HVX 3ccd @ 1080?

[Graeme_Nattress]

But if you'r resigned to 4:2:2 anyway, sacrificing some chroma rez for luma rez is not loosing you any resolution at all, perhaps only in the very most of extremely saturated video, something you're very unlikely ever to see or record, or even notice a slight loss of luma rez.

Also, thinking about scaling != pixelshift. Pixelshift is very much closer to de-bayering than scaling.

Graeme

 

[Barry_Green]

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

As far as an average customer making a choice, I think that's correct. As long as customers are conditioned to think "more = better" (witness how many think 1080i is "real" high-def and 720p "isn't"!) How many people think 720p is just a "stopgap" on the way to "real" 1080i? Fact is, broadcast 720p looks better in almost all circumstances than broadcast 1080i does.

More is not always better. Would you like more beestings, or less?

but to say there is no relevance to the image with regards to the CCD specs seems to be stretching it.

I didn't say there was no relevance to the image. I said there was no relevance to the consumer's ability to decide which is better based on knowing what the spec is, and I stand by that. If someone bought just based on specs, they'd buy a JVC HD1 over a Sony FX1. And that would be a horrible, horrible mistake! :laugh:

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

No, they're asking for an arcane little tidbit of info that will mislead many of them into thinking that one is a superior performer over the other, when in all likelihood the exact opposite is true. Going for native pixel size gives the HD100 some nice resolution, but it also brought out the split-screen effect. I think that's a far more relevant little tidbit of info and I'd rather know about it than know or care how many pixels they used.

You've to to also recognize that this 1/3" high-def world has changed the rules. Pixel size was never so important as it is now. Diffraction is such a powerful influence over the image, it's now something that has to be actively dealt with, whereas previously it could fairly safely be ignored. A 1/3" high-def chipset may resolve 1000 vertical lines at f/2, and may drop to something like 300 lines at f/16! There's a very good reason the Canon XL H1 stops the iris at f/9.5. This is something that shooters have never had to deal with before. Pixel size is far more important on 1/3" high-def cameras than it ever has been on any other CCD size.

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.

Panasonic has never said it. I'm the one who's saying that the vast majority of customers don't have the base of knowledge to know why one arrangement is better than another (and nobody has ever said "too stupid" -- there's a difference between recognizing that the customer isn't fluent in camera engineering philosphy, and saying that they're "stupid"!)

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.

And how many customers did the Arri group sell to, vs. how many mass-market customers that these prosumer cameras go to? And what was the pricetag of the Arri product, vs. the HVX? I think you're looking at, for the most part, an entirely different caliber of customer. They may sell 100 535's a year, to Paramount and Lee Utterbach and such. Panasonic will be selling that many HVXes per DAY, to people who are trying to save $50 by buying from Buydig instead of B&H. The level of understanding you're going to get from the average Arri customer probably far exceeds the level of understanding from the average 1/3" HD camcorder buyer. And when you're dealing with the mass market, like it or not, your mass market marketing materials have to be mass-market friendly. Publishing an accurate but fundamentally irrelevant number that can be misconstrued by the mass market is just not a wise choice.

Ideally the manufacturers should be publishing RELEVANT statistics, things that actually matter, rather than IRRELEVANT and potentially confusing statistics. Which is more relevant: the pixels on the chip, or the lines of resolution in the image? Forget the useless pixel count statistic, and instead find a standardized resolution chart and publish the actual resolved resolution in the live image and in the recorded image. THOSE are numbers that we can ALL use, without getting mislead. Those are a case where "bigger = better" and is actually relevant, whether someone's a beginning aspiring filmmaker or an accomplished programmer or physicist or academy-award-winner.

IMO, The question should be, why shouldn't they release the specs?

Because it is irrelevant, and potentially misleading. Like lux ratings -- they're not standardized, and most pros know that the lux ratings on camcorders are fantasy, but customers read 'em like they're the gospel.

I'd instead say "why don't they release relevant specs?" Standardized, useful specs. I'd like to know how many footcandles it takes to render 55 ire at f/5.6 off an 18% gray card with no discernible image noise. That's a spec that matters, not some bogus "f8@2000 lux" spec where you find out after the fact that they had the camera on 18dB of gain and there's grainy noise the size of snowballs on the screen. I'd like to know how many horizontal lines of resolution a camera records and use that as a basis of comparison, rather than hearing that one is "1920x1080, except it actually records 1440x1080, off of a 960x1080 pixel-shifted CCD, which of course is interlaced so you don't get an actual 1080 lines of resolution, it's more like a max of 820, but then again the lens can't resolve that much anyway." I mean, seriously, what is that supposed to mean? How can you possibly judge that based off the printed specs? Would you know that your 1920x1080 image off a 960x1080 CCD is actually delivering somewhere around 700x775 in terms of actual resolved definition? And what good does it do you to know that the CCD is 960x1080, when the relevant spec is the resolved resolution of 700x775, and whether it has a split-screen effect running down the middle?

I'm all for specs. I just want them to be meaningful and relevant, not some trumpeted marketing blather that internet discussion groups argue over for "bragging rights". The "bragging rights" specs are meaningless, misleading, and irrelevant.

 

[Graeme_Nattress]

Measured resolution on test images is meaningful, CCD rez is not. Way to go Barry!

I've said it before, but the way some people think of resolution is akin to copying a VHS tape to HDCAM and now saying it's a high def image as it has 1440(or 1920 )x 1080 pixels in it. Yet it's still very low definition. Definition = (resolution*contrast*detail) or near enough in my book - you need all three components. I want high definition not resolution.

Graeme

 

[Spiff]

This whole argument is just defensive bollocks, trying to justify what will eventually be a lower pixel count on the HVX200 CCDs.

The facts are:
a) We expect some pixel shift in the HVX200 CCDs
b) We have a good understanding as to the benefits (lattitude and sensitivity over higher rez chips, resolution over similar chips without pixel shift) and compromises (conditional resolution gains, lower resolution compared to native rez CCDs) of pixel shift

The argument for not presenting the information is that b) is not true. Panasonic is stating that they don't believe that consumers understand the benefits of pixel shift enough to consider it "a good thing".

CCD resolution is meaningful. It is one of several defining specifications of a camera... but it is by no means the ONLY one. Don't close your ears, go "la la la" and say it isn't meaningful. We wouldn't have this discussion at all if it wasn't.

-Spiff

 

[Graeme_Nattress]

CCD resolution is not meaningful though!

What is meaningful is the full resolution of the lens+ccd+dsp+recording+playback system. Just looking at one component of that system and assigning a single performance figure to is like measuring the total performance of a decathalon athelete on how fast they can swim 100m.

Graeme

 

[braw]

I aint takin sides, and I aint sayin Barry or Graeme are wrong, but we would like to know regardless. I am not really comfortable as to the explanation being that we just wouldn't understand or get why it is the way it is. Like everything else, we would hash it out here. Barry and Graeme would be able to set us straight as to how it's doing what, but simply not letting anyone know is a little suspect.

What are they going to do? never release the specs, or release them with the camera? If they ever intend to release them, then what's the point of holding out? So it's probably true that most of us would make uneducated assumptions at first. We already have a billion times over. I made a few about the DVX before I got here. I promptly learned from the few that truly knew what was up and then BAM, I was educated. Same applies to the HVX.

Why is it that all cameras list the CCD size as a staple feature listing? Because it is. While I understand as probably most at this point that there are some very crucial variables...that doesn't stop us from wanting to know. My decision is not going to be based on pixel count, but by footage. Still doesn't stop me from wanting to know and feeling a little toyed with by Panasonics marketing machine.

 

[Graeme_Nattress]

CCD resolution is an "easy" spec - that's why it's listed. However, put a high rez CCD behind a poor lens, and now that CCD spec is meaningless as the lens doesn't pass enough detail through. Put a a great lens on the front, but a poor DSP on the back, and again, you've got poor detail. Put a VHS recorder on the back and you'll wonder why you ever bothered. That's why the only resolution spec worth reading is the one that measures the resolution of the entire system - not just the CCD spec.

The fact is the people buy digital cameras by their megapixel count and never look at the image - and it's the same with video cameras - and by that argument, a Sony Z1 produces a "better" picture than a Varicam, a top end DigiBeta and an SDX900, and do you know what - it doesn't.

Graeme

 

[ImBiggie]

CCD spec should be known.

CCD spec should be known.

Hi,

Have been reading the heated discussions about HVX CCD size for months now. Although, I don't care about the cam pixel count, I do feel that potential buyers have the right to know the spec (CCD). It's up to potential buyers to decide if certain information is relevant or not after specification as been published. If certain spec might cause confusion or misunderstanding, it's up to the manufacture to make their case. Not that it's true but Panasonic seems not to proud of the pixel shifting technology. Why are they hiding the spec of an essential component?

Barry,

I have learned a lot from your posts. I have tremendous respect for your knowledge but you starting to look very defensive on behalf of Panasonic. You starting to sound like Panasonic can't make a mistake. When the HVX200 was first announced the most requested spec was the pixel count. Panasonic should have realized the importance of knowing the CCD spec then and make it available soon after to potential buyers. Panasonic made every other spec details available except for the CDD, Why?

 

[Barry_Green]

What are they going to do? never release the specs,

Maybe. That's their preogative. Perhaps the particular layout is somehow indicative of some trade secret or proprietary process that they don't want to reveal? I don't know, all I know is they ain't talking. At least not yet.

or release them with the camera?

Maybe. Or maybe six months after. Or maybe never. We don't know.

If they ever intend to release them, then what's the point of holding out?

I gave one compelling scenario a while ago. Probably said it better then, but I'll try to paraphrase here.

Let's say the pixel count is something preposterously low, like 400x300. And let's say that they released this pixel count back in April when they first introduced the camera, and said "but just wait until you see the footage!" What would people's reaction be? I'll tell you exactly what it would be: they'd say "bah. Can't possibly be any good. I'll buy me a JVC instead, at least it has NATIVE PIXEL COUNT." (of course, then they'd say "what's with this split-screen? what's with these lens aberrations? why is the image noisy? ... but at least they'd have their precious "native pixel count!")

Then, 8 months later, Panasonic finally has a working model and releases footage, and all the guys who said "bah" would say "I can't be bothered... if it doesn't have NATIVE PIXEL COUNT, it can't possibly be any good." and the others, who might be tempted to look at it, and see jaw-droppingly awesome footage that blows them away, would then say "wow, looks great, wish I could have waited, too bad it's too late, I already spent my money on the JVC."

So in scenario 1, "release pixel count first", they lose across the board.

Let's go to scenario 2: forget the pixel count, just release the footage. On December 8th they release footage that stops you in your tracks and makes you say "wow". Makes you say "boy, am I sure glad I waited -- now I can get REAL high-def footage ... look at the edge-to-edge sharpness, look at the lack of chromatic aberrations, and hey -- no split-screen! Look at the better color fidelity and the better color sampling and the complete lack of motion artifacts and macroblocking. Why, this is the camera for me, I simply must have it!" And then, after you've made your decision, then they say "oh, and by the way, we accomplished all this magic with a pixel count of just 400x300."

What would the reaction be? Would anyone seriously stop in mid-transaction and say "oh. well. I'm afraid that pixel count is just too low. Sorry, but I'm now going to walk over to the JVC booth and buy an HD100 because it has NATIVE PIXEL COUNT, and that's more important to me than actual better footage." Do you think that would happen? Not a chance. After they've seen the footage, they won't care what the pixel count is, other than to say "really? You got that kind of results out of such a low pixel count? Amazing!"

So obviously this scenario only works if the footage is that superior, etc. Yes I'm using the extreme to illustrate the example. But I think that a few years of watching internet forum behavior has convinced me that this scenario is basically exactly how things would play out.

 

[Spiff]

What is meaningful is the full resolution of the lens+ccd+dsp+recording+playback system.

Which is why I want all the specs.
Lens resolving power in lp/mm
CCD size and resolution
DSP bit depth and sampling criteria
CODEC formats and recording options etc.

You act as if the final measured values are all that's important... but every time a resolution chart comes out 50 people line up and read it differently. Not to mention no one knows if the camera was set up or focused properly, at the right gain, with a good tripod, no vibrations, what the colour settings were, etc. etc. etc. Give me a single measurement from a system as complicated as the HVX, and I couldn't tell you much about it at all.

Give me the raw component specs and I can derive the results for most tests, make theoretical predictions and test those. Without it, I'm in the dark and have no idea what to expect. "Point it at something and see what you get" isn't acceptable.

Basically, all I see here is:
"We're afraid people won't like the CCD res"
and
"We think they're too dumb to figure out our camera will be great"

Frankly, Barry - you're hyping the damn thing. To me it just reads as "I like Panasonic! I'm a fan! They r0x0rz!" I'm glad you've seen footage and you like what you've seen. But this is just getting obscene. Your examples are theoretically impossible, and your attacks on the JVC and HDV in general make it seem impossible to get good results from any format but DVCPRO-HD - which is inferior to nearly every other codec at the same bitrate.

We're not stupid and we're not uneducated. We do understand a good deal of what's going on thanks to helpful posts from you and your comrades on the net over the last several years. No one is going "wahhh!! Native resolution only!" No one is saying "I won't buy it because of..." They're just saying "release the damn pixel count".

-Spiff

 

[Barry_Green]

but you starting to look very defensive on behalf of Panasonic. You starting to sound like Panasonic can't make a mistake.

Oh, they can definitely make a mistake. Panasonic's been responsible for some of the biggest bonehead moves in digital camera history. The original AG-EZ1 DIDN'T HAVE A FIREWIRE PORT! What's the point of a DV camera if you can't access the DV data? Dumb, dumb, dumb move. They also introduced DVCPRO-P, a 480/60p version of DVCPRO. Footage looked great. Not one editing system ever supported it. Stupid move on their part, they should have lined up editor support before ever putting the product on the market. And of course they're well-known for the classic M-2 disaster. And DVCPRO/25 in PAL countries uses 4:1:1 instead of the universal PAL 4:2:0 sampling used by every other digital format. I know why they did it, but it's still a mistake, it was a mistake the moment they did it.

This rant isn't in "defense" of Panasonic, it's more against mindless sheep mentality. I bought an HD100. I used it. I sent it back. I know what it's capable of, and I know what it couldn't do. I heard all the complaining about JVC but I took a chance anyway. I got burned. But I still had to know for myself. I pre-ordered a Z1. I spent a week with the FX1, and I canceled the Z1 order. I think it's a fine product for some uses, but not the ones I wanted to put it to.

I've seen HVX footage. It was significantly superior to anything I saw from an HD100 or FX1. It's very difficult to sit still and see people get worked up in a lather over something that is, fundamentally, irrelevant when they should rather devote their energies towards something that has some bearing, some meaning, some relevance in any way.

When the HVX200 was first announced the most requested spec was the pixel count.

Sure it was. And what would people do with that spec?

Why, they would compare it to the other cameras, that's what they would do.

And would that comparison be meaningful?

Not by any stretch of the imagination.

Would it be misleading? Probably.

So -- please, tell me -- what value would it have?

Panasonic made every other spec details available except for the CDD, Why?

See my post above, I think it probably illustrates the reasons quite nicely.

 

[braw]

However, put a high rez CCD behind a poor lens, and now that CCD spec is meaningless as the lens doesn't pass enough detail through. Put a a great lens on the front, but a poor DSP on the back, and again, you've got poor detail.

I absolutely agree with you. My point is that witholding a spec based on an assumtion that consumers are dlinically retarded and therefore can't have the scisors is pretty insulting to the kids that are just in special ed because they have a screwy home life, and can actually be potty trained.

Well...if that made sense. My argument isn't with what's important about the camera, it's that Panasonic is withholding a very basic spec supposedly based on it's assumption of our level of understannding. I am assuming that's the reason. Maybe it is higher than we expect and want to "surprise us" by causing us to doubt it. Who knows what their tactic is.

The fact is the people buy digital cameras by their megapixel count and never look at the image - and it's the same with video cameras - and by that argument, a Sony Z1 produces a "better" picture than a Varicam, a top end DigiBeta and an SDX900, and do you know what - it doesn't.

Again, I wouldn't begin to argue with what you say. You're absolutely and depressingly accurate in your statement. I guess I just assume that since this isn't an Xbox 360 or PS3 and won't be in Best Buy for Christmas, we as "prosumers" or whatever the hell you want to dub us could get a little more credit than the 280lb SUV driving soccer mom buying "Whore Mangler 3" for their 6 year old kid.

 

[braw]

Fair enough Barry. I guess I can only really gripe if they never release the specs. I can't really make a buying decision until it's released and you along with a handful of others run it through the obstacle course.

 

[GGman]

[QUOTE= "release the damn pixel count".


Agree. It's my money and my personal desire to know some info. Now if any other person out there that keeps defending Pana and telling some of us what we only need to know and should not know, then maybe they should pay for the camera and give it to me. There will be no explanation or analogy presented by anyone to make me feel I just do not need to know something to make a decdsion on a purchase. Hoo Hah.

GGman

 

[harlan]

Anyone remember MII (M2)?

 

[Isaac_Brody]

I hope they never release the pixel count. :evil:

 

[Graeme_Nattress]

MII gave a better picture than BetaSP, so it wasn't a technical failure. I had an MII and it was lovely. What stopped it was the broadcaster mentality that "only" Sony is good.

As for 4:1:1 on PAL - I don't see that as a mistake. 4:2:0 is a ghastly abomination in interlaced video, 4:1:1 is not.

Graeme

 

[Barry_Green]

As for 4:1:1 on PAL - I don't see that as a mistake. 4:2:0 is a ghastly abomination in interlaced video, 4:1:1 is not.

Graeme

4:2:0 is indeed ghastly, and I said in the post "I understand why they did it" -- but to make their format bitstream incompatible with every other flavor of DV was, I still contend, a mistake. Yes it may look a little better, but who cares if nobody's buying it?

In America DVCPRO has had great success; according to Panasonic it's captured 70% of the news market away from BetaSP. I don't know if it's done so well in Europe.

 

[pertierr]

I think I completely understand why they are keeping the pixel count quiet for now. Everybody will obviously eventually know one way or another.

However, there is this incredible fixation on all the internet forums with resolution, it's amazing. People really do think that anything with more pixels will look more like film, and that couldn't be farther from the truth. The reason being is that when you increase pixel count, you may gain resolution thus improve in one area, but then you decrease performance in multiple other areas such as noise and dynamic range, which some would argue are more important to getting your footage to look like film. This is especially true when the data gets pounded on by heavy compression codecs such as DVCPROHD.

Knowing this, the smart people at Panasonic probably attempted to keep the pixel count as low as possible, and I expect the DVX to yield better sensitivity, noise performance and dynamic range than the competing HDV cameras out there.

Of course, the one down side is that this goes against the public fixation with resolution numbers. But this can be alleviated just by keeping it quiet and letting the images speak for themselves.

I think Panasonic is probably confident that when people see the images, the pixel count will make little difference in their decision to purchase the camera.

Just my 2 centavos.

 

[ImBiggie]

Sure it was. And what would people do with that spec?

Why, they would compare it to the other cameras, that's what they would do.

And would that comparison be meaningful?

Not by any stretch of the imagination.

Would it be misleading? Probably.

So -- please, tell me -- what value would it have?


See my post above, I think it probably illustrates the reasons quite nicely.

Barry, most buyers know better not to buy a camcorder base on one spec (CCD) alone. But the information should be readily available if needed. Panasonic have many documentations explaining the internals of DVCpro HD, why not pixel shifting. My point is the spec should be available and if there is a misperception educate the consumer. If Panasonic does half of the pixel shifting explanation that you do, a lot of the CDD questions, if not most questions would simply disappear. You are doing a better job explaining their product then they are. You don't see Canon or Nikon hiding the spec of their sensors. Foveon has 3.2 megapixels sensor and they call it 10 megapixels. At least on their website www.foveon.com they explain how they came up with 10 megapixels from 3.2 megapixels. Panasonic should do the same, no question ask. Panasonic don't tell me I don't need to know, it's for me to decide. Just make the damm information available.