JVC Says that the split screen can be fixed

[Michael Maier]

Originally Posted by Chris Hurd

It's okay to paraphrase or summarize something that comes from another site, but posting "the whole information" is never allowed, as that's a copyright violation that I don't care to deal with. Giving the link as Barry did was the right thing to do.

I wasn't talking about posting the full article, but the complete information. If he had added that Steve mentioned JVC has a way to eliminate SSE, it wouldn't be any copyright violation. It would just reflect what Steve wrote, which is, firmware won't help, and hardware won't be changed, but there's a solution. But his post can be understood by people as, firmware won't help, hardware won't be changed, we are stuck.

[Barry Green]

Michael, to each his own. I have spoken to JVC myself. They do not consider the split-screen a defect. They figure that every camera has compromises, and this is one that theirs has, and people can choose to work around it or not.

I thought that point had been adequately made. Perhaps in the dozens of times that Steve has posted his list.

But whether Steve considers SSE "not a defect," and whether JVC considers it "not a defect," is not necessarily the end of the issue, for many people. Many people here consider it a defect, and have held out hope that there will be some firmware or hardware fix that makes it go away once and for all.

That isn't going to happen.

They've been waiting for an official statement from JVC -- apparently Steve got one. And that's what the "news" is here. Not that there are yet more workaround rules, but that (as I say again): "it is what it is." Work with it or don't, but don't expect it to change because it's not going to.

[Diogo Athouguia]

So, from what I understood... and I'm getting a bit confused here, SSE on PAL units can be better controled by latest firmware, so firmware updates can fix SSE. The split on NTSC units is not better controled by latest firmware, so firmware uptdates are useless.

PAL units can be fixed, NTSC units can't, is this correct? If this is a limitation and not a defect why some cameras have it and others don't?

Steve Muller's post is being interpreted in many different ways.

[Panos Bournias]

So, coming back. JVC Japan send via e-mail the firmware to JVC Indonesia (Alphateck). The guy reistalled the new software to my camera. This took 15min. Then with a vectroscope and a monitor he went through all the modes of frame rate of the camera and through its color. After a few hours of work he managed to eliminate the split screen and the haze? phenomenon. At the same time he said that the color aberation would be improved. Yesterday we shot with the lense against a 1000W Fresnel light and the reflection looked better than before. It looked almost as in a 35mm lense. Before the fixing of the cam the direct reflections had this boiling feeling or they ended up us a laser spot in the image. I can tell you for sure that the split screen is gone, the haze is gone. JVC guys told us that a new firmware is also under development and as soon as they get it from JVC they will upgrade our cameras again. Another colegue of mine had his camera repaired yesterday. He called and said that he came back from a shooting and his footage was really optimized. My opinion is that in the USA you will have the same solution soon. It is impossible that JVC will stay in a limbo at its biggest market. I will not have the time to ask for technical details as tomorrow morning i will fly to Sumatra for a shooting project and for 2 weeks. As soon as i will get back i think that you will have your problem solved. Panos Bournias

[Barry Green]

Originally Posted by Diogo Athouguia

Steve Muller's post is being interpreted in many different ways.

Well, yes -- it's not just Steve's post, it's that what Steve was told seems to directly contradict what Panos is reporting.

So I guess the confusion will continue...

[Huiy Tang]

If it takes a few hours to do the firmware upgrade, we can all expect to be shipping our cameras off again.
at least it's an improvement. I hope.

[Stephen L. Noe]

Panos is reporting first hand (seems like) with a technician. I was told that a "proceedure" was being done on the camera's in QC. Maybe it's the same proceedure Panos' technician was performing. I do know (from JVC) that the HD-100's are selling by the pallet load.

[Panos Bournias]

We had the same stress with the camera and we were really cautious when shooting, in order to avoid the split screen and the double lines around of some objects.. and even around human portraits sometimes!
The cameras are indeed fixed.
There is no split screen anymore, at least as far as we tested them and all the other obvious defects are noy there anymore.
The technician adjusted our cameras in NTSC and PAL modes, then in 24p mode, each color by the time so it took him around 3 hours to finish with the vectroscope. He was coming back in a monitor pointing the camera towards a white surface. I saw the split screen and how gradually it disapeared from the image. For its mode of the cam separately.
We have the HD101E, that is PAL, but there was an NTSC part and i saw it posted on the screen after he went into the extention of the menu, after NTSC, PAL, as expected, and all the other possible modes of the cam in DV and HDV.

The reality is that the cameras are really serviced and perform well without these serious defects, that have been a bad dream for us...
We are shooting HDV 30p by tomorrow. We will capture and edit in FCP. We will come back hear as soon as we finish. I will try to get a technical description from the JVC guys and post it hear in 2 weeks time. All the Best
- Panos Bournias

[Greg Boston]

Originally Posted by Panos Bournias

We had the same stress with the camera and we were really cautious when shooting, in order to avoid the split screen and the double lines around of some objects.. and even around human portraits sometimes!
The cameras are indeed fixed.
There is no split screen anymore, at least as far as we tested them and all the other obvious defects are noy there anymore.
The technician adjusted our cameras in NTSC and PAL modes, then in 24p mode, each color by the time so it took him around 3 hours to finish with the vectroscope. He was coming back in a monitor pointing the camera towards a white surface. I saw the split screen and how gradually it disapeared from the image. For its mode of the cam separately.
We have the HD101E, that is PAL, but there was an NTSC part and i saw it posted on the screen after he went into the extention of the menu, after NTSC, PAL, as expected, and all the other possible modes of the cam in DV and HDV.

The reality is that the cameras are really serviced and perform well without these serious defects, that have been a bad dream for us...
We are shooting HDV 30p by tomorrow. We will capture and edit in FCP. We will come back hear as soon as we finish. I will try to get a technical description from the JVC guys and post it hear in 2 weeks time. All the Best
- Panos Bournias

Let me throw my technician hat on for a moment and explain what's happening, given the description of the technician's actions that Panos described. The firmware updates are most likely adding more 'calibration points' than were originally available. This makes it possible for the technician to use the scope and monitor to calibrate both sides independently at more points in the response curve so that their outputs match up in all modes under all conditions. BTW, the high tech term for a calibration point is 'fudge factor'. :-)

This is how you overcome the inherent flaw or limitation of the original design. Had the same scenario with a new to market machine in our wafer fab a few years ago. The factory had to revise the software to allow for more calibration points in the magnet response curve so that what was actually happening at the physical level would jibe with what the front panel monitor displayed. Without those extra points, the machine couldn't accurately auto tune itself at various points in the AMU curve.

-gb-

[Guy Barwood]

Here's the bite. It sounds like it could come down to how skilled and motivated your local JVC technician is when it come to 'eliminating' SSE from your camera if what Panos had done holds true for everyone.

3 hours is a lot of time to spend in service for every camera sold, I wonder how long these techs can stand doing these calibrations for? Will they get better with time or just get p*ssed off doing so many...

If this process is a successful one, do you think JVC will start properly calibrating every single unit as it comes off the factory floor? Or only for those who complain?

Would each firmware update require a complete recalibration? That would be nasty for Victor.

I think it is a bit of a joke though that JVC don't consider this a 'defect', even if it becomes a documented limitation, it certainly isn't a deliberate feature. Even Steve has mentioned he hopes they 'fix' it, which means even he sees it as a defect. "If it ain't broke, don't fix it", must be broke then...

If Greg is on track with his analsys, then it is as good as any fix I hoped for(would be nice to be an automatic process but so be it). I really don't care how, just if.

[Stephen L. Noe]

Originally Posted by Guy Barwood

Here's the bite. It sounds like it could come down to how skilled and motivated your local JVC technician is when it come to 'eliminating' SSE from your camera if what Panos had done holds true for everyone.

Also consider, the more of them they fix, the faster they become at fixing them (and more skilled). They'll find some commonalities to the 'fixing' process, that's human nature.

[Thomas Smet]

If this is mainly a calibration issue I don't see how every single camera that comes from the factory could be the same. Perhaps where JVC is coming from is true. The camera using 2 processors and that is that. I think what they do not realize is how bad off some of the camera are. The position they are taking is pretty ignorant considering there is a better way to calibrate the cameras. Usually for the stupidest little issues companies recall cameras to make sure they all work the same.

I do not think the HD100 is defective at all but they are just calibrated badly. The sad part about this is we ourselves cannot do it. If JVC continues to have this position of "hey it is what it is deal with it" I think it will be only a matter of time before somebody posts the info for the service menu so we can calibrate the cameras ourselves.

The funny thing about all of this is that these rapid firmware releases are not just falling down from the heavens into a few european techs hands. Not only must they be coming from JVC but they sure are pumping out a lot of them at one time for the camera being "fine". To say there is nothing wrong with the camera and pump out 5 firmware updates in a few weeks kind of makes no sense.

If this issue can be fixed this easily in a few weeks why wasn't JVC able to come up with this firmware before the camera came out? Were the engineers hoping we wouldn't notice and were too lazy to deal with it or could they really not tell. Or maybe it just goes back to the cameras coming out of the factory with different levels of calibration and nobody at JVC is going to test each camera for 3 hours to calibrate them.

This issue may have been easier for people to take in if it was a common issue with 1/3" HD cameras but as it stands the HD100 is curently the only camera with this issue.

[Greg Boston]

Originally Posted by Thomas Smet

If this is mainly a calibration issue I don't see how every single camera that comes from the factory could be the same. Perhaps where JVC is coming from is true. The camera using 2 processors and that is that. I think what they do not realize is how bad off some of the camera are. The position they are taking is pretty ignorant considering there is a better way to calibrate the cameras. Usually for the stupidest little issues companies recall cameras to make sure they all work the same.

I do not think the HD100 is defective at all but they are just calibrated badly. The sad part about this is we ourselves cannot do it. If JVC continues to have this position of "hey it is what it is deal with it" I think it will be only a matter of time before somebody posts the info for the service menu so we can calibrate the cameras ourselves.

The funny thing about all of this is that these rapid firmware releases are not just falling down from the heavens into a few european techs hands. Not only must they be coming from JVC but they sure are pumping out a lot of them at one time for the camera being "fine". To say there is nothing wrong with the camera and pump out 5 firmware updates in a few weeks kind of makes no sense.

If this issue can be fixed this easily in a few weeks why wasn't JVC able to come up with this firmware before the camera came out? Were the engineers hoping we wouldn't notice and were too lazy to deal with it or could they really not tell. Or maybe it just goes back to the cameras coming out of the factory with different levels of calibration and nobody at JVC is going to test each camera for 3 hours to calibrate them.

This issue may have been easier for people to take in if it was a common issue with 1/3" HD cameras but as it stands the HD100 is curently the only camera with this issue.

Thomas,

Any complex piece of equipment, be it a video camera, or a multi-million dollar ion implanter (the machine I referenced in my earlier post), can have unforeseen issues after getting into real world use. The cameras were probably calibrated at the factory with the available points they had to work with. They have figured out that the original number wasn't enough and have added more to keep both sides matched up in their output. There's a trade-off usually made where just enough points are used and the space between is interpolated, vs. needing more actual reference points and less space between for interpolation. The latter making it more time consuming to set up (3 hours in Panos' description). Also, you have to have non-volatile memory to store that in so it becomes a matter of how much available memory space do you want to devote to holding calibration curve points.

But it looks promising and I think JVC is probably already calibrating at the factory with the newer firmware and probably getting better results out the door than with the first cameras shipped. Time will tell.

Guy mentioned it would be nice if were an automated process and that got me to thinking. Remember the advice to do a manual white balance? That's sort of an automated process using the current lighting and camera set-up which helps to remove SSE for that scenario only. It gives the two sides a single common point in their response curve. The problem is, if you swing the camera and change the light enough, the two sides will become different again because their response curves aren't matched. And that's where I think the new firmware is helping the techs get the camera to behave by having more points to cal the response curve with, thus keeping both sides producing an identical OUTPUT under changing conditions.

Just my humble opinion based on years of dealing with this type of thing under different circumstances.

-gb-

[Steve Mullen]

Greg, I think your experience with complex equipment is helping you "grok" SSE very well.

There is no one "cause" and no single "fix." There are 9 microprocessor controlling the HD100. Each has it's own firmware. Any one of these firmware loads could affect something that in turns effects something else which in turn effects SSE. (By the way, it is other issues we are totally unaware of that is generating these firmware updates.)

It is an assumption that JVC has been working on SSE "fix" firmware. It may be the calibration process that has been getting the attention.

As anyone who has calibrated anything wil tell you -- time is critical and time is money. It's entirely possible the latest SSE firmware (which the USA got and which Europe is installing) did nothing more than make calibration easier and faster and more realiable!

THE BOTTOM LINE IS THAT TALKING ABOUT FIRMWARE IS FUTILE. By now almost every HD100 has the latest "SSE Firmware" and has been adjusted and inspected.

That does not mean firmware that, for example, made AWB more accurate in low light might not help the shooter avoid SSE. It does not mean that a 3 hour calibration won't make ANY production-line camera look better.

You are all free to debate forever. I review production equipment. I'm able to proceed with my review without worrying about V1.172345J. It doesn't exist. (Although, I know all products have on going firmware updates.)

Perhaps more importantly, JVC is willing to have the HD100 reviewed as it has come from the factory.

[Steve Mullen]

Originally Posted by Michael Maier

I wasn't talking about posting the full article, but the complete information. If he had added that Steve mentioned JVC has a way to eliminate SSE, it wouldn't be any copyright violation. It would just reflect what Steve wrote, which is, firmware won't help, and hardware won't be changed, but there's a solution. But his post can be understood by people as, firmware won't help, hardware won't be changed, we are stuck.

I'm glad Berry did NOT quote all of it because the word "eliminated" was a poor word choice. What I mean, is that under some conditions, when you see SSE, a very tiny exposure adjustment up or down will cause SSE to "dissappear." Someone, in fact posted they could make it come and go by slightly rocking the iris. They were right!

Also, my rules do work, but not for all the reasons I thought. But, I'm working on simpler rules. These will not eliminate SSE, yet they may help "prevent" SSE.

Think of SSE as Vertical Smear. Smear is a characteristic of IT CCDs. CCDs can be made exhibit less. But, it can still happen. CMOS, on the other hand, does not have Smear.

So if you buy a CCD camera you need to learn to CONTROL smear. You'll need to learn to control SSE if you buy an HD100. We are not stuck -- but there is no magic bullet coming that will make us totally free of the need to control SSE.

It sure is nice to get out of the rain!