HVX 3ccd @ 1080?

Thanks Barry!

a) I hope that we should see a lot more bigscreen movies shot in DVCPRO-HD when the HVX arrives...what do you think about the combo HVX + Guerilla35 for a use like that? (I read a beta tester review from you in their forum and you well know the weight to us that your opinion has).

b) Besides Graeme's enlightenment (that native CCD resolution is not so important to the final result), is it possible to count with a native CCD resolution @1080 or will it be always from pixelshift method?
 
In Jan's own words, until this last weekend, we have it:

@Nov 19, 2005 at 4:19:31 pm here

@Nov 20, 2005 at 12:33:40 pm here

@Nov 20, 2005 at 1:11:33 pm here

@Nov 20, 2005 at 1:53:41 pm here

@Nov 20, 2005 at 11:50:26 pm here

Quoting Jan: «The CCD is an analog device. There is not a line for line readout in any CCD camera.»

Quoting mezelf27: «Huh? A CCD has a line for line (well, row per row) readout in its nature. The voltages are read-out analoguely, but the matrix-dividing is done beforehand, making it somewhat digital, i.e. quantisized in nature.»

So, what do we have until now??
 
Emanuel said:
In Jan's own words, until this last weekend, we have it:

@Nov 19, 2005 at 4:19:31 pm here

@Nov 20, 2005 at 12:33:40 pm here

@Nov 20, 2005 at 1:11:33 pm here

@Nov 20, 2005 at 1:53:41 pm here

@Nov 20, 2005 at 11:50:26 pm here
Click to expand...


I imagine Jan is probably looking forward to the camera's release more then anyone else now. Having to chase down all these HVX misconceptions and misunderstands as everyone speculates is a job no one should envy. :(
 
Well, it sounds like Jan's decided to commission a posting on defperception.com that will explain the spatial offset method once and for all.

All I'd encourage you to do is -- don't get hung up about it. Don't even worry about it. Who cares what the chip res is, so long as the picture delivers legitimate high definition. I said that back when I postulated that the chip pixel count could be as low as 960x720, and I'm saying it again even in the face of Panasonic statements calling it a 1080/60p chip. It *doesn't matter* how the chip does what it does, so long as it does what it's supposed to do and it delivers the goods. Spatial offset gives substantial gains in sensitivity, dynamic range, signal-to-noise, and legitimate increases in actual perceived resolution. Isn't that what you want?

Okay, look at it like this: the lens can only deliver a certain level of optical resolution to the chip, expressed in terms of line-pairs per millimeter. Because it is a fixed-lens camera, the engineers know exactly what lens will be used with it, and know exactly what the resolving properties of the lens are. Therefore they could design the chip to perform at maximum efficiency with that lens. If they find that the lens could deliver, say, a maximum of 150 line-pairs per millimeter, then they know that they don't have to design the chip with any more than 300 pixels per millimeter -- any more than that would be wasted and inefficient.

The larger they can make the pixels, the better the video will be, in all ways but one. Larger pixels (read this properly: FEWER pixels on the chip!) mean better video in all ways but one: overall resolution. Better SNR, better sensitivity, better dynamic range. The biggest pixels possible would yield the best performance, excepting resolution.

So -- how much resolution do they need? As much as the lens can give them, and not one pixel more. What if the lens could only deliver 700x700 worth of resolution? Then the chip would need no more than 700x700, right? Obviously any more than that would just be stupid, correct? Okay, what if the lens can deliver 1280x1080 -- then the maximum they'd need would be 1280x1080, right? Any more than that would just be a stupid waste of light-gathering capability. What if the lens can only deliver 1000x800 -- then the chip should have no more than 1000x800, right? Any more would be stupidly wasting all the other benefits they'd get from having bigger pixels, right?

Do you see how this works? They designed the chip to work with the lens, at the peak of efficiency that the lens can deliver. They didn't design the chip for internet discussion groups, they didn't design it to look good on a spec sheet, they didn't design it for bragging rights, they designed it to produce the highest-definition, best-looking picture they could get.

So the ultimate answer to the question "what's the resolution of the CCD" is: IT'S AS MANY PIXELS AS IT NEEDS, AND NO MORE. Whatever that actual count is, is *irrelevant*! It's a completely irrelevant speculation, because it cannot use more resolution than the lens gives it! So why have more? And why have less?

Because it's designed to work with the lens, and only that lens, they can totally optimize the system to deliver the absolute maximum performance in all categories: resolution, signal-to-noise, latitude, and sensitivity. And that's what they've said they've done. The camera has an optimal recording format of 1080p, so now we find out that -- guess what, they optimized the chip to scan in a native 1080/60p format! Isn't that exactly what you'd hope they'd do? And, surprise surprise, that's what they did.

Of course that's what they did. To do otherwise would be utterly boneheaded. They told us from the beginning that it would be optimized to provide the best picture possible, striking the best balance of resolution, sensitivity, SNR, and dynamic range. And from the footage they showed at ResFest, it sure seems like they've done a good job of it!

A 1920x1080 1/3" chip would not accomplish that. A 1920x1080 1/3" chip would have horrible performance. And it would be exceptional overkill. There's probably not a lens made that could resolve 1920 pixels across on a 5.24mm-wide chip -- that would take a resolving capability of 366 pixels per millimeter, or almost 200 line-pairs per millimeter. No lens can handle that. So they didn't do that. To do so would have been stupid. And if some people are disappointed because of it, all I can say is that they should educate themselves on the reality of the situation rather than bemoaning that Panasonic didn't cripple the camera just so they could claim some statistical "bragging rights".

Sorry if I seem like I'm blowing a gasket, but this whole chip discussion is starting to exceed my patience level. Have you seen the footage from ResFest? Have you read the reports of those of us who reported back about the footage from ResFest (meaning me, KHMuse, Evinsky, Rush, etc)... It looks really, really, really good. It's incredibly sharp, it has extremely low noise, it has rich lush color with twice the color res of HDV, and it has *no* motion artifacts whatsoever. It delivers a real, genuine "high def" picture. So -- is there any legitimate reason why anyone should spend even one more half-second's worth of neuron activity contemplating what the CCD resolution is?
 
I know you have said all that before Barry but you have never said it better. :thumbsup:

You should put that in it's own thread and sticky it to the top of the HVX section as a Camera Spec Primer.

With 16 days until, hopefully, all this speculation is behind us, how many more times do you think this topic will be discuss? :grin:

(Won't even contemplate how much it will be discussed after all the info and footage is officially released)

Cheers
 
wabbit said:
I know you have said all that before Barry but you have never said it better. :thumbsup:
Click to expand...
That's why I've been saying that you well know the weight that your opinion has to us! Thanks again for your help: your conviction says everything (I was just waiting for another post in order to have you as last post because all we know the audiences that your name have). But now it's possible to say also:
wabbit said:
Won't even contemplate how much it will be discussed after all the info and footage is officially released
Click to expand...
After the info and footage to have been officially released?... Not so much! We hope...
wabbit said:
You should (...) sticky it to the top of the HVX section as a Camera Spec Primer.
Click to expand...
:thumbup: Good idea! :):cheesy:
 
Barry - Nice Recap. All that being said, I expect to have those specs. With a product the $ the HVX is, such a basic spec is a requirement of the manufacture to provide - not a favour to the customer. How the customer evaluates the spec is up to them.

The fact that a large part of their customer base is asking for it, is reason enough (IMO). I'm willing to wait, even to the extent if they want to show footage before releasing the spec. But it is a requirement for them to release it, not a request. :)
 
It *doesn't matter* how the chip does what it does, so long as it does what it's supposed to do and it delivers the goods...
So -- is there any legitimate reason why anyone should spend even one more half-second's worth of neuron activity contemplating what the CCD resolution is?
Click to expand...
Because some of us actually want to know HOW, for no other reason than understanding. The use of pixel-shifting and lower resolution sensors gives conditional gains in image quality, and if you understand how these are affected, it gives you the power to optimize how you shoot material.

Suppose for example 1/2 your resolution was due to horizontal pixel shift, and you were shooting a scene to be "lit" with entirely red lights. If you did in fact use completely red lighting, you would compromise the picture quality over shooting it white and shifting the hue in post.

-Spiff
 
Well, you can't get 1/2 your rez from PixelShift, only a max of 1.414 factor, or a 41% increase, not at 100% increase.

So, similarly in the worst scenario, you can't loose more than 41% if things didn't work out right.

So lets try and figure out what happens when you shoot pure a red scene. Lets assume only any light gets through to the RED CCD. In this case, you're going to loose rez. But, a) would you notice it, and b) is it possible that only light hits the RED CCD and hence create this condition. And yes, I think shooting it white and CC'ing in post might just work better.

Graeme
 
D_and_G said:
All that being said, I expect to have those specs. With a product the $ the HVX is, such a basic spec is a requirement of the manufacture to provide - not a favour to the customer.
Click to expand...
Well, again, we all would love to have the info, but as to whether it's a "requirement" that they release it: I point you to the example of Rolls Royce who, with their $100,000 price tags, when specifying horsepower they would list it in the brochure as "ample." :laugh:
 
maurowww said:
ehm...ehm...
http://www.adcom.it/prod/13610_AG-HVX200E.html
till now ADCOM never lies...
Click to expand...

Caratteristiche tecniche:
3 CCD 16:9 HD 1920x1080 Pixels reali.
Formati supportati: 1920x1080 50i e 25P; 1280x720 50iP e 25P (con FrameRate variabile da 12 a 50 FpS), DVCPro50 ( Pal 720x576 25P e 50i 4:2:2:), DVCPro ( Pal 720x576 25P e 50i 4:1:1:), DV ( Pal 720x576 25P e 50i 4:2:0:)

means

Tecnical specs:
3 CCD 16:9 HD 1920x1080 real pixels.
Supported formats: 1920x1080 50i e 25P; 1280x720 50iP e 25P (con FrameRate variabile da 12 a 50 FpS), DVCPro50 ( Pal 720x576 25P e 50i 4:2:2:), DVCPro ( Pal 720x576 25P e 50i 4:1:1:), DV ( Pal 720x576 25P e 50i 4:2:0:)
 
maybe but usually they publish the informations that are on the manuals and in this days they are doing public demonstrations comparing the HVX with others HD low cost cameras (i'm not hostile, i'm just enthusiastic)....
 
maurowww said:
Caratteristiche tecniche:
3 CCD 16:9 HD 1920x1080 Pixels reali.
Formati supportati: 1920x1080 50i e 25P; 1280x720 50iP e 25P (con FrameRate variabile da 12 a 50 FpS), DVCPro50 ( Pal 720x576 25P e 50i 4:2:2:), DVCPro ( Pal 720x576 25P e 50i 4:1:1:), DV ( Pal 720x576 25P e 50i 4:2:0:)

means

Tecnical specs:
3 CCD 16:9 HD 1920x1080 real pixels.
Supported formats: 1920x1080 50i e 25P; 1280x720 50iP e 25P (con FrameRate variabile da 12 a 50 FpS), DVCPro50 ( Pal 720x576 25P e 50i 4:2:2:), DVCPro ( Pal 720x576 25P e 50i 4:1:1:), DV ( Pal 720x576 25P e 50i 4:2:0:)
Click to expand...

These specs are incorrect. I can't tell you what they are but this is wrong. Adcom is mistaken. If one thing is clear from what Jan has told us, there is zero possiblity that they are using a ridiculous pixel count like that on a 1/3" chip. So let's just put that behind us. :grin:

Best
 
Barry_Green said:
Well, again, we all would love to have the info, but as to whether it's a "requirement" that they release it: I point you to the example of Rolls Royce who, with their $100,000 price tags, when specifying horsepower they would list it in the brochure as "ample." :laugh:
Click to expand...

that doesn't apply here. If I could afford a Rolls, I wouldn't be dicking around in forums for a prosumer camera. :-) So, some of us prefer to know more details about what we might purchase in the near future, just as if it was a Chevy or a Ford. Keep asking the questions, Toke. It does matter to some people. :-)
 
Barry_Green said:
They may not "lie", but they definitely made a mistake, because Jan has gone on record as saying the CCDs are most definitely not 1920x1080...
Click to expand...

All I can say is thank God they didn't try to make a 1/3" sensor with a 1920 x 1080 imager array!

Can you imagine how bad the noise and sensitivity of such a camera would be?

More ain't necessarily better in every situation.:thumbsup:
 
You must log in or register to reply here.
Back
Top