Thanks Vic. The description of his methodology is good and gives one confidence in his conclusions, noting his caveats about unavoidable subjectivity. Resolution is only part of the story. His general conclusion that all the cameras were good HD value for the money is important for those who are easily upset about what is HD and what isn't. Everything is a trade-off. I read those reports some time ago and found them most useful.

While it's certainly avery thoroughly done test I have to wonder if existing tests are still enough to really 'test' the latest crop of cameras. There's so much happening not just with advanced encoding technology but also dynamic signal processing that static res charts I'm certain don't tell the whole story.
This is not just stuff for the geeks like me.
Under certain conditions the HDV cameras can produce simply amazingly good images (typically the same ones as you test with res charts) but also pretty sad results as well. I think it's a bit like some of the tests run on PCs, every manufacturer knows what the benchmark tests are, so they optimise their CPUs for those tests.
I did see a website some time ago where they were testing home cinema systems and I was amazed at how complex the tests were compared to what we subject cameras to.

Bob.

Well I personally was "smacked in the face" (and it hurt) with the so far un-necessary "need" for good glass.
On another post I mounted results of HDV vs SD shooting - where I shot the same exact "typical" talking head with two cameras.

I then did optical "zoom-ins" in Vegas to both HDV and SD. The SD footage was shot on a good SD camera, not converted in any way from the HDV original.

The SD was exactly what I expected - it basically was incapable of anything but minimal enlargement, before it began to pixilate and just look nasty.

What hurt was that the HDV allowed WONDERFUL enlargement, but for the first time, I could see the NEED for good glass. The lens of my particular camera couldn't keep up with the resolution of the HDV format. Now that's a FIRST!

I've come to rely on the crappy imaging of current standard definition NTSC television - well not "rely" but utitlize. As long as you don't need to "blow-up" your images (and a good cinematographer shouldn't need to very often) - you are solid!
But the current imagery and resolving power of HDV is pretty damn amazing!

And in fact, I recently had the opportunity to see my SD footage (put onto a DVD) screened in a movie theater, and IT looked amazingly good - better in fact, to my eye, than a 16mm print (converted to DVD) did.

So I can't wait to see my HDV (projected as HD DVD) stuff blown up. It was just a "shock" to have to once again consider lenses as part of the overall production methodology.

I'd say, if you're thinking of projected "feature" productiion, it would behoove you to get "good glass" on a camera with interchangeable lenses, where you can buy or rent expesive gear to put in front of your HDV sensor.

For my current documentary work, however, I'm sticking with the Z1. My new stuff looks beautiful.

v

The quality of the lens tells the story. There is a reason why cinema lenses can cost in excess of $30,000. Just for glass... If HDV cameras had better glass they would hurt the sales of the higher end HD cameras. I think this is part of the plan with manufacturers. Your final image can only be as good as the glass your shooting through.

Vic and Muhlstock make good points. It's funny that only Canon and JVC have interchangeable lenses on their sub 10k cameras, they don't have any other HD models to protect.

That's not quite true, Canon makes pretty much the same camera with and without interchangeable lenses, the XL-H1 and the XH-G1.
Adding a lens mount adds a lot of cost to a camera and you really cannot afford to cut corners otherwise you introduce problems like shifting back focus. Canon attempts to mitigate this problem by having a system that remembers the back focus setting for each lens but this can only be done with servoed lenses.
Once you add that you've pretty much got a captive market for your lenses, it's no longer just a matter of the physical mount, you need electrical interconnections to the lens.
Even so who would put a $30K lens on a $5K camera?
It really wouldn't make any sense, the imager and it's size introduces problems for the lens designer, limits resolution, latitiude and noise.
Bob.

It may very well be that the second coming in "affordable mid range HD" is in 1/2", thanks to Sony's CineAlta PDW-530 for example.

2/3" HD cine lenses are incredibly expensive, but it seems that Fuji and the other manufacturers are sharpening their pencils to come out with 1/2" lenses that have good to very good performance (very difficult) at a reasonable price (less than a 2/3" SD broadcast lens).

Of course if you go for 16mm primes the cost of lenses isn't so enormous. Zooms are very convenient for doco work but really an unnecessary expense when working under more controlled situations. The Switars standard on the Bolex are very good lenses and you'd be surprised how cheaply you can buy a Bolex fitted with 3 primes. The difficulty is that these aren't designed to work with the prism systems of 3CCD cameras, so until you get 2/3 inch CMOS sensors you really can't tap into that pool of excellent 16mm lenses.

Don't forget the inexpensive HVR-A1. I just redid some shots for a real-estate video shoot where all the previous footage had been shot with a Z1. I used my trusty little HVR-A1 and there was no lighting other than the ambient lighting in the house. The footage mixed perfectly and nobody could really tell what was shot with what. I would venture to say that for all but the lowest light cases, the extremely inexpensive HVR-A1 is in the same ballpark as the other HDV cams:

Severak low res versions of the project I'm talking about is here

Serena really makes a good point - there are tons of 16mm lenses just sitting around out there.

One of the reasons I got the JVC GY-HD100 is that JVC has come out with a re-imager for the camera that has a PL mount and will take 16mm lenses. It's called the HZ CA13U, it's all optical and it will retain the field of view and depth of field of 16mm lenses. I have yet to see any tests done with it so I'm a bit skeptical but JVC insists that this is what it does. One of these puppies and a set of decent 16mm primes could be very interesting indeed.

Google HZ CA13U for more info (not sure about linking to JVC stuff on a Sony site :)

yeah but...

16mm isn't 35mm and 16mm lenses while WAY better than 1/3" fare, still won't give you that 35mm "film" look -

as far as primes go, most features are shot with primes.

I hardly ever used zoom lenses (until the Cooke zooms - but that's another story...) when I was shooting. A prime and a dolly - now that's "old school".

v

It's called the HZ CA13U, it's all optical and it will retain the field of view and depth of field of 16mm lenses.

Lenses don't have a field of view and depth of field of their own...

You may never have used a 210mm wide angle lens, but I really enjoyed the improved field of view compared to the 300mm normal lens.

That was when used on an 8"x10" view camera of course.

If that same lens had been mounted to that JVC camera with a suitable adapter, would I have gotten that great "8x10 wide angle field of view that I would expect from a 210mm lens?"

No.

Why not, if it's all in the lens?

Because FOV and DOF are about focal length in relation to frame size. This means that for a given frame size, say 1/3" or 1/2" or 2/3" or 16mm or 35mm cine, a 10mm lens or a 50mm or a 100mm lens will always have the same FOV and DOF regardless of the frame size.

The only "cheat" isn't a cheat. By using a 35mm lens to create an image on a matte screen then shooting that screen from behind with a macro adapter using an HDV/SD/16mm/whatever camera, then you can get real 35mm FOV/DOF. That is simply thanks to the 100% 35mm primary optical path.

I agree with you Coursedesign - the lens does not create the DOF.

According to JVC this thing is a re-imager, just like a Redrock Micro or PS technic, except of course for 16mm lenses. The advantage is it is all optical with no spinning ground glass or moving parts:

"List price is $4395.00. Remember, this is not just an empty tube. There are some high-grade optics inside that allow the 16mm format lens to retain the same angle of view and DOF on our video cameras, as if it were on a film camera. This is not an easy task, thus the cost.

Regards,
__________________
Carl Hicks
JVC Professional Products Company "

I don't want to defend this thing because all I have seen is a prototype, I don't even know if it works but I sure am intrigued. I am very interested in the 35mm re-imagers but they seem to have a bunch of downsides, light loss, grain, keeping the ground glass clean, focus issues. If JVC can deliver on what they are talking about it would be very cool, not to mention I have access to a bunch of 16mm lenses :)

>>>16mm lenses while WAY better than 1/3" fare, still won't give you that 35mm "film" look <<<

Agree with all you say but I wasn't speaking about DOF, just resolution and contrast for cost. I think many people have forgotten that primes give better results because they're easier to get right in design and manufacture. Obviously there are very good zooms out there, but they do cost more than the cameras (as a rule).

Perhaps this calculator will help clear things up for you:
http://www.dofmaster.com/dofjs.html

Once again - depth of field is a fixed number based on:
Sensor (film) size
Focal length
Point of focus
Iris

The bigger the sensor, the longer the "normal" lens and the smaller the depth of field at any given iris setting. So you have way less depth of field on a "normal" lens in 35mm than on a 1/3" sensor in HDV at (for example) f2.8
That "selective focus" is one of the hallmarks of modern cinematography.

v


v

And again Victor you've ignored an important thing on that web site, Circle of Confusion.

The circle of confusion is different for a HDV camera to a SD camera, hence the DOF is different even with everything else the same.

EDIT:

In fact on that very web site if you scroll right down to the end of the camera listings under Viewpoint photography you can dial in different CoF values and see that....DOF changes!

Thanks for the link Vic - that is a cool calculator but I dont think I quite understand how it works. It's the first category that has me stumped:

"Camera, film format, or circle of confusion" I'm not sure what the JVC, or I guess any 1/3 inch chip camera would be. Also what would a 16mm camera be?

Hi Vic, this is tough argument to win because there is so much miss-understanding of how lens are designed by the video crowd. Most don't know that the last couple of elements correct a number of optical problems as well as creating the focal point. Another critical problem is the "surface" design of CCD, as well as the design of the prism( for 3CCDs). Hoperfully, the new CMOS designs(3/4 color layered) will replace the need for prisms and simplify lens.

Has anyone seen the V5 and V7 HDV video cameras from Sony sporting the new colorspace?

Too late at night to try to compute CoF for a HDV camera.

However what Victor is saying works quite well for film. Film has a physical resolution limit that's the same number of line per inch (or DPI), just like you ink jet printer. So a frame of 35mm film will have more resolution than 16mm and 8mm, assuming the same emulsion etc.

But this breaks down with video cameras, a PD 170, DSR 570 or a DSR 709 all have the same number of pixels even though the CCDs are bigger (that's roughly speaking), they're 1/3", 1/2" and 2/3" CCDs. Now throw a Z1 into the mix, it has the same size CCDs as the PD170 but way more pixels, so the Circle of Confusion is smaller for the Z1 than the PD170 IF you view the HD from the Z1 at native res on a big screen.

Why does the last bit matter?

Because the CoF is affected by many things, optics, imager resolution, display resolution and viewing distance. Again with film these things were keep pretty damn constant. Cinemas were designed to have the same angle of view etc, etc.

But with HDV we're using bigger screens and sitting closer to them, all this affects the CoF values and hence perceived DOF. I'm not saying it's a dramatic thing, I'd guess the Z1 has around 1/4 the DOF of a PD 170, which is still way, way short of 35mm. BUT it can be enough to make focussing way more critical.

Bob.

The resolution of film emulsions is stated as lines/mm at specified contrast, so the same emulsion performs the same in all cameras (small niggle). As farss says, the CoC depends on viewing conditions (size of screen, dist from screen, etc) and does involve a subjective judgement on when a point becomes a perceived circle (ie the point can be seen to be out of focus). In the film world there are accepted sizes of CoC for the various film formats and these became smaller as screen sizes grew. If such standards weren't agreed, it wouldn't be possible to publish useful DOF tables.

This article here explains the concept quite well. Lots of nasty maths however if you scroll down to Pixel Size Constraints you'll probably get an idea of where I'm coming from in regard to HDV cameras compared to SD cameras.

Bob.