Jon Fordham's HD10 complete review, Parts 1 & 2

[Les Dit]

This is getting funny. Anyone that says the HD10 is capturing less detail than *any* DV camera is using questionable methods.
I did some analysis of the HD10 footage, and the thing resolves quite a bit more. I had to blur the crap out of the HD footage to make it look like the Dv stuff.
I guess some posted stills illustrating this may be in order.
Shoot some frames of, say some power lines or a chain link fence and then tell me the DVX has the same 'sharpness'.
I use the words 'sharpness' detail' and resolution with the same intended meaning: seeing a non out of focus appearing image.

-Les

[Craig Jones]

Depth of field is a function of magnification and aperture only. It is not dependent on sensor size or focal length. f-stop is not aperture, it a normalized aperture used for exposure calculations only. DOF varies in the HD10 with aperture just like any other camera. You can measure it if you want to.

Therefore, saying that two cameras having identical sensor sizes have identical DOF is nonsense. Saying that 1/3 CCD's have identical DOF is also nonsense. Sensors have nothing to do with DOF. It's the optics. When comparing like sensor sizes, you can interchange f-stop with physical aperture, though, since focal length will be the same for matching perspectives.

If these two cameras have identical sensors sizes and shot identical scenes at identical lens focal lengths and f-stops, then the evidence proves that there is something wrong with either the observation of DOF or the optics in the DVX100. I would suspect the former unless the test didn't fix f-stop.

[Ken Hodson]

Yes, Jons comments in regard to resolution are very puzzling. It goes against everything I have seen and read so far.
Jon, I would say you have held up on posting some stills/clips of any of your comparisons long enough. We are all desperate to see the results of Heaths seperation from his HD10. We were all intrigued to see the HD10 vs. HD comparisons, but now that you compare its resolution unfavorably to the DVX100 (THE hd10'S real competition), we just have to see!!
KEN

[Jon Fordham]

One more time, for all those who aren't listening or refuse to actually read what I've been writing...

Les -

I never had any methods, nor did I ever intend for the two still frames that I emailed heath to ever become public debate.

A few posts back, these were my comments:

"Nothing scientific or planned out. Just some random shots for fun before I put these cameras head to head in a production situation."

and...

"Like I said, this was not a scientific test. Nor was it meant to be. This was simply me using a few minutes of my free time to satisfy my curiosity."

Twice, I've stated that there were no "methods".

Sharpness, resolution and detail are all different things. related, and similiar and often referred to in the same context, but different. The Sony HDW-F900's 1080 HD image is sharper than 35mm film. But 35mm film has the ability to achieve higher resolution. That higher resolution can result in revealing more subtle detail in the images than the sharp image of 1080 HD. At the same time, I've been on shoots where we had to have the art department windex the windows of a car four times to remove all dust and fingerprints just so we could shoot a shot through the car window. The DP bitched the whole time that we wouldn't have had that "problem" with film.

Ken -

A few posts back, these were my comments:

"There aren't any clips or stills currently posted from any of my side by sides. Primarily due to the legal issues still being worked out with the film "3 Days" where the HD10 was put next to the Varicam. And I have yet to shoot the film with the DVX100 that will serve as the HD10 vs DVX100 side by side. I mentioned this in my post, Jon Fordham's HD10 Review II, "Run and Gun". Please be patient. The footage is coming."

So, in case you missed it, in either the post on this thread or the other post in my above mentioned thread that explained that there are legal issues surrounding me posting the footage (read, I don't own the footage, I have to receive final written permission from the production company), I'll say it one more time. I'm sorry the footage hasn't been posted yet. Please be patient. It's coming, I promise.

I took some time to download the footage that Barry had posted and took a look at it. TO MY EYES, the footage seems to show similiar results as my informal and unscientific test. The HD10 is overly edge enhanced making it look sharper than it is. And the DVX100 with its manual control yields a much more pleasing image with similiar looking resolution.

Listen guys, if you're still reading this, I will say that I'm not trying to be a total asshole. But come on. Let's all chill out and relax. It's just a $3,000 handycam. And all the test and opinions and scientific bullshit is meaningless. If you can use it to do what you need it to do, then great. Do it. In fact, go do it and stop debating whether or not it can be done.

I'm just trying to tell you what I saw and what I've experienced. If your experiences differ then fine.

[Chris Hurd]

Here are two stills from Jon Fordham -- see:

http://www.dvinfo.net/jvc/media/HD10vsDVX100.jpg

and:

http://www.dvinfo.net/jvc/media/DVX100vsHD10.jpg

There is some serious misinformation in this thread concerning DoF (depth-of-field). CCD size as well as focal length do have an impact on DOF, it is not simply a function of magnification and focal length. For more information about how CCD size and focal length can impact DoF, see our Ultimate Depth-of-Field Skinny, part of our Optical Science series by instructor Jeff Donald.

Also, the DVX100 is *not* the HD10's "real competition." These are two entirely different camcorders, intended for entirely different markets. I'm aware that the HD10 has the interest of some filmmakers, but JVC produced it solely for the high-end consumers of the D-VHS marketplace -- it was never meant for anything else. The HD10 represents an experimental foray into next year's forthcoming HDV format; the DVX100 is a compact SD camcorder with a 24p feature from Panasonic's professional camera division. They both take tapes and make pictures; beyond that I would resist any comparison, which would be neither fair nor appropriate. Hope this helps,

[Ken Hodson]

I myself along with many others are trying to make a decision in regards to this cam. $3000 may be nothing to you, but its $5000 +tax in my country, + matte box + batteries + filters ect... . Or put in another way a lot of money for someone who doesn't get paid to use ultra high end cameras all day.

We all value your input as you are possibly the most qualified person who has ever commented on this cam at any length. Some of the things you have said, the DVX comparison for example, are 180 degrees from some others. We all here have a high degree of enthusiasm for this cam, so we tend to question everything. Usually with no disrespect intended.
Ken

[Craig Jones]

I don't recall saying small sensors were capable of superior overall performance. I simply said that sensor size did not determine DOF. This fact is not controversial but it is often misunderstood by non-technical types.

A very good discussion of lenses, including DOF, can be found here. You will see in the DOF equations that sensor size of a format is never mentioned. This is intuitively obvious, of course, and fully supports my claim.

Nevertheless, the claim that was made was that both cameras in question had identical sensor sizes, yet one exhibited significantly different DOF than the other. If that is true, then there are only three explanations: (1) the observer is mistaken, (2) f-stop and focal length, and therefore physical aperture, were not held constant, or (3) there is something wrong with the DVX100 lens. The third option is unlikely. More likely it was a testing error or a consequence of the substantial difference in video format. Generally speaking, the higher resolution format would be expected to give less perceived DOF, the opposite of what was reported.

On a separate issue, I found this quote unusual:

Bottom line is, the DVX100's full 480P vertical resolution is comparable to the HD10's HDV 720P resolution.

Sure they're comparable but the comparison is unfavorable! In order to accept that the DVX100 provides similar or better vertical resolution (as was implied) you would have to accept that the JVC is a tremendous underperformer for 720p (roughly 1/2 to 2/3 the potential of the format). I doubt that seriously. It would be pretty bad if the JVC couldn't resolve better than 360x240 lines in 1280x720 pixels.

[Chris Hurd]

Gentlemen,

If I have to do any more editing to this thread, I'll simply lock it and be done with it. Please retain the highest degree of mutual respect and civility and courteousness at all times. If you see something rude, don't bother responding to it because it won't last very long (flames are always edited out of dvinfo.net).

For Craig:

Just because that photo.net article doesn't mention sensor size, doesn't discount the fact that sensor size *does indeed* impact DoF. Once again, please see Jeff Donald's Ultimate Depth-of-Field Skinny. The man has been teaching optical science for many years and his article spells out very well the factors which determine or alter DoF (including CCD size as well as focal length). Or, see this page on photo.net: "larger sensors have less depth-of-field." Hope this helps,

[Frederic Lumiere]

Jon,

Did you turn the TELE MACRO setting to ON (default OFF) on the HD10U before shooting the samples we got from Chris?

[Craig Jones]

Originally posted by Chris Hurd :

There is some serious misinformation in this thread concerning DoF (depth-of-field). CCD size as well as focal length do have an impact on DOF, it is not simply a function of magnification and focal length. For more information about how CCD size and focal length can impact DoF, see our Ultimate Depth-of-Field Skinny, part of our Optical Science series by instructor Jeff Donald.

Sorry, Chris, but you are incorrect. I read through Jeff's article briefly and it looks OK, but I prefer this article since it is far more complete. In any event, focal length only effects magnification (target size in Jeff's article) and the proportion of front-to-back DOF. Magnification is already mentioned, and front versus back doesn't matter in the discussion of total DOF. Jeff's article largely supports my claims.

The fact is that "target size" changes somewhat between formats for identical perspectives, but at higher magnifications it is offset by the lowered acceptable diffraction tolerance. Larger sensors are actually capable of better DOF in macro than smaller ones. I suspect people here don't concern thenselves with macro too much, but macro DOF issues are more difficult (i.e. intractable) than wide or normal perspectives.

The real concern, and the one Jeff makes an example of, is trying to restrict DOF. Prosumer devices have integral lenses that place limits on the range of DOF they can achieve. This is a practical concern and does not reflect what any sensor is capable of. It's a mistake to extrapolate an experience with one or several specific cameras into a generalization of any camera with that sensor size, yet that's just what Jon has done.

Larger sensors make it easier to achieve limited DOF not because the sensors have inherent DOF qualities, but because they allow the use of lenses that are more appropriate for that kind of shooting. This is a practical issue, not a technical one.

[Les Dit]

With regards to the two stills:
First of all, I would have expected the JVC frame to be 1280 across,not 720.
To compare detail of a lower spacial resolution image with a higher one, the lower one is typically resized up to the bigger ones size. It's safer to upsize, because you are not crushing out possible detail that way.
The older clips posted were done the right way, the DXV footage was sized up to the 1280 size of the JVC.
Naturally, the DV diehards hate that type of test, as DV looks totally out of focus that way. That's life. When up do a film out, it all gets sized up to academy or whatever, so no biggie.

The DOF is very different, OK, got that.

The HD10 image still looks like it shows more detail in the logo on the green bottle, and the sunglasses case. Just a tad. Hard to tell at 'web cam' resolution (720), it's just too mosaic-ed in the details.

As far as the picture rights and posting stuff, too bad you didn't have a few seconds to swing the camera 'off set' and at a road case or grip truck, or a porta potty, to eliminate the rights issue. I understand the situation was probably pretty stressful anyhow, you were probably pushing peoples patience with the 'toy' camera there anyway.

If anyone is in LA and would like to shoot some tests, I have a HD10 to test with. Heck, I can shoot some 35mm movie film too, which I'll scan as 2K Cineon format, just to see what that looks like, of the exact same scene. Some of the new Kodak stock as pretty much no grain, at 2K res. ( 2048 x 1556 pixels )
-Les

[Jeff Donald]

The size of the sensor has an effect on Circle of Confusion (CoC). CoC is a factor in DOF. Magnification is a part of DOF as DOF is determined by distance to subject and the focal length of the lens.

Craig's analysis of the observed differences is probably correct. We are dealing with a very subjective media. The five factors that determine DOF include the subjective nature of personal observation (see factor E in the article). The critical analysis that many of you are looking for need to be observed under a more controlled situation. Don't make inference's from personal observations that are not your own. It's an argument that is never won.

[Craig Jones]

<<<-- Originally posted by Chris Hurd :
For Craig:

Just because that photo.net article doesn't mention sensor size, doesn't discount the fact that sensor size *does indeed* impact DoF. Once again, please see Jeff Donald's Ultimate Depth-of-Field Skinny. The man has been teaching optical science for many years and his article spells out very well the factors which determine or alter DoF (including CCD size as well as focal length). Or, see this page on photo.net: "larger sensors have less depth-of-field." Hope this helps, -->>>

The photo.net article is definitive as far as I know. I'd be curious to know if Jeff thinks it is incorrect. At first glance, I'd didn't see anything in Jeff's article inconsistent with it, and I'd be curious to know if Jeff disagrees with it.

The Kevin Borden reference just demonstrates than many people believe sensor sizes have inherent DOF field properties. They don't. They just record the image rendered on them. The size does determine the required magnification for a given perspective, but that is all.

P.S.

Twice now I've been caught composing while someone else was posting. This time by Jeff.

I think there is some difficulty with terminology between here and the still folks (although the optics work the same). Sometimes terminology alone can cause big arguments.

[Barry Green]

Sensor size has nothing to do with DOF. Craig's right.

Think about it: the lens resolves the image and then projects it onto a surface: it doesn't know nor care what that surface is. It could be a 1/3" CCD, or a 2/3" CCD, or a piece of film, or a piece of ground glass, or a blank wall, it doesn't matter. All DOF is resolved and settled within the lens. The sensor just records what the lens projects.

The CoC is related to the amount of magnification an image will undergo. The image (as projected by the lens) will be identical either way. The circle of confusion is just a guideline to adhere to when thinking about how much magnification you're going to put the image under.

Here's another way to look at it: put a Nikon lens on a 35mm movie camera, and shoot a frame of film with it. Then put that same lens, in the exact same position, on a Bealieu Super 8mm camera with Nikon adapter. Shoot the same scene. Then take the 35mm film frame, and cut out the central 6mm x 4mm area, and compare it to the Super8 frame. THEY WILL BE IDENTICAL. Absolutely grain-for-grain identical. Blow them up to the same size, and they will look identical.

Finally, regarding Jon's two shots: I have no doubt whatsoever that the JVC had a smaller aperture than the DVX did. That is, quite simply, the only rational explanation for the differences observed in these shots. Especially when you consider that the DVX would have needed to be zoomed out *wider* than the JVC in order to capture the same framing and field of view (due to the 4:3 DVX vs. the 16:9 JVC).

[Jeff Donald]

Sensor size has nothing to do with DOF.

Until you print it, project it or transmit it to a TV. Then differences in CoC, as a variable of film size, chip size etc., will have an effect on DOF.

Since this is a video forum and people's work will be viewed on a TV or projection screen (magnified), it is misleading to ignore CoC and it's impact.