or was regional. I personally had a "no questions asked" exchange of two units on Sept 7 with the GM of JVC Pro Sales in Canada, and I felt I had been given the "The Perfect Experience" when I left with two brand new cameras.

It is too bad JVC doesn't have someone like Jan from Panasonic to talk straight to us. However, it doesn't make any sense to quote an anonymous JVC representative in the U.S. and state "HD100U Split Screen Will Not Be Fixed" when dvinfo and dvxuser members have reported at least THREE separate firmware updates in the past four weeks in North America release models, and UK customers are reporting that the latest firmware seems to have licked the problem completely. This could be due to operator error (not knowing how to properly stimulate it) but time will tell.

I have been thoroughly testing the SSE in my v1.14 firmware HD100 in a controlled environment and have discovered some very interesting things (that I will fully report on when I have completely analyzed my results.)
My first unit was v1.12 or older (if memory serves) and the SSE was visible in slightly underexposed situations: totally unacceptable.

My replacement camera body (v1.14) showed no signs of SSE when I picked it up and tested at JVC Canada. However, four weeks ago we all thought it was a result of gain JVC was using +9dB as the tolerance of rejection for sales. The other problem was that the "test" at the time involved closing the iris and cranking the gain until a split was visible.
However gain is only one factor. Others have suggested that you need a stop of at least ƒ2, and a manual white balance will solve the problem. That's not exactly right either. There is no easy solution, but I have been concentrating on finding the cause.

All that matters is the output level of the chip, no matter how you choose to light, expose, WB or otherwise capture your image. If you're base signal level hovers around 7.5IRE, there may be a problem.

In my many, many hours of observations this past week inducing the SSE (with v.1.14 fw) using controlled light on an out-of-focus gray card & measuring with waveform monitor, I have found that when MASTER BLACK is set to normal, the split presents itself in a very tight range of 5 and 10 IRE. In fact, it seems to be most prevalent when light levels are averaged on the 7.5IRE dashed line.
I remembered a couple curious observations from last week's gamma curve mapping that the DVX100 added the 7.5IRE setup level to the firewire signal that could be digitized, but the HD100 did not. Also, the DVX100 seemed to be capable of illegal superblack, but the HD100 clipped at 0 no matter what. I also got thinking about the fact that 7.5IRE is a North American thing and that NTSC in Japan doesn not use 7.5 IRE setup, they use 0.

So, I know I'm ranting on, but... long story short... I tried the setup at 7.5 and 0 (only affects analog out,) adjusted MASTER BLACK to every conceivable setting and retested and found the magic number of MASTER BLACK -3 removes SSE. This brings the curve down about 2/3rds of a stop, but can be easily compensated for with black stretch3.

I now cannot stimulate a split screen at gain levels 0, +3 or +6 under any conditions (on my v1.14 Firmware.) My wide latitude scene file (which used MASTER BLACK -1 anyway) still works well with Master Black -3, and it reduces the amount of "banding" compression anyway.
I will once again repeat that these results were with a version 1.14 firmware HD100. Those with v.1.12 should probably just plain return their unit, and those with 1.15,1.16 or 1.17 should try other Master Black settings and see what happens.

[Tim Dashwood]

Now, back to the old theories, and why they did hold partial truths.

circa July/August: Barry, Nate & Charles test the HD100 with the mini35 and observe the split in a couple of shots. However at the time it is chocked up to a "pre-release unit."

August 20-22: I receive two brand new HD100s and do not notice the split because I was shooting some bright outside footage, and frankly wasn't looking for it. Barry asked me if I noticed it, so I turned my camera on at night in my office and saw it on my studio monitor but not on the digitized footage. Barry prodded me to look closer at my digitized footage and after boosting the mids and blacks... there it was.

August 23-30: My dealer gets on the case with JVC and talks directly to the head Canadian engineer and another on loan from Japan sent to QC every camera for release. I find out that the first batch hadn't all been QC'd before delivery, so I may have a couple of lemons.
I was also told that it was a calibration problem to do with gain and that if any camera presented SSE at or below +15dB it would be replaced. The basic test was to close the iris, gain up until it was visible, and then note the gain value. This test made sense at the time, but now that I have discovered that SSE doesn't happen below 5IRE (black=black on both sides) it makes sense that the gain grain was just raising the level high enough to trigger the SSE.
I arranged a day to exchange my cameras.

Sept: The reports start coming in from around the world that everyone seems to be discovering the phenomenon. We all start doing the "lens cap on test" to see what level it appears at. Mine at the time were triggering at +6dB.

Sept. 7: I spend two hours at JVC Canada head office exhanging my two cameras, chatting with the QC engineers and the GM of Pro Sales, and testing the new ones before I leave. I found a dead pixel on one, but the very nice QC engineer from Japan graciously showed me how to clone it.
Both new units only showed SSE at +18dB with the "lens cap on test" so I was a happy camper.

Sept. 14: I shoot an impromptu night side-by-side with a DVX100 to test low light sensitivity, grain in gain, streaking, and MAX gamma at 0dB.
The test was very promising, but I noticed a couple moments of SSE just as headlights flared the lens from off-screen. This makes sense to me now because I had rich BG blacks hanging around 0-5IRE and the flare brightened the whole image up just enough into 7.5IRE to trigger SSE.

Sept. 16: I realize that I can fix the split in post by slightly adjusting the blacks (0-50%) slightly. This once again suggests that it is a "setup level" problem and not a gamma level problem.

Sept. 19: I shoot a more practical nighttime camera test/screen test with an actor and a director friend. It is all available street light in an industrial "ghost town" of sorts. I have now tweaked my "low-light" scene file and do not notice the SSE in the field at all. My exposure is consistent, our "key" light is only reading 2 to 5 footcandles in the hotspot (ambient light doesn't even move the needle) and the camera performed exceptionally well. I would have rated that curve at 800ASA.
However, upon review of the footage SSE does show up from time to time... and not on a shot for shot basis. It "flickers" on and off within a shot depending on tilt and pan. But why? The BG is the unchanging night sky, iris is full open, nothing automatic is turned on, and I was on 0 gain!
Once again the culprit seemed to be lens flare raising the overall level of the blacks into the 7.5IRE range. This would be similar to "flashing" film negative after it has been exposed to bring up the blacks.

Sept 23: Steve Mullen gets a camera in his hands and explains the technical magic involved in JVC two-chip method of image capture and processing. The calibration issues all make sense to me, but it worries me that the problem is intermittent and flickers in and out - suggesting more than just a stable calibration problem.

Sept 26: Steve ponders that he saw the split on the LCD monitor, but not on a properly calibrated monitor. This fits with my 7.5 observations as a HDTV monitor would use 0 black.

Sept. 29: Steve coins the phrase "SSE" and I hope it doesn't catch on. I think "Special Skills Extra" everytime I read it!
Anyway, he suggests that white balance is tied to the problem, and you must set your gain and then manually set WB. I'm not too sure about that one considering 95% of my nighttime shoot was OK, even though I used preset 3200K with various odd colour temp street lights.
However Steve does concede that the split screen problem is not only related to high gain levels. Correct... but this is about where all hell broke loose on the forums. In addition, Steve suggested it is caused by insufficient light, and that you will "need at least a reading of F2 - AVERAGE," and that if there is a BG in frame that doesn't have at least F2, then we are just "SOOL" (paraphrased by me;) That's just plain nutty! What would happen if you hired a DP like Darius Khondji and told him that you really liked the way he shot "Se7en" but the producers want to shoot with the HD100 and we can't go below an average exposure of F2 on this movie?
Anyway, this is about when all of the "the HD100 isn't for me" comments started with a fury and haven't stopped.
I knew that this "need more light" theory didn't make any sense because I had already disproven it two weeks earlier. I know Steve was doing his best to come up with an easy to use "rule-of-thumb" for ENG shooters and wedding video guys, but IMHO it just doesn't hold any water.

In conclusion:
A camera is one of the cinematographers tools, and we need all of our tools to work as expected at any time. I believe that these past 8 weeks have been tough for new HD100 owners. We have felt a little left out in the cold, but if the new reports out the UK are true, it may seem that the most recent firmware update has finally worked the bugs out.
Even though I would like to know what it happening behind the scenes, I also think that it actually makes corporate sense for JVC not to acknowledge the problem publicly or do a recall.
Didn't anyone ever see "Fight Club?" "Take the number of vehicles in the field 'A,' multiply it by the probable rate of failure 'B,' then multiply the result by the average out-of-court settlement, 'C.' AxBxC=X. If 'X' is less than the cost of a recall, we don't do one." This is how corporations think... at least it isn't a life or death decision with JVC.
If they just wait it out, work their asses off to fix the problem by the end of October and there is no bad press, just a little negative word-of-mouth from some members of the newsgroups, this camera will sell well and the whole thing will be forgotten.

In the meantime my solution of setting the master black to -3 (on v1.14 FW) works well for me, and isn't undesirable, but I still want to make sure I get the firmware updated so that I don't have to ever think about or discuss it again.

'nuff said. Tim

[Steve Mullen]

[QUOTE=Tim Dashwood]Steve suggested it is caused by insufficient light, and that you will "need at least a reading of F2 - AVERAGE."

Actually, I later raised this to F4 to compensate for the possibility that you might have a statistically "non-normal (non bell shaped curve) where a few bright lights caused you to set the exposure higher than it should be for the larger dark areas. This was also a simply way to keep the F-stop at the setting where the lens offers the most rez as measured by you and/or Berry. (There's nothing wrong with F2 or F2.8 in "normal" situations, but you are scarificing resolution.)

Now my logic may be wrong, but the results I get speak for themselves. No SSE and max rez.

I'll admit I did not think about those who are not into reality shooting. And, at that time most of the discussions were about shooting reality -- weddings and docs.

You feel my "rule" would prevent creative -- dark mood -- shots. I don't think is true.

One always must set exposure with regards to either the brighter (reversal film) or darker (negative film) areas based upon the film's latitude (reversal film) or sensitivity (negative film). If one doesn't, then either the highlights burn-out or the dark areas on the negative wind-up being clear. In either case detail is lost. No matter the desired mood -- it's very rare that a filmmaker wants his/her source film to REALLY have no detail as there is no way to get back details once lost!

What they want is to CREATE the LOOK of dark. Which brings up the old confusion of film students -- sometimes they hear someone use "F-stop" to mean an amount of light (as used by a director), other times "F-stop" means a setting on a camera. I think you have confused the two meanings. I can dig out a quote from one of my Cinematography books that explains this much more clearly than I can.

Once one lights for the bright or dark area(s) in a way that keeps detail (or with the HD100, prevents SSE) -- one then lights the other area(s) to keep the light-ratios of all parts of the scene at what one wants to create the mood. There are whole chapters on how -- once you choose a negative film and know the senstivity of your film PLUS the latitude of your film, you use a light meter to compute these ratios and set a CAMERA F-stop. THEREFORE, WHEN ONE IS LIGHTING FOR MOOD, ONE WOULD OBVIOUSLY NOT FOLLOW MY "F4" RECOMMENDATION.

Nevertheless, by treating the camera AS IF it has a lower sensitivity than it measures (by video testing) -- light levels will be high enough to prevent SSE while the ratios will still establish the mood you want. (Given, of course, you keep total scene contrast within the camera's latitude.)

-----------------------------------------------

"In the meantime my solution of setting the master black to -3 (on v1.14 FW) works well for me, and isn't undesirable, but I still want to make sure I get the firmware updated so that I don't have to ever think about or discuss it again."

1. I don't think that even with a firmware update you will be able to forget that SSE is a possibility if you shoot without sufficient light. Unless there is an even newer NTSC firmware than I have!

2. What I "sense" you are doing by dropping MB is forcing the darker areas to become fully black. Since at first, it seemed it was the dark, but NOT black, areas that got SSE this idea works. Until you realize that you can get SSE on a white wall if the lens reads OPEN. Clearly, lowering MB by a 3, will not bring that white wall into black.

This raises three possibilities:

1) We are both seeing "black level" as somehow very related to SEE. Perhaps we are close to a solution, but not yet there. Alternately, for example, perhaps the JVC camera unlike Panasonics doesn't do a true BB with the WB but instead, as does Sony, uses the pixels outside the CCD image area to measure black level. This may not be fully adequate under all conditions and so will limit HOW the camera is used.

2) We both have stumbled upon "solutions" but they are not universal. In which case, we have tried but failed.

3) Ideally, JVC will allow their engineers to publish a White Paper on SSE and HOW TO MINIMIZE IT. As long as the internet is buzzing with folks like us trying to solve THEIR problem without a deep understanding of the nature of the problem -- SSE will stay a thorn in their side!

Alternately, every product has "recommendations" so there is no shame in making them clear.

Here's an example: "The camera will output 100IRE at no gain, for a sensitity of X lux. However, for maximum picture quality with the fewest image artifacts, use no more than +YdB gain and at least Z footcandles of light on a 80% gray card."

The gain value would be selected by looking at: chroma saturation that remains proportional to luma level, luma noise, plus AM and PM chroma noise. There must be a point where all these values go to hell. JVC could simply choose the gain under that point.

[Michael Maier]

Originally Posted by Chris Hurd

At the risk of bounding dangerously off-topic, just wanted to point out that you really don't have to use the clutch to change gears in a manual transmission. All you need is a working tachometer, and to know at which r.p.m. the shift points are.

Actually Chris, unless you really don't care about your car, you will want to step in the clutch to change gears, or you will destroy your gearbox. :)

That's in a new car.

[Guy Barwood]

"Maybe that's why I find this "I can't shoot with limitations" talk so whimpy."

So now Steve has resorted to calling everyone who doesn't agree with him as 'whimpy'.

What were you saying about personal insults Chris?

"This "Community" seems to be filled with an overwhelming amount of negativity regarding this camera"
I disagree, it is much more directed at JVC themselves or only a single but significant to many of us issue, that in theory, has the potential to be fixed. We are angry at JVC because they have brought us so close to what we want then stuck a thorn in our side without warning. Its not like its a design feature such as a limited CCD resolution which everyone would just accept (such as Sony's half res CCD).

Also, people don't generally discuss success stories (sure you get a few) so you are much more likely to see problems and complaints in forums, its in out nature...

"and much of it by folks who don't seem to own the unit."
but not people who don't want to buy the camera if it wasn't for a particular issue, there is a big difference. If this issue didn't exist most here talking without the camera would probably be much lower.

We just want this one issue fixed so we can get on with buying one.

How many cameras are JVC replacing? that must be costing them a small fortune?

[Brian Duke]

Originally Posted by Tim Dashwood

In the meantime my solution of setting the master black to -3 (on v1.14 FW) works well for me, and isn't undesirable, but I still want to make sure I get the firmware updated so that I don't have to ever think about or discuss it again.

Hey Tim,

Thanks for your thoughts and comments. I am gettting my camera this week. I assume the "Master black" is a setting in the menu, correct? As a novice I am not completly clear on your settings. Can you explain it in laymen terms for me and others who are going to be using it. Maybe a step by step guide to "safe settings" to avoid SSE. What is v.1.14 FW? The lens you use?

Thanks

Duke

[John Mitchell]

Originally Posted by Guy Barwood

"I know you assumed that the chip was full 1920 x 1080 - no such luck amigo"
Actually I know the CCD specs very well, which is why I detailed Sony having only 12% extra pixels, not over 100% extra (as 1920x1080 would have calculated)

"And the irony here is that Sony make the GY-HD100 CCD"
This is the first time anyone has indicated who actually makes the CCD. You know this for a fact or just guessing? I'm just curious thats all. I think it is not the CCD iteself that is the problem, it is potentially the circuits just after the CCD, so it could be the CCD itself is just fine.

"I think it's patently unfair to accuse them of being technically incompetent"
I didn't, but I did say they weren't up to this particular task. If being innovative produces cameras with faults like this then perhaps being Innovative isn't such a great thing, and perhaps Sony are onto something.

"to date there has been no official announcement from JVC on SSE"
This is a major concern to many in here, its like publically they completely deny the issue (but in private will discuss it),and that doesn't provide any confidence in them from an end users stand point. I guess it would be worse if they came out publically and denied any problems...

1. G'day Guy - you're still missing the point - it's the horizontal that is the critical factor not the vertical (as Sony scan interlaced they only have to scan half the vertical resolution at a time, but even that is irrelevant) and the facts are that the Sony chip is not as wide as the JVC by a factor of 25%. I repeat it is not simply a problem of the number of pixels - sorry I didn't do the math on the 12%.

2. I do know for a fact that Sony make the JVC CCD, because JVC told me and I'm certainly not he first on the forums to mention that fact ( it may have been the HVX forum... I can't recall). Hey - maybe JVC lied to me for some obscure reason. You're right it's not specifically the CCD, but the technical dilemma posed by the CCD.

3. Was that you who made the technical comment - I'm losing track? If you read what I wrote I pointed out that Sony DID have an enviable field reputation precisely because they don't like to innovate too much - so you are agreeing with me, right? Mark my words - this is an innovative piece of kit, but you are perfectly entitled to wait until JVC sort out their problems. Even then if I were you I wouldn't buy it, because (as I've already pointed out to you on another thread) this is not the right camera for low available light shooting - they are significantly slower than your current camera (2 - 3 stops). You'll probably need to wait for the next generation of one third inch HD or spend the money and buy a larger chipped camera, to get anything like the same performance.

4. No official announcement - once again you're just agreeing with me. Personally I think it's a public relations nightmare for JVC not to have come out an made some kind of official statement, but I was merely pointing out that someone started a thread that says "Split screen will not be fixed" based on what they were told by one person from JVC, not an official announcement. Are you disagreeing with that point Guy? Or merely stating the obvious - of course everyone on this forum would like to see an official statement from JVC, myself included. And BTW, I believe Brian is buying one.

Guy, bottom line is I have dealt with both Sony and JVC here in Australia on a professional level (and I guess you have too) and I couldn't be happier with JVC's response locally - they have been totally proactive (unlike Sony), replacing my unit with a brand new one when it developed a fault and the head service engineer personally testing a bunch of cameras until he found one that passed muster. And I repeat they are individually testing and rejecting units that don't come up to scratch when they come into Australia.

[Chris Hurd]

Originally Posted by Michael Maier

Actually Chris, unless you really don't care about your car, you will want to step in the clutch to change gears, or you will destroy your gearbox. :) That's in a new car.

That is true, Michael, you are right, and I would never recommend that practice. But it is possible. Damaging, yes, but possible nonetheless. Sorry for the off-topic response -- many thanks to Tim Dashwood and John Mitchell for their excellent contributions to this thread!

[Guy Barwood]

1: I'm not missing the point at all, I took your point and corrected your assumption about what you assumed I thought.

2: I was only asking out of curiosity. I havn't seen it detailed in this forum before and I have previously asked if anyone knew (which went unanswered)

3: I know the problems of sensitivity, and perhaps this problem with this camera will help me in the long term by making me wait longer, possibly long enough for a more sensitive camera. Unfortunately short of manufacturers releaseing larger chips at lower prices in the next 12-24months, I can't see sensitivity improving a great deal very quickly. Even DV cameras sensitivity improvements has slowed quite a bit lately. It will continue but I think it is something we will have to live with for some time yet. Hence, why wait when I really don't think it will improve, might as well get to know how to work with it now. It's really no worse than a Z1 (better in some cases) and others are already surviving with those in weddings. Luckily I can wait, but don't want to.

4: Probably stating the obvious, but what is obvious to me doesn't always seem obvious to everyone else, so sometimes it doesn't hurt to 'state the obvious'.

I am interested in your experiences with JVC service in AU. I am still waiting for a firmware upgrade to my DV500 to stop it pulsing the iris with photographic flash's I was promised 2 years ago...

[Chris Hurd]

Originally Posted by Guy Barwood

Unfortunately short of manufacturers releaseing larger chips at lower prices in the next 12-24months...

Short answer: there will always be a significant jump in price between 1/3" and 1/2" (or larger) cameras.

Originally Posted by Guy Barwood

Hence, why wait when I really don't think it will improve, might as well get to know how to work with it now.

If you're waiting, you're not creating.

Originally Posted by Guy Barwood

It's really no worse than a Z1 (better in some cases) and others are already surviving with those in weddings.

"Surviving" is hardly the right word... "thriving" is much more like it.

[Werner Wesp]

Originally Posted by Marty Baggen

I'm thoroughly confused by the metaphorical soup, but I do know that reading Robert and Werner is a lot more FUN than most of the other posts on this thread!

Ha, thanks Marty! we're doing our best! :-)

[Werner Wesp]

I'll have to try that master black -3 thing. Mine is a HD101E and it should be arriving very shortly... Don't know what the latest firmware is (that it probably be on)...

By the way: I know you don't have to use the clutch, but how smooth is the drive then? (especially the 'take off'...)
On the other hand: not using the cluch DOES have nice features: lots of revving sound, smell of "almost"-burning rubber when starting, more regular stops a the garage (so your car will be in better condition?), ... but most of all: if you drive a little (too) sporty WITH your clutch, and that cluch gives in.... man that smell... It is a good thing you're saving yourself from that :-)

I'm still thinking how I can make the analogy back to camcorders....Well let's say: you can, have 2 different kinds of setups in one shot. Always a bonus... and since it's HD, you can use it for 2 different SD images