Give JVC and their Lens a break - It passed a tough test

[Graeme Nattress]

Michael, would it be possible for your contact to be able to publish the lens specifications for us, so that we can see them here? All I'm going off is actual HD100 footage on my system, (and my visit to the JVC exhibit at NAB, which I don't think is the best environment for judging something like this).

What would help us all is to get some real measurements of what resolution a lens needs to be, taking into account the size of a CCD. All I know is that the smaller the CCD for a given resolution, the higher the resolution needs to be. So, indeed, the lens MTF could be "HD", but that might not take into account the extra sharpness needed for the smaller CCD. I don't know, but that's why I'm asking for information.

As far as I understand, making a good 35mm lens is hard, making one for a 2/3" CCD HD camera is very hard, and doing so for a 1/3"CCD HD camera is very, very hard indeed, and not a problem I expect to see solved for such a cheap camera / lens package as the HD100.

To add to this, I also doubt that Panasonic can make their HVX200 lens sharp enough also, and we could be very well in a similar "lens limiting" situation there, especially if they try to get 1080p out using pixel shift, which if I understand the maths behind it correctly, requires a high enough resolution getting through to the CCD that they alias, and I just don't think this is going to happen. Look at the Z1, it's also lens limited. They try to pixelshift, but it doesn't work as the lens won't pass enough detail to get the CCD to alias. All 1/3" HD cameras are in the same boat here - lens limited due to too small CCD.

I guess we'd all better start saving.....

Graeme

Originally Posted by Michael Pappas

Hello Graeme.

At JVC, I have made contact with a very nice gentlemen named Ken Freed. Ken has been answering questions when I need them. I called him today. The MTF for the Fujinon lens is in the HD MTF specs.

I have worked with 60fps material 15 years ago. At Showscan, a 65mm 60fps format, if you don't have it right, it will be obvious very quickly at 60P.

The NAB display of 4:2:2 uncompressed at 60P on CRT was the best to see if a lens is going to fail. The next step would be on an optical bench.

That's my point, this lens achieves well beyond what it critics say.

The bi-product of JVC's NAB display of their new HD camera was the lens was getting it's butt put on the line with regards to going out to the CRT's like that. It recieved amazing feedback about the live image on CRT. That says a lot about the removable lens.

michael pappas

[Steve Gibby]

Graeme,

"Well, "HD quality lens" is a marketing term, not a scientific one, and is essentially meaningless. When physics says you need a sharper lens for HD on a 2/3" chip than you do for 35mm film, and that you will therefore need a sharper lens still for a 1/3" chip HD camera (going to 720p will help a little here) I severely doubt any SD lens (yes, I know, an essentially meaningless term, but you get my meaning) will be of any use on a 1/3" HD camera."

You still didn't address the issue I quoted. Fujinon is one of the most reputable lens manufacturers in the business. They got that way by creating quality products that satisfy the needs of the production industry, as did Canon, Zeiss, etc. They didn't get to that position by lying to industry users questioning them about their products. I use lenses from several manufacturers. What I get from your posts, including your quote above, is that you feel Fujinon is lying to the production industry about the "HD" quality of the new T16x5.5 lens, and that JVC is lying about their position that the three Fujinon 1/2' SD lenses and two Canon 1/2" SD lenses on the HD100 System Chart will adequately resolve HD images from the HD100. Would a company with the reputation of Fujinon risk their good reputation by creating a substandard lens for the HD100 - a lens that would not resolve HD images when they claim it will? I don't think so. As JVC is trying to carve out a market position in the indie and TV market with the HD100, can they afford to claim a new lens for the HD100, and existing 1'2" SD lenses shown on their System Chart will resolve HD images, if that isn't the case? Is Graeme right and Fujinon/JVC wrong?

"Yes, I did go to NAB, did see the footage, and have indeed, which I suspect nobody here has, viewed the footage in their own system (http://www.lafcpug.org/reviews/review_decklink.html), and indeed did so before NAB, this thread and associated articles / comments.

I find your words quite hurtful that you seem to invalidate my opinion while bolstering your own by appealing to the awards you and others have wone and their industry status. AFAIK, you don't get tested on your eyesight and ability to determine picture quality to SMPTE standards before they allow you to win and Emmy or hold a job in Network television."

Wait a minute! You are the one who sought to invalidate the HD100 uncompressed output opinions of other prior posters on this thread (5/2 @ 1:41pm) by essentially trying to "school" us on the difference in how footage looks on CRTs vs. LCDs. The essence of what you posted was that we really didn't know what we were looking at and how it should look on various monitors. Before you go saying something like that you better know the background of who you are accusing of not knowing how HD footage should look on monitors. My quoting of my background and others backgrounds was simply a refutation of your position, and to let you know that YES we do know what we're looking at because we've done a lot of looking for a lot of years! Your recent statement about eyesight testing before someone wins an Emmy or works for a networks is simply ludicrous. If someone spends several decades producing, directing, editing, and shooting network programming, and winning scores of national awards at it, they don't how to judge the quality of footage or a live output to a monitor? What kind of a ridiculous position is that? I'm online right now. on your web site. I've looked through every page for a biography on you - an indication of a background in network television or indication of having won any television production awards. I also don't see an optical engineer background for you listed on your web site. What gave you the credentials to start this whole exchange by telling prior posters on this thread that we don't know how to judge footage? What I do see on your web site is lots of pictures of Panasonic HVX200 cameras and Final Cut Pro plugins and articles. Where's the HD100 pictures on your web site? Based on the info on your own web site it sure appears that your business is based around Panasonic cameras and Final Cut Pro. Do you have an agenda in debunking the lens resolution capabilities of the JVC HD100 on this forum?

"No, it doesn't hold much validity at all, because physics says different. And as you quote me saying "I'm not sure about this, but..." I must add that I have enough knowledge about the physics of optics to know that for a given CCD resolution, the smaller the CCD, the higher the resolution of the lens needs to be. That is why I comment that it might not even be possible (or at least economically possible) to make an HD lens sharp enough for a 1/3" chip CCD camera, and even if it were, I doubt it would sell for what the stock HD100 lens sells for. I discount the opinion of someone who has seen the camera footage on show floor conditions on an CRT, when I have the footage, have looked at it closely on a higher resolution display. And indeed, my opinion is that it still looks like the best HDV footage I've seen, but that it's softer than Varicam footage I have, and is most likely, for the phyical reasons I give above, lens limited."

Lets analyze this: You're not an optical engineer, and yet you don't believe the claims of a Fujinon optical engineer? Fujinon and JVC have both claimed that the T16x5.5 lens was custom engineered for the HD100 and will resolve HD images - but you claim they are both wrong or are lying? Fujinon and JVC have both claimed that the better 1/2" SD lenses, when used on the HD100, will resolve HD images - but you claim they are wrong or lying? You claim you have some original footage from the HD100 at full resolution, and have tested it, which totally sidesteps the original issue brought up on this thread - that the 16x lens resolved excellent HD images on the uncompressed 720p60 output at the NAB booth, and that if it wasn't an "HD quality" lens, it wouldn't have been able to resolve those images. Your strategy has been to attack the capability of others who viewed the footage as being unqualified to judge the CRT output, and to try to establish LCDs as better than CRTs. If you have the "footage", hasn't that footage been subject to MPEG2TS compression, with it's inherent compression artifacts? How can that footage then be judged as demonstrating the top-performance of the 16x lens? Isn't the live uncompressed output we all witnessed a better indication of the resolving capabilities of the lens than your compressed footage? None of the rest of us have seen the footage you say you have - but many of us have seen the uncompressed output and it looked exceptional. Again you mention a Panasonic product (Varicam). FYI - I love the Varicam and use it regularly. It's an exceptional tool for acquiring great images. I also love the DVX100a and use it regularly. But I also love the JVC DV500 and DV5100, and use them regularly. For that matter, I also love the Sony F900, and use it regularly. That said, I find it interesting that you have the HVX pictures on your web site, no HDV camera pictures, and now refer to an $85K (with lens) Panasonic product to debunk the quality of a $6k JVC product. Doesn't you're using the Varicam in the same sentence as footage from the HD100, and only saying that the HD100 footage is "softer", tell us readers that the HD100 must be remarkable, and the 16x lens very capable, if you're not making a more harsh judgement of the difference?

"I'm trying hard not to take your comments wrong, but I'm fairly well known, and also fairly well know for having very balanced, reasoned opionions, that might not always be the "party line", but at least have a lot of logic and reason behind them. I did not shoot off a post saying that everyone was wrong, but that because I have had the benefit of viewing the footage under controlled conditions (ie better access to available evidence), and to do some comparitive analysis, at least of the footage, I'd be listened to a bit more respectfully.

I think my key point is that making an affordable HD lens for 1/3" cameras is going to be very, very hard indeed. And picture quality-wise, you may be much further ahead with a 2/3" HD camera with a cheap lens, than a 1/3" HD camera with an expensive one, and I'm going to be very interested to see how a 2/3" SD camera (with decent lens) compares with any 1/3" camera in terms of real, actual, measureable detail / definition, as I suspect they might be much closer than many people think."


Again, if you've viewed "footage", how does that equate to us viewing live uncompressed output on good monitors? Isn't the live uncompressed output a better indication of the quality of the 16x lens than "footage" you have that has been subjected to MPEG2TS compression artifacts? When the HD100 and 16x lens ships and can be tested at length, in multiple scenarios, isn't that a better way to judge the lenses ability to resolve HD images? So why debunk the capabilities of the lens, call Fujinon and JVC liars before the facts are in and can be analyzed? It was your post of 5/2 @ 1:41p that triggered this whole exchange.

Graeme, this has been an interesting exchange of ideas and positions. FYI - I have used, and do regularly use products from almost all manufacturers: Sony, JVC, Panasonic, Ikagami, Apple, Fujinon, Canon, Media 100, Avid, and on and on. Among other editing applications, I use Final Cut Pro. I simply match the paintbrush to what I'm going to paint.

I'm not interested in clogging this thread by continuing this tennis match online. If you want to debate the issues futher, simply e-mail me at steve@cut4.tv and we can continue offline.

[Barry Green]

Whether someone claims that a lens is HD, and whether it actually is HD, is something that can be determined by simple testing. Sony refers to the Zeiss T* lens on the Z1 as an HD lens, but look on the FX1/Z1 board here for a thread running right now that pretty much acknowledges the lens is demonstrating chromatic aberrations.

I thought the live footage at the booth looked great. But, let's be fair, it was on small CRTs that were overhead. It's not like we were critically analyzing a 42" display, we saw small CRTs that were six or eight feet away, right?

I was quite impressed with what I saw, on the CRTs. I thought what was on the plasma was disappointing, but the CRTs looked good. But that's a first impression, at a trade show. Before declaring that the lens and image system are good enough, I would want to test it thoroughly. I think anyone would.

Graeme knows more about images than I think any of us here ever will. Images are his business. He may not be an optical engineer, or a television producer, but he's freaking brilliant and he knows what he's talking about.

Attempts to discredit someone publicly are poor form, as is bashing a product that's not even on the market yet. The simple fact of the matter is: the JVC lens will either hold up under testing, or it won't, regardless of what JVC's engineers or Fujinon's engineers or Graeme or Steve or I say about it. It either will or it won't. And that's a determination that cannot be made until we get our hands on it and put it through some testing.

[Graeme Nattress]

As far as I can tell (looking back through my R&D notes from working with Panavision) the 2/3" CCD on the HDCAM F900 (I think it was that at the time) has pixels of 5x5 microns, and a corresponding resolution of 100lp/mm. Does anyone have the CCD dimensions of the JVC HD100 1/3" camera? I took a guess based upon the above dimensions of a 2/3" 16:9 sensor and figured out that the pixels on the JVC are going to be around 3.7 microns square. That equates to a resolution of 133lp/mm.

Now, Fujinon don't have any MTF specs for the lenses, and I don't see any on the JVC site, but I did see the specs for the Zeiss DigiPrimes:

rated at 100lp/mm, but with a 90%mtf at 56lpmm. Usually, for the rated resolution they take the 5% mtf point, so I'm guessing that it has no resolution beyond 100lpmm.

So, for a 1/3" 720p camera, we need a lens rated at 133lp/mm, which to me means that..... Well, I've said my piece already, but I think anyone with a little mathematics can draw their own connclusions, but simply put, there's going to be quite a few figures after the $ sign.

For what it's worth, again at a rough guess, the Z1 CCD is about 100lp/mm horizontal but 200lp/mm vertical (again if anyone has exact CCD measurements, I can refine these figures).

Given this, I don't hold much hope of any 1/3" CCD camera not being lens limited for the forseable future, no matter whether it's Sony, Panasonic, or JVC, and whether it's lens comes from Zeiss, Fuginon or Leica.

Graeme

[Steve Gibby]

Barry-

After reading your last post I find I agree with most of what you said. If you really analyze it we've circumlocuted around to the original bottom line: without an actual hands-on test of the HD100 with a T16x5.5 lens attached, everything anyone says is conjecture. That said, the only thing experienced professionals in the TV/video industry can use to draw preliminary conclusions is the camera and lens manufacturer tech sheets and what we've actually seen first-person thus far. My purpose in even bothering to post on this board, or any other, is to share impressions and information with other interested parties, in essence, to network. I'm not on a mission to convert anyone to my line of thinking. I'm also not interested in being told that something didn't look as good as my eyes observed it. When the camera and lens are released, and thoroughly tested, we'll have definitive answers. Until then, everything is merely opinion. And opinions are like noses - every one has one.

As for someone having more knowledge on images than me, I don't think you are in a position to make that judgement. After contributing as a producer, director, editor, and cameraman to over 700 national television programs, SD and HD, that aired on 12 different networks, I have a pretty good handle on technology and images...

[Aaron Koolen]

Graeme, that was English right? ;) My head hurts now...Where can I go to learn something so I can, at least superficially, understand what you just said?

Aaron

[Douglas Spotted Eagle]

Steve, Graeme, Barry,
I don't think anyone of you mean to be escalating a heated discourse, but it's somewhat starting to sound that way.
But as you've each pointed out one way or another, it's ALL conjecture for everyone at this point, until you've got the glass and body in hand, and can make comments on what you *are* seeing, not what you've read, what you saw, think you saw, or might have seen at a tradeshow in less than ideal, or perceptually the most ideal situation.
It seems like this discussion can't really go anywhere else for a few months, right?

[Steve Gibby]

Hey Spot-

Well said. I agree and voiced the same in my last post. It's kicking a dead horse right now- nothing will be solved.

Let's move on...

[Michael Pappas]

Hi Steve, Can I have a try at this.

It's actually like kicking an unborn horse.

Since, it's still in the creation stage.....

Originally Posted by Steve Gibby

Hey Spot-

Well said. I agree and voiced the same in my last post. It's kicking a dead horse right now- nothing will be solved.

Let's move on...

[Brian Wells]

Originally Posted by Barry Green

I thought the live footage at the booth looked great. But, let's be fair, it was on small CRTs that were overhead. It's not like we were critically analyzing a 42" display, we saw small CRTs that were six or eight feet away, right?

An expert on lighting design I am not, but it doesn't take a genious to notice the way these sets are lit at NAB: None of them had any blacks and everything was lit to textbook perfection and limited to within a couple of stops dynamic range. Every camera looked good at NAB.

These companies hire serious Set and Lighting Designers to make the cameras shine. Just an observation from a middle class film worker..

Brian

[Steve Mullen]

Originally Posted by Michael Pappas

Steve do you have any info on your site about short GOP and also on mpeg/HDV

I can't find a definition. Is "long" anything longer tham a 1 or 2 frame GOP? Is "long" 12-15 frames?

I cnider 6 frames "short" not "long" but I could be wrong since all the folks (like Apple) never define LONG?

The MTF definition of what is an HD lens is based upon 1920x1080 resolution. This is defined as 800 TVLines. Turns-out, 720p definition is at 530 TVLines. So, when we say a lens is HD we must qualify which system we will be using. A true 1920x10 lens may be very expense -- a 1280x720 HD lens could be dramatically cheaper!

An SD lens is only 400 TVLines -- so it really doesn't take much more quality to make an 720HD lens!

[Steven White]

trying to compare the glass between these two cams is a completely backward way of researching a major purchase decision.

This gave me a good chuckle!

The glass is the first thing the light sees - if the bottleneck in your system is here, none of the rest of the stuff Chris mentioned: format, workflow, etc., would matter in the slightest :)

EDIT: By the way, I just looked over Graeme's posts. His math looks good to me. Worth reading.

-Steve

[Luis Caffesse]

This gave me a good chuckle!

The glass is the first thing the light sees - if the bottleneck in your system is here, none of the rest of the stuff Chris mentioned: format, workflow, etc., would matter in the slightest :)


Steve, you may be right if all things are equal, but I think you're taking Chris's point out of context a bit.

He said:
Choose your format first... trying to compare the glass between these two cams is a completely backward way of researching a major purchase decision.

And that is something I would completely agree with.

The JVC shoots HDV, the Panasonic will shoot DVCProHD.
People don't go out and pick a camera based on the lens, they choose generally based on the format they want...then from there they choose which camera specifically fits their needs.

An HDV camera and a DVCProHD camera are completely different animals.

If we were to take a camera based solely on the best lens first then I suppose that the JVC would win out, seeing as it's the lowest priced HD camera with a removable lens...meaning you could get the best lens on it.
Of course, then you're left with MPG2 compression, no 1080 HD, no 720P60, compressed audio, and a 4:2:0 colorspace.

My point is just that obviously you are right, the lens does matter, but you can't just disregard the rest of the camera, much like you are saying that the lens can't be completely disregarded.

Also, I would add that I do think it is a moot point to compare the glass between these two cameras seeing as no one can get hands on with either of these cameras/lenses and do any side by side comparisons until the end of the year.

[Steven White]

I didn't taking his post out of context at all - I read the whole thing. My amusement, and hence laughter, came entirely from reading the sentence independently... and as you can plainly see with it quoted on its own - it's a bloody hilarious way to look at the world.

The notion of starting at any one point in the camera and saying it's "backwards" place to begin struck me as very odd. Obviously you have to examine the whole package. But DVCPRO-HD coming out of a pinhole wouldn't garner much excitement now, would it?

In perfect context, I challenge both yourself to write the following essay:
"Garbage in, garbage out: Why is starting with the input is a backwards approach."

:: grins, winks, etc ::

-Steve

[Luis Caffesse]

My amusement, and hence laughter, came entirely from reading the sentence independently... and as you can plainly see with it quoted on its own - it's a bloody hilarious way to look at the world.

Yes, by itself it is funny.
My point was that in the context of the conversation I think it's an apt point to make. These cameras are very different, with very different strengths and weaknesses. There are larger differences between them than the differences between the lenses (the lenses which we haven't been able to test yet).

The notion of starting at any one point in the camera and saying it's "backwards" place to begin struck me as very odd. Obviously you have to examine the whole package. But DVCPRO-HD coming out of a pinhole wouldn't garner much excitement now, would it?

Of course not, but I also don't think the differences between these two lenses will be great enough to ignore differences between the formats, recording times, etc.

I think we're on the same page here.... we both agree that you have to examine the whole package. In this case though, considering the lens on the HVX hasn't even been seen, and the lens on the JVC hasn't really been tested...and neither camera has been released, I would agree with Chris that comparing the lenses seems a bit 'backwards.'

Then again, comparing anything on these cams right now seems a bit ridiculous to me.
:)