TRV950 (& PDx10?) Looks great, less filling. But...

So I just got the new TRV950. First the good news.

The picture can be really outstanding, actually rivaling the pd150's in good light. It's video quality is way ahead of the trv900/pd100a, and while it's about a good stop slower than the cam it replaced, it's gain is so much cleaner that there's no contest. In any decent level of interior light (say, that you would care to read fine print under), it does fine. Certainly it's capable of making the best video of any small camera out there.

And small it is. Measuring without the lens hood (as w/most small cams specs), it's more like 6 3/4" long, than the 8" spec I've read. Most of it's width is considerably less than 3 3/4", too. Plus, mini Info-lithium's. This is like Half the size of a vx2000 or GL2.

The downside?

Vertical smear. The chipset can't always seem to handle the contrast between a bright source and darker background, without some pretty impressive banding. The owners manual even mentions this twice, in the troubleshooting section. "This is not a malfunction", it says. -Sure.

Is this going to be fixed in the PDx10? I know some of you will be getting that cam, please aim it around some sunny interior windows and report back.

[Jeff Donald]

Vertical smear is a CCD phenomenon. It can be minimized by optical coatings and the use of 3 CCDs and other filtering methods. But, it never goes away completely.

Jeff

[Mike Rehmus]

You are correct that the smear is from the CCD technology used in that camera.

3 CCD cameras are just, if not more, as apt to smear as a single CCD camera. But 3CCD cameras are more likely to use smear-resistant CCD technology. As far as I know, only the FIT CCD technology is truly smear-resistant.

Optical coatings can reduce reflection artifacts and therefore reduce 'lens flare-type of distortions. But they do nothing for the vertical smear that is associated with bright lights in the visual field.

If you look at the more expensive cameras, you see them placing brand-names on the side of the camera with regard to the CCDs they use. HAD, Power-HAD, & FIT are some of the names (and technologies) they use.

[Jeff Donald]

Yes, 3CCDs in and of themselves will not aid the problem. But as you point out more expensive cameras have 3 CCDs and will use better technology to help reduce vertical smear. FIT (frame interline transfer) CCDs reduce vertical smear by as much as 80% to 90% Panasonic claims. FIT CCDs use a masking or filtering technology to prevent the spill over of light from one Photo Diode (PD) to another. Improved multi-coatings on the micro lenses, used on the surface of CCDs, also help reduce vertical smear. But nothing elimates it 100%, which a simple test will prove. Aim your camera at headlights at night and the vertical smear will still be present.

I read an article several years ago about FIT chips having a problem with 16:9 size. I don't know if Panasonic has overcome the problem or if they are moving away from FIT technology as we adopt 16:9 chips.

Jeff

[Mike Rehmus]

The smearing is not caused by light spilling from one charge site to another but by the discharge of an entire readout shift register IIRC. CCD's are read row by row IIRC and basically, a row gets wiped out by the spread of the discharge caused by the light.

Or some such nonsense. anyway, it is the way that the charge sites (pixels) are read that make the difference in high-brightness sensitivity. The FIT technology costs more and probably has other tradeoffs as well.

According to the head techn at the local Sony Auth Svc Cntr, he can probably adjust the smear it if it's out, "on any camera". He referred to it as "bloom".
I hope this is true. I was under the impression that only replacing the chipset could make a difference... Any techies know the answer?
It seems a little too easy to induce the smear (looks extra bad when bright red...) on this camera when there's a bright enough window in the frame. Maybe this is just one of those early-production run kinks. Otherwise, this cam is like a stealth pd150. Picture's maybe even a touch better. I'd hate to have to get rid of it already...

BTW, the tech I spoke to (on the phone) did mention adjusting using a bright 5600°K source behind a slit, so I know at least we're on the same page. I didn't get to talk to him long, but he seemed pretty sure that adjustment was possible, the new-ness of the trv950 to him irrelevant. I asked if the adjstmnts changed the performance of the camera in any other way, and he said No, unless adjusted to the point where there was no picture at all.

The little I know of the technical side of smear I copy below, from Tektronix' site:

>The camera's iris is first adjusted so the output signal is at the reference white level. It is adjusted again until a predetermined smear threshold is met. The F-stop settings at these points are noted and used to derive the result. The result is expressed in*dB, relative to the selected reference white level. The equation used is:

20 log(ref. level/thresh. level) + 20 log(ref. iris/thresh. iris)<

I believe this is just describing measuring a given camera's smear performance. -What can the tech adjust to improve the outcome of that equation, if in fact mine is out of proper adjustmnt?

[Mike Rehmus]

Not much as it is a function of the CCD. Any adjustment he makes, unless the camera has adjustable image processing capability, will effect the picture in one way or another.

I suspect they can chop the high level of the signal off but only on the output of the CCD.

Mike, could you elaborate please.
I do not want something done that will degrade image quality in any way for the sake of diminishing the problem.

[Jeff Donald]

Mike,

This is a quote from a Panasonic White Paper on FIT CCD. This quote covers smear "When you shoot object with intense spot light, you may recognize white vertical line appear on above/below it. Such a phenomenon is known as SMEAR from which you can not evade as long as adapting solid state for pick up device instead of vacuum tube. How is SMEAR caused in Charge Coupled Device (CCD) which is most popular solid state pick up device ? The light leaking from PHOTO DIODE (PD) or crossing over it yields electrical carriers below PD (inside of Silicon chip)."

psurfer1,

Bloom is different from smear. If your having problems with saturated colors like red, that is possibly bloom and there are adjustments that can be made that will not effect the overall picture quality. I would send your camera back to Sony.

Jeff

[Mike Rehmus]

That may be their white paper but back in the 80's, when my company produced the first digital cameras, the phenomenon was well recognized. It had to do with an excess of electrons bleeding into the readout channels (which themselves are just part of the semiconductor sub-structure) and overwhelming the readout structure. As I recall correctly, the bleed-down of the charge on one array element would effect others connected by virtue of their common read-out.

Furthermore, this phenomenon did not exist in single-line arrays that are, in all other respects, similar to 2-D arrays. In 2-D arrays, the pixel sites are read in rows or columns by shifting the charge out to the charge 'reader' using a bucket-brigade technique. I believe the FIT devices do not use that approach.

In any case, if the solution were to just add optical filters to the arrays, there would not be different CCD technologies. Everyone would just put filters in place and a HAD would be as good as a FIT. Tain't so. Optical filters don't cost so much that Sony wouldn't use FIT on the 1/2 inch and lower-end 2/3 inch cameras. And who among us wouldn't pay more for a PD-150 using FIT instead of HAD?

The easy way to 'fix' the problem, were it light leaking (which does happen, BTW) would be to lay down a neutral density layer around the pixel sites although I believe semiconductor material is a bit opaque anyway. This is the same thing as the neutral density in film backing. It serves to cut down on light spreading through the plastic.

All very interesting. But I'm still trying to find out if there is adjustment that a tech can do for the problem of Vertical Smear.

(To repeat, when shooting a contrasty scene w/bright sunlit window in it, I am getting distinct vertical lines - sometimes they are all white, at other times the lines are all red. Sometimes it's very bad in very high-key situations, and it shows as red vertical bands across most of the picture. I am Not talking about oversaturation; there is no red colored subject in the shot at all with this light-show-like effect.)

This is a brand new camera that operates fine in every other regard.

[Jeff Donald]

Send the camera to Sony. Sony cameras only have a 90 day warranty for labor. Let their qualified service technicians evaluate your camera. I would document the problem thoroughly. Send a tape that shows the severity of your problem. Sony doesn't want defective products in the market place any more than you want to have a defective camera. So, send it in before your 90 days expires. In my experience Sony is pretty strict on their warranty terms. After 90 days your out of luck for the labor cost and labor is usually the most expensive part of the repair.

Jeff

Next week I'll have the local Sony service cntr take a look, but I'm afraid I probably will end up having to send it in, waiting a month or two(?) for it to be sent back with some xerox'd "Sony videography tipsheet", with the line- "...avoid strong contrast" circled in yellow.

[Bryan Beasleigh]

Can you return it as defective?