Calculated estimate of HVX200 CCD resolution

[Bob Grant]

I gotta ask, does anyone here know how a CCD works?
ALL CCDs transfer the charge from the photodiodes to the CCDs an entire frame at a time. The stored charge on the CCDs can then be clocked out in quite a number of ways, has zip to with it being 24p or 50i. One can have a CCD with 4 taps feeding 4 A/D converters or in the case of the HD100, I suspect two A/D converters, the big problem is making 100% certain that the A/D are precisely matched else the SSE problem. So technically one can scan a CCD at 48Hz and still produce an exact 24fps progressive sequence. Whatever the photodiodes have transferred to the CCDs remains until a new transfer takes place. In an interlace system, the whole frame is transferred to the CCDs but only every second line is scanned and then another frame is transferred and the alternate lines are scanned. If you don't transfer the frame a second time, the result is 24PsF which will yield 24p.

All that aside there's gotta be a dead simple way to resolve the CCD resolution issue for any camera. Open the thing up and stick a logic analyser in there to count how many times the A/D converter(s) are being clocked per frame, that'll give the total pixel count, looking at the row and column clocks will also reveal the horizontal and vertical resolution. A way more accurate approach than relying on resolution charts, you remove the lens issue and the human factor. You'll probably remove your warranty too.

[Alister Chapman]

Originally Posted by Bob Grant

In an interlace system, the whole frame is transferred to the CCDs but only every second line is scanned and then another frame is transferred and the alternate lines are scanned.

Not always, many interlace systems create the the two fields by using alternate PAIRS of lines, so for example field 1 would be created by averaging lines 1&2 then 3&4 then 4&5 etc then field 2 would be the average of 2&3 then 4&5 then 6&7 etc. This system doubles the sensitivity and reduces aliasing. It is also why most manufactures are struggling with progressive as most current CCD designs are based around line averaging.

[Jeff Kilgroe]

Originally Posted by Bob Grant

All that aside there's gotta be a dead simple way to resolve the CCD resolution issue for any camera. Open the thing up and stick a logic analyser in there to count how many times the A/D converter(s) are being clocked per frame, that'll give the total pixel count, looking at the row and column clocks will also reveal the horizontal and vertical resolution. A way more accurate approach than relying on resolution charts, you remove the lens issue and the human factor. You'll probably remove your warranty too.

But CCDs are analog devices and direct pixel for pixel or photosite mapping doesn't necessarily apply. Factor in the green layer pixel shift too... It's easy to see where a 720p CCD with a vertical green shift can approximate a 1080p censor and the analog to digital converter can easily sample and interpolate the CCD data at virtually any resolution desired. Both Panasonics marketing propaganda and Jan have stated that the HVX "scans at 1080p". But in terms of what that really means, nobody knows.... It means they are scanning and interpolating a 1080p image from an analog device. Whatever the actual resolution or number of photosites on that CCD may be is not necessarily a direct relation to the native CCD resolution. Simply trying to test column and row clocks from the AD converter will give us the interpolative scan size. Which we already know to be 1080p (when the camera is in 1080 mode) and horizontal interpolation is most likely somewhere between 1280 and 1440 as the camera encodes 1080i/p as 1280x1080 per the DVCPROHD spec.

This question of CCD resolution will probably remain unsolved until Panny finally comes forth with the information (doubtful) or until someone has a junk HVX to take apart and they put the CCD under a magnifying glass and count the photosites.

...1024x768. That's still my guess. There are a few CCDs on the market with native 16:9 and 1024x768 resolution that could meet the requirements of the HVX. But Panny may be using a custom CCD block too. Jan has said the resolution is most certainly not 960x720... It also looks pretty obvious that the vertical resolution is AT LEAST 720 as some res charts can be shot showing 700+ lines.

[Stephen L. Noe]

Desktop scanners advertise the same way. On the box it's 2400 x 2400 resolution but the reality is that it's 600 x 600 optical resolution. Nevertheless, it's listed as 2400 x 2400 scan.

[Jim Giberti]

Originally Posted by Chris Hurd

By the way, what happened to just looking at the image?

Sage words from the Hurdster.
This does get nearly absurd at times. Shooting side by side on set today with the Canon and the HD100 and looking at the HD footage from the HVX...what's the big whoop. They literally all look great - just different flavors.
Now a meaningful discussion would seem to be more on the aesthetic differences (without scientific reverse engineering as to how they get their looks <g>.
For instance as a guy who really likes and has had great success with the Canon in each generation, I really didn't think I would be comfortable with the JVC (just stupid presumptions on my part).
However, using a TV shoot today to shoot the same talent on the same set with the Canon followed by putting the HD100 on the sticks there was no question that there is a substantially "different" look to these cameras using similar settings.
All I can say is the HD100 is smoother in skin tones and overall image and as much as I've loved the Canon, the JVC is, to me, a much more filmic tool.
It feels, looks, handles and produces images that look remarkably like 16mm film.
Not saying that the JVC is better, but for a straight "film" camera, this might be the most realistic tool below the Varicam level.

[Steve Mullen]

Originally Posted by Jeff Kilgroe

It also looks pretty obvious that the vertical resolution is AT LEAST 720 as some res charts can be shot showing 700+ lines.

How can it be "obvious" when the only professional test of ALL the cameras clearly indicate V. rez. is only 540 TVL. However, given that Adam does mention there is a difference in the test charts -- I'm willing to consider that ALL the cameras measured low. I just don't have any reason to switch data sets now.

However, I did try my model with 1080-row and 1080-column chips running PROGRESSIVELY. The "1080-column chips running progressively" option came close enough that I tuned my model to get zero error over all the cameras -- plus the 1080i60 mode of the HVX200 and the Sony in CF30 mode. (The latter I hadn't included before.)

This new model estimates ALL cameras in ALL modes with ZERO total average error. When I enabled both Horizontal and Vertical Green-shift -- as I have heard the HVX200 uses Green-shift -- the exact CCD resolution was 960x810.

Of course, without Green-shift, CCD resolutions increases. I was able to get the model to also work perfectly with 1080x810 (1.33:1) CCDs. Note that this finding is important because this model could be described as "1080 at up to 60p."

So the most likely resolution is 1080x810 running progressively. My model estimates H.rez. will be 561 TVL/ph and V. rez. at 536 TVL. The actual measures were 550 and 540.

Lastly, I don't think Panasonic blew-it by using lower than usual resolution CCDs. There are two VERY good reasons for doing so:

1) Much better sensitivity -- which the test confirmed.

2) No hot CCDs so no need for clocking each half separately so no SSE.

If this is all true -- there was never any reason to hide the CCD spec. Just provide the market with an explanation!

[Stephen L. Noe]

A new wrinkle is that the rez chart was shot at F5.6. I didn't know this either. They need to test it around F2.8 and see the results. Still if the rez dumps to 550 lines at F5.6, its not a good showing.

[Steven Thomas]

Originally Posted by Stephen L. Noe

A new wrinkle is that the rez chart was shot at F5.6. I didn't know this either. They need to test it around F2.8 and see the results. Still if the rez dumps to 550 lines at F5.6, its not a good showing.

Maybe, but this could explain why Barry got better results from his test he perfromed recently. Detail coring was also set to +3 during the camera comparison. Maybe that setting was giving up to much rez for noise improvement.

[Shannon Rawls]

Rez Charts, Smez Charts. People are just fanatical! Who says it was at f5.6? Was it a guess, or was it rock solid assurity? Focal lengths, positions of the camera, the girl moved, noise here, no noise there, what was the other camer settings? maybe the Canon was at f9?? who knows? skin tones were good, skin tones were bad, interlaced scanning, progressive scanning, blown highlights, SSE, no details in the shadows, chromatic abberations, it look slike a chainsaw, it looks like a football, it looks likea bazooka, blah blah blah. And now.....it's all about the MOJO. lolol LMAO LOLOLOLOLOLOL hahahahaha. My god. Now it's the offical "Austin Powers" Camera. *smile*

If you really are on the fence.....THEN WAIT FOR THE TEXAS SHOOTOUT @ THE ALAMO! *smile*
This test should be objective, documented correctly, in an ideal location and under no time constraints or pressure. Cameras will be set correctly and the settings will be recorded by a dedicated person. And honestly people, whatever the test results come out to be.....you already know what you want. So just go buy it.

I wonder.....how many people actually use these cameras professionally outside? I mean like really outside the house on real gigs with real people and real money being spent. People like Barry Green & DSE & Heath & Nick & Chris & Jon Fordham & Steve & Barlow do come here and report because they are Nerds. But Nerds are cool! why? because Nerds get paid, Geeks do it for free! *smile* These guys aren't Geeks. They actually leave their houses and get paid using this daily!

Stop being geeked out over these tests and go buy a camera already. 1 month down 11 to go. You're running out of time in 2006. How many things did you do in the month of January while you were waddling around on the internet waiting for 10 second clips from Germany and test charts from Japan? Don't be an "Internet DP".

Think about it....it's just a camcorder. And it costs less then that 42" plasma HDTV you bought 3 years ago! I bet you didn't geek out over that waiting for tests and argue with people and lose friends over that HDTV before you bought it! So why now over a camcorder that costs less? My wife just bought an all electric automobile (GEM e4) to drive around on the weekends for $11k grand. You think she waited on test charts & argued on the internet with people who owned other type electric cars? LOL. So why now over a camcorder that costs less? I hope you get my point friends. We should never HATE or DISLIKE eachother here. Especially over these dag-on camcorders. It reminds me too much of the pitfalls of human nature. We should all be COOL and FRIENDS and work TOGETHER and combine forces.....we can all get rich(er) that way!

Look at it this way.... If you buy an HVX200...YOU WILL NOT RETURN IT! That camera is bad-ass!!! The image is cold-blooded, and for $6k, you will be OVERLY pleased. If you love the DVX.....it's a no-brainer. So forget the res charts and all that. It's meaningless.

- ShannonRawls.com

[Stephen L. Noe]

We're just guys talking about cameras here......

[Jeff Kilgroe]

Originally Posted by Shannon Rawls

And it costs less then that 42" plasma HDTV you bought 3 years ago! I bet you didn't geek out over that waiting for tests and argue with people and lose friends over that HDTV before you bought it!

Uh... Er... no comment...

Heh. Anyway, you're right. This is pointless, but I think the discussion has drifted to these topics because we're all still waiting for our cameras. Hopefully I'll have mine within the next 10 days. A large shipment is supposedly leaving from Panny QC to several of the larger dealers/distrubutors TODAY. :)

It's amazing how worked up some people get over this camera (and other camcorders in this price range). Sheesh...

[Walter Graff]

"It all suggests to me Panasonic knew review measurements were going to be bad -- and felt giving out the real CCD specification would only add fuel to the fire. I think it's now becoming obvious that hiding information from your potential customers and the press is a bad idea."

Steve, you are closer to the truth than most. They marketed something that they simply did not have and made it sound better than it really was and now it is all backfiring. And now wait till folks find out that 8gig cards don't exist in any quantity for the masses. I was told today that two of the largest retailers are being inundated with folks wanting to cancel their HVX order for other cameras now that the cat has started to come out of the bag. Seems to me to be yet another marketing blunder by Panasonic. THey never were good in that department, broadcast or consumer. Their website disguised as a person who "used to work there" defperception.com and their poor handling of it being discovered as a site that was nothing more than astroturfing should have given folks a clue as to their goals.

That said, any of these cameras produces a great picture when given to the right person. In fact they all perform no better than each other just as their SD DV counterparts do.

[Jeff Kilgroe]

Originally Posted by Walter Graff

And now wait till folks find out that 8gig cards don't exist in any quantity for the masses.

If that's news, then you obviously haven't been paying attention. 8GB cards weren't even supposed to start shipping until mid/late February, but they are available, just scarce. It will take some time for them to show up in quantity as manufacturing catches up. The question I have about 8GB cards is the price... Here in N. America MSRP is $2199 or whatever it is and the cheapest I've seen it from a retailer is like $2089... So how come I can import them from a Japanese shop for $1795 + about $25 to ship them to me in the USA? There were 3 sellers on eBay yesterday selling 8GB P2 cards from Japan at $1800 or less and a friend of mine picked up his HVX200 in Tokyo last week and only paid the Yen equivalent of $1750 for 8GB cards from a retail store.

I was told today that two of the largest retailers are being inundated with folks wanting to cancel their HVX order for other cameras now that the cat has started to come out of the bag.

Told by who? There's no massive flood of cancellations with B&H or EVS... I'm still at about #100 with EVS and I'm #4 (out of 18) at my local shop. Satan will be skiing to work before I get my HVX from EVS... I pitty the poor souls who are in the 200's.. I might get one this week from the local store's current shipment or I may have to wait for the next shipment, they haven't received a confirmed number of shipped units just yet.

[Jeff Kilgroe]

FWIW, I can find 8GB P2 cards easier than I can find an XBOX 360, so it's not like they don't exist.

[Walter Graff]

"If that's news, then you obviously haven't been paying attention. 8GB cards weren't even supposed to start shipping until mid/late February, but they are available, just scarce."

And will be for some time. And then don't expect the next jump to 16gigs anytime soon if at all.

"Told by who? There's no massive flood of cancellations with B&H or EVS..."

By folks at two of the largest retailers. I didn't put EVS on that list.