HVX 3ccd @ 1080?

[wabbit]

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)....

OK, to put it another way. Who do trust more? Jan, who will look like a completely untrustworthy idiot if what Adcom is saying is correct, or, Adcom, who has no more inside information then any other retailer and probably just made a mistake in reading the spec sheet?

It's silly to think Jan would mislead everyone and lose her credibility (and probably her job) as apposed to Adcom making a simple mistake. It's even sillier to think Panasonic would design a chip like that with all the obvious flaws that would come with that.

Cheers

 

[Barry_Green]

and then, ok for 1920x1080 CDD but also for 1080 25P?

1080/25p is guaranteed.

 

[maurowww]

some "fantatechnology"... and if Pana had uncovered Pixel clean opposite to Pixel shift?

 

[Emanuel]

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:

Comparing HVX with a Rolls is a must...:happy:...today you are specially humor-inspired...

 

[stephenvv]

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.

Amem brother. Very eloquently stated and thanks for blowing a gasket. The CCD question is so incredibly annoying I personally hope Panny never releases the spec and forces people to actually look and measure the picture to decided it it has "enough" pixels.

More importantly, you perfectly articulated why after owning a XL1 with 4 lenses I decided that fixed lenses were a better design paradigm for low-end DV cameras. Unless you have the bucks for a higher end camera with larger CCDs and real glass options, the engineering advantages of having the CCD imaging and lens system perfectly mated is well worth the compromise.

You would think with all the incredible awful megapixel cameras on the market, both still and video, people get it but obviously not.

We're obsessed with numbers and it may be no accident that with so many males on camera boards and males are obessed with size, numbers and well you know. :evil:

Bottom line, is people just need to grow up and look at the images as they come out and forget the damn CCDs.

But when Panny reveals the specs, there will be tons of people complaining and comparing CCD blocks regardless of what it measures on a rez chart.

Oh well...:embarasse

 

[mezelf27]

I'm with spiff here. The question of the res the CCD's have isn't something to go mad about just because someone asks. If you don't care: fine, but don't shoot anyone that's interested.

You guys are all big fans of Jan, but remember what her job is. She won't be lying, but she won't be stating obvious restrictions of a panasonic product either. And of course there are lots of limitations in the HVX200. If there weren't the cam wouldn't be just $6000.

Some of the things in the posts of Jan over on creative cow are incorrect - but probably meant well to be talking in a positive way about a product without saying anything, really (I'm a scientist. I need facts beforer I can call something 'talking'). I personnally prefer just saying it like Barry sometimes does "if the image is fine, it's fine for me - whatever the amount of pixels" - I take such a comment over a pseudo-technical explenation that isn't quite correct at all. There can't be a mistake: not enough pixels: you need upscaling. Intelligent methods or standard algorithms? doesn't matter: it's upscaling (although 1 way is a whole lot better then the other).

And if people like to be extremely positive and only focus on the good parts of low-res CCD's (lattitude, ...): fine! If they say: if the end-result is fine, I don't care about the pixels: fine! But it is also fine to ask for the specs of a product. nothing to get all defensive about to prove how nice this product is... Stephenvv, YOU might be happy just to look at the image, perhaps we don't? Or we might look differently at it? This is a forum for everyone I guess? Not just for people that are easily satisfied e.g.? Tolerance would be nice. I tolerate those who don't care about the technical specs. You should respect thos who do.

Talking of numbers? Is it okay to talk about the data-rate and compression numbers? Everyone knows by now the JVC records at 19.7 and the DVCPROHD codec is 40 - 100... Isn't there anyone yelling at this? That it doesn't matter? that we shouldn't ask? just look at footage? everyone knows 20 Mb/s interframe holds more info then 40 Mb/s intraframe, I'm guessing that's nothing unexpected? So let's not mention that again then? We just need to judge the footage and shut up? ... Mmmmh, I wonder if there would be a lot of a forum then?

 

[Graeme_Nattress]

Well, scaling != pixelshift. My degree is in Mathematics, I develop scaling alogrithms, among others, and pixelshift is not scaling. It may produce a similar end result that you get more pixels, but it doesn't work the same. I mean, a plane and a car can get you to the end destination, but a plane is not a car!

As for 20Mb/s interframe holding more info than 40Mb/s intraframe - you just have to compare what the compression artifacts are like on the JVC to the Varicam (because until we get an HVX200, that's all we've got) and I'm sorry, but the JVC codec is not twice as efficient as the DVCproHD one. It may be close on static stuff (but nothing in video is really static) but on any movement of any kind at all, it's well below DVCproHD quality. You've just got to watch the video to see that. And although the luma on static subjects is very close (although you don't see any extra definition from it's practically useless higher resolution) the chroma artifacts are significantly higher. So when it comes to compression you can't talk numbers, only visuals as you're not comparing like with like.

Graeme

 

[mezelf27]

Well that's it, isn't it? not comparing like with like. exactly my point. The function for relevant information divided by storage space - data rate (which is exactly the mathematical 'quality' of some compression). This does tips the scales in favor of MPEG2 - due to 'higher order correction', in this case: the time. In DVCPRO-HD, all this excess information is clouding the datarate and using it not very efficiently, except for very, very, very fast motion (e.g. when the cam is on, but you don't know).

Anyhow, I disagree with you on pixelshift. Of course you don't need a mathematical upscaling, but the upscaling has been done by a 'trick' in CCD positioning. As it's nature in physics it has some advantages, but some distrinct disadvantages. In no way it is a substitute for the full resolution. Actually, I personally have a high-enough-order spline interpolation on some shots over pixelshift.

I am very well aware what pixel-shift is, and I am very very aware of compression algorithms. As a master in Physics, my work was developping the algorithms ro reduce the info (datarate) but maximize the relevant info for pattern recognition etc... It is absolutely different from trying to keep the visual quality, but it is very much related.

 

[Graeme_Nattress]

It doesn't take fast motion to flummox MPEG2 - almost any motion can do it. MPEG2 as implemented in HDV cameras is not efficient enough to give at an advantage over HDV, and the pictures show that.

As for Pixelshift, by using larger pixels you gain elsewhere. I don't like high order spline interpolation due to ringing artifacts.

Graeme

 

[mezelf27]

As for Pixelshift, by using larger pixels you gain elsewhere.

Well, this is exactly what I'm saying: if you like the larger pixels and the lattitude that comes with it better: fine. But some of us might go for the full possible resolution... Now, it almost sounds like an excuse for lesser pixels. It isn't and it shouldn't be. Okay, it might be somewhat softer, but it has better lattitude. fine. it's the pro's and the cons. People aren't always wanting to hear the cons. no problem for me, but that's no reason to pretend there are no cons, or just plainly deny them (this is NOT meant particularly to you, graeme, sorry)

 

[mezelf27]

It doesn't take fast motion to flummox MPEG2 - almost any motion can do it. MPEG2 as implemented in HDV cameras is not efficient enough to give at an advantage over HDV, and the pictures show that.

And how exactly do you take into account that the one is a 1280x720 and the other a 960x720? SD MPEG2 with a datarate of 20 would be quite nice....

 

[stephenvv]

Stephenvv, YOU might be happy just to look at the image, perhaps we don't? Or we might look differently at it? This is a forum for everyone I guess? Not just for people that are easily satisfied e.g.? Tolerance would be nice. I tolerate those who don't care about the technical specs. You should respect thos who do.

I very much care about technical specs that are relevant to image quality, usability, price, reliability etc. However, the exact CCD rez and pixel count of a camera is not one of them, for the reasons cited here and on many other still and motion digital sites all over the world.

I'm pretty obsessed with what the CCD sensitivity, noise and latitude are. But unless you or others can cite reasons that the exact CCD pixel count matters I don't see why thread after thread complains that the spec is "missing" and infers that Panasonic is doing something untoward.

Ironically, all the CCD rez discussion confirms to Panny that they are right in not disclosing it :laugh:

 

[khmuse]

Ironically, all the CCD rez discussion confirms to Panny that they are right in not disclosing it :laugh:

A voice of reason... unfortunately, this is seems to be a rare commodity.

I am reminded of a conversation that I heard while attending a trade show. When being asked yet another irrelevant question in a long series of ill-formed queries, the engineer turned and replied to the attendee...

"If I were to provide you with this information, what would you do with it?"

 

[stephenvv]

"If I were to provide you with this information, what would you do with it?"

LOL - that's hilarious and a wonderful summary of the point.

An extra stop of latitude, now that's something I can use. An extra few thousand pixels, well I suppose I could use them for digital image stabilization :evil:

 

[D_and_G]

Ok, so let me get this straight. There's NO relevance to knowing the specs on the CCD's rez and pixel count.? Fantastic ! Perhaps we should inform all the other camera manufacturers. I'm just a neophyte, but to me that sounds ludicrous. I understand the reasons not to jump to conclusions (thanks to Barry ), but to say there is no relevance to the image with regards to the CCD specs seems to be stretching it. Applied relevance yes. Zero relevance no.

Customers aren't asking for some arcane little tidbit of info. , that has zero impact on the cameras process.

I very much care about technical specs that are relevant to image quality, usability, price, reliability etc. However, the exact CCD rez and pixel count of a camera is not one of them, for the reasons cited here and on many other still and motion digital sites all over the world.

I'm pretty obsessed with what the CCD sensitivity, noise and latitude are. But unless you or others can cite reasons that the exact CCD pixel count matters I don't see why thread after thread complains that the spec is "missing" and infers that Panasonic is doing something untoward.

Ironically, all the CCD rez discussion confirms to Panny that they are right in not disclosing it :laugh:

With regards to your last sentence - Any company that even intimated that they won't disclose a primary element in their product because the customers are too stupid to integrate the information properly, is one company I won't be purchasing from. Frankly, that is insulting and condescending. Hopefully, Panny isn't that shortsighted.
When i was working with Arri group, they were willing to get into minute technical details, let alone primary systems. Not once did I hear , well why do you want to know, or you're too stupid for that information.

I think those sentiments don't give the customer enough credit.

IMO, The question should be, why shouldn't they release the specs? A few people who will jump to conclusions is not enough of a valid reason.

Anyway, differing opinions. :beer:

Cheers.

 

[stephenvv]

Ok, so let me get this straight. There's NO relevance to knowing the specs on the CCD's rez and pixel count.? but to say there is no relevance to the image with regards to the CCD specs seems to be stretching it. Applied relevance yes. Zero relevance no.

I think we are arguing semantics here as usually is the case. Obviously if HVX had CCD of res 640 X 480, there might be an issue with rez (as if they were 10,000 X 20,000 which introduce all sorts of artifacts as well).

The issue is that Panny is most likely tweaking effective vs. total number of pixels as well as precise pixel shift and other CCD techniques. The 60i footage is already out and we know for a hard fact the image is plenty sharp for 1080i footage. Barry's point is the point. How many pixels? Enough.

With regards to your last sentence - Any company that even intimated that they won't disclose a primary element in their product because the customers are too stupid to integrate the information properly, is one company I won't be purchasing from. Frankly, that is insulting and condescending.

Well, I agree that it may not be the wisest policy for some customers, but then again count the HVX Latitude threads vs. the CCD rez threads and it's clear that many people are not that well informed about what really matters with a CCD block as I was when I first got into DV.

One thing Panny could do is as they did with 24p when the HVX came out is publish the CCD specs with a nice whitepaper explaining this issue for folks.

But like 24p, despite all the info about it after all these years, there are still people complaining about it being "too strobey" compared to film...:shocked:

 

[Tzedekh]

Ok, so let me get this straight. There's NO relevance to knowing the specs on the CCD's rez and pixel count.? Fantastic ! Perhaps we should inform all the other camera manufacturers. I'm just a neophyte, but to me that sounds ludicrous. I understand the reasons not to jump to conclusions (thanks to Barry ), but to say there is no relevance to the image with regards to the CCD specs seems to be stretching it. Applied relevance yes. Zero relevance no.

Of course the pixel count is relevant. You can't get an HD image out of SD CCDs without radical spatial offset, such as with Reel Stream's Andromeda DVX modification. Will it be 1,920 x 1,080? No. Panasonic tried to engineer what they considered the best compromise, given a host of factors (lens resolving power, CCD size, light sensitivity, cost, etc.). So they are trying to explain why the HVX's CCD won't win the pixel-count prize yet will yield an adequate image. But is it true that a 1,920 x 1,080 third-inch imager produces a really bad image, as some here have claimed? Not necessarily. By all counts, the CMOS chip used in the Sony HC1 yields a pretty good image. It's effective 4:3 pixel count is 1,920 x 1,440, and 1,920 x 1,080 at 16:9. At 60 lux it suffers compared with the Z1's CCD, which is to be expected. But one must remember that it's a first-generation consumer HD chip -- the technology will most certainly improve. My point? Don't believe the claims made by either side. Follow the advice often given in this forum: Let your eyes be the judge.

 

[Luis Caffesse]

Ok, so let me get this straight. There's NO relevance to knowing the specs on the CCD's rez and pixel count.?

That's right.
There is NO relevance in knowing those specs until you can see footage from the camera.

Tell me, what would you do with that information if you had it right now?
How would it help you exactly?

 

[Barry_Green]

Talking of numbers? Is it okay to talk about the data-rate and compression numbers? Everyone knows by now the JVC records at 19.7 and the DVCPROHD codec is 40 - 100... Isn't there anyone yelling at this?

EXACTLY EXACTLY EXACTLY! You've just crystallized the point perfectly.

You see people latching onto 100, comparing that to 19.7, and they say "ah ha, 100 is better, only an idiot would consider 19.7". See? See what throwing numbers around with people who don't understand the ramifications accomplishes?

That's why I'm saying: what good would it possibly do, to 98% of the potential market, to go trumpeting these numbers? There are a very, very few here who could make valid use of the information. And there is a vast majority who could (and probably would) draw the wrong conclusion, under the prevailing "wisdom" that "more = better." And to them, publishing a lower number would harm their case and be a just plain bad PR move.

You've just got to watch the video to see that. ...(although you don't see any extra definition from it's practically useless higher resolution) the chroma artifacts are significantly higher. So when it comes to compression you can't talk numbers, only visuals as you're not comparing like with like.

Again, exactly. Graeme and I are both programmers with extensive experience in video, guys who made our livings pushing numbers around, and he's come to the same conclusion that I have -- you can't talk numbers. Visuals is the only thing that matters, the only thing you can rely on to make a qualified judgement. Just looking at numbers will lead one to draw conclusions that simply aren't borne out by looking at the visuals.

"If I were to provide you with this information, what would you do with it?"

Brilliant, and exactly the point. If one is qualified to use the information, it's great to have. If one is not qualified to use it, and it can be construed to be misleading, then -- why publish it? How many of us, honestly, are qualified camera engineers who truly understand CCD design and have a full grasp of the tradeoffs required to engineer the best result? Anyone? Anyone?

Okay, different question: how many of us are qualified to look at the images and say "wow, this one looks a lot better than that one for my purposes"? And that includes the average customer as well as the physicist, the programmer as well as the filmmaker.

I don't doubt that they will publish the number someday. But I also have no doubt whatsoever that they will wait until they have representative footage samples and even comparison footage samples. We've seen enough posts by "Radek" decrying the VariCam vs. HDV, whining about how HDV will look "better" because of the Varicam's subsampling. Well, now we can empirically prove that the VariCam kicks HDV into the weeds, and the numbers do not tell the story, the visuals do. Once there are adequate visuals out there to tell the HVX's story, then they will probably reveal the numbers. But they won't do so before then -- it'd be stupid and foolish on their part to do so, because a) obviously the number is lower than some of the competition, and b) because of that, people will leap to the wrong conclusion. Guaranteed.

 

[stephenvv]

Of course the pixel count is relevant. You can't get an HD image out of SD CCDs without radical spatial offset, such as with Reel Stream's Andromeda DVX modification.

That seems to be a contradiction? And Sony single CMOS is not really a good example as the HVX is a 3 CCD camera.

Everyone seems to be a camera engineer. I still don't know why if the image looks sharp and resolves HD rez, why does the CCD count matters?