If you go all the way to the bottom, the charts look like the DVX or the XL2 has the best stuff. However, IMHO, the FX1 seems to have the best looking frame grabs from the scenes that were shot. (granted these were shot without setting up for the best look)
While it's a review found on another page, it should be discussed there. It's not fair to the authors to use this as a forum for discussion.
However, one thing I want to know is how they matched size, aspect, and exact angles of the higher resolution and wide FX 1 vs the other two. This strikes me as a camera shot in DV mode, not HDV mode. I ain't buyin' it.
Either the FX1 was improperly squished into 4:3 mode, or the DVX was not in 16:9 mode and the same frame attributes were selected. I haven't compared it with the XL2, but I'm familiar enough with the Z1 and the DVX from two days of shooting them, that there is no way these images could be that close in quality if you were shooting HDV in one, and widescreen in the other.
LOL---there is something here that doesn't meet the eye---check out the highlights on the XL2 shot---no way those cameras were set up similiarly. As I have said before, I don't want to see any of these whacky tets---I will go down and make up my own mind.
I'm fast learning a thing or two in this game game:
1) When YOU know you're on a winner go with it.
2) Keep your mouth shut about it.
I came within a mouse click of buying a DVX100A and I still think it's a great camera and if I had enough work I'd buy one or its liitle brother the DVC30.
Having seen FX1 footage I KNOW the only way I'd get better 16:9 SD images is if I spent 10x more. So why do we all feel the need to shout this from the rooftops? Who are we helping? Sony sell cameras?
Those of us in the know have got a distinct competitive advantage, make the most of it and let the competition work out for themselves why they're loosing jobs to you.
All that aside a lot of the guys who frequent DVXuser.com are pretty seriously into their craft, the knowledge they've acquired means they can turn out better product with any camera. However I'm certain once we get to the same in depth understanding of the HDV cameras then we'll really start to see the difference, remember though, your audience comes to watch a story, not count pixels on the screen.
LOL, I guess it wasn't that great a link after all! Makes me think: I've got an old half megapixel still camera that blows all those newfangled megapixel still cameras out of the water... and I've got the test pictures to prove it ;-)
WOOO man, you guys are just begging to conjure up the mystery poster HDV and Barry Green debate. ; )
I saw the tests and frame grabs , and while I think Mr. Green is a good guy and respect alot of what he has to say, there's still a DVX slant to the article. Even though he does point out the many good aspects of the other cams.
As he states in the article,to me it does boil down to what you want to do with the cam, and how you use it.
Nor did I think he was trying to put down the other two cams.
I know what I want, and the article did not change my opinion.
"It's not fair to the authors to use this as a forum for discussion. " Where did this come from???
Once the authors published their views, everyone, at least in the US, can openly discuss their article, no approval required. The question is, did they get all three manufacturers to sign off for the article and the testing format.....
I would be pleased to discuss any aspect of the testing, the setups, the procedures and the methodology used.
These are the tests we conducted, and our findings. Those are the images that the cameras delivered. The only processing on any images, at all, was to split-screen them and output them to DVD. The HDV footage was downrezzed using Vegas 5.0, using "best" quality.
If anyone wants to conduct their own tests, I could re-capture bits of the HDV footage (right now it's all captured in several-hundred-megabyte clips) and make them available to anyone who wants to play with 'em, next time I have an FX1 on hand.
However, I do not appreciate accusations of "the test results were bogus". You can say that you don't think we knew what we were doing, or whatever, but to imply lying is... well, heck, it's rude.
The cameras were not set up to extract the "best results" under any condition, they were set to middle/default settings across the board. Our intention was to find out what was really true, rather than what "should be". And that is what we found. Anyone is free to conduct their own test and document their methodology and if they find different results, I'm sure we'd all love to see them.
As for aspect ratio, obviously you can't compare 4:3 against 16:9 without one suffering. Since the XL2 and FX1 are both 16:9-native, we conducted 90% of the tests in 16:9. The DVX was usually in electronic squeeze mode, we used the anamorphic adapter for about 10% of the testing. And we did about 10% of the testing in 4:3 mode.
Since conducting this test, and witnessing the ensuing discussion, it's become clear to me that the only test people will likely be satisfied with is a three-camera side-by-side showdown, with each camera piloted by its most vocal advocate -- ideally I'd like to see Spot on the Z1, Chris Hurd on the XL2, and me on the DVX. Unfortunately it's proven difficult to get us all in the same place at the same time with these cameras, but I very much look forward to the opportunity.
If I'm wrong in my observations, I'd like to know it. But I'd like it to be demonstrated by scientific objective proof, such as we attempted to employ. Hopefully someday soon we'll get a chance to conduct such a test that would deliver results that would be absolutely above any claims of "bogus results".
If you ever compare cameras, people yell at you. I don't know why, unless it's the fear that their purchases are the next betamax vcrs.
I have no experience with HDV yet, so I'm not taking any stand. But I thankyou for doing your tests!
Barry, you can not call your written opinion a demonstrated scientific objective proof. The simple fact that you did not properly describe the pixel density on the ccds at 16x9and at 4x3 is one serious flaw. The fact you used a moving light displays at night on NY is a major problem, where are the light meter readings to demonstrate a stable lighting scene for each camera or proof the cameras were syn. . ETC.......
Its nice that you gave your opinion, but its not a "proof" by a long shot. The only thing I can agree with you is the 100A is a nice camera for the dollar in 2004 for IDF makers who can not afford good glass or a good 16x9 camera.
Obviously a written opinion is not proof. What I offered was the pictures generated by the cameras, without biasing any of them towards any result, just let the chips fall where they may. By "proof" I'm referring to actually having put them to the test, rather than theorize that "this one should be better because it has more pixels" or "that one should be better because it's HD" or whatever.
The pixel density of the CCD's is an irrelevant statistic. What matters is how the footage looks. If one wanted to compare cameras based on pixel densities, you wouldn't need to even have the cameras, you can just get some tech papers about them and compare notes. But that doesn't tell you how the footage looks. And that's all that matters, isn't it? And what would pixel density prove? Is it taken as axiomatic that more = better? Because through direct observation, more does not always equal better. Many people bought XL1's thinking that it had great video quality which they preferred over the VX1000, yet it had a much lower pixel count than the VX1000. Denser pixels yields lower latitude and worse low light performance. The XL2 has denser pixels than the DVX, and is about a stop slower. The FX1 has denser pixels than the XL2, and is about a stop slower. Those are the kinds of things you find out when you put them in the ring side by side, which may not be apparent if you just read the specs.
If one's concerned about sharpness in 16x9 mode, you can look at the Grand Central Terminal picture we put in there. The XL2 is noticeably sharper than the DVX in squeeze mode is. You can see that in the picture, regardless of how many pixels one camera has or the other one does; just like you can see overall sensitivity being reflected by the f-stop and brightness level each camera exhibits.
As far as the "moving lights", I'm not sure what that is referencing -- are you talking about the f/5.6 and f/1.6 picture? We extracted an identical frame in time from each camera, so whether the lights were moving or not is a complete non-issue -- the frames were taken from the exact same moment in time, so whatever lights were moving would have been in exactly the same position, making the frames directly comparable. The exposure was set by the 100% zebras on the DVX, and the FX1 was wide open. How is that shot not a valid representation? Or are you referring to something else?
There are dozens of other differences, Barry.
What was sharpness set to in the FX for instance? That could and would make a huge difference.
I'm not saying the test results are "bogus." What I AM saying is that it appears you've taken the best of one, and the average of another, and that's not a fair comparison. There is simply no way the two shots were that close in resolution based on the FX being set up correctly. Just setting the aperture the same doesn't mean much. There are so many other variables that come into play. As I said in the other forum, I appreciate you taking the time to set this up, but not posting the exact settings by itself, invalidates it as a "test" and only validates it as an informed but biased opinion. Shooting all three cams in auto mode is the only benchmark that can accurately be used as a "test" given that you've got to cripple one or the other two of the other cameras to make the test be reasonably similar. I'm not saying at all that the DVX is a weak, poor, or bad camera. But you can't expect to have the resolution differences alone barely show up. There are plenty of folks here who have played with the footage supplied from various websites and the DVX or XL2 alone that have seen different.
Either way, I didn't say the results are "bogus," but I do suggest that they are skewed. That's not a test, but you also know I have tremendous respect for you and would never suggest you created false images.
Barry, I don't really know what there is to prove, is a 1/2 promist better than a polariser? Stupid question and a lot of stupid comments here. It's the same with film stocks, each one has its quirks that some prefer and others hate and some use which ever one suits what they're shooting.
The comparison is valid, it shows what'll happen if you do xy and z with each camera. Sorry but I don't think that proves anything, it shows a lot but possibly if you did ay and z the results might be quite different.
These sorts of arguments are as silly as all the hot air over microphones, mic preamps, cars and batsmen. I pretty regularly bump into one old timer whose shot on everything from VHS to 35mm for Hollywood, he's shot on what the budget allowed and adjusted accordingly, the results had as much to do with his skill, the lighting, the script and the talent as the camera.
"What was sharpness set to in the FX for instance? That could and would make a huge difference. "
Of course it would -- as it would had we changed it on the other cameras too. As listed in the article, all three cameras were set with all settings to middle or "normal" settings. So sharpness on the FX1 was set to "8", on the DVX detail was set to "0", on the XL2 it was set to the middle. Color matrix was set to "normal". Gamma was either set to "normal" on all three, or "cine" on all three, depending on the shot, and notations made of what setting was set when. Gain was always at 0db. The only things we ever changed on the cameras was 4:3/16:9 mode, or Cinegamma/Normal gamma, or frame rate - and always equally. The goal was specifically *not* to let custom settings influence the picture, but to deliver an equally "sharpened" or "not", equally "enriched" or "not" picture. Yes the FX1 could have been used at detail setting 11 or 12 or even 15, but then the DVX could have been set to +7 instead of zero, and the XL2 could have been set at "max" instead of "mid"... where do you draw the line, and decide which gets which? That's why we went with the middle settings, at least it was a common reference point and should have been roughly comparable on each.
There was one one test where we did put them in different modes: DVX/anamorphic/24p, XL2/16:9/24p, and FX1/HDV/60i, and we did that specifically to show each camera performing at its highest-resolution, most uncompromised mode, because we felt that was something viewers would want to see too (that shot is the "tunnel" shot). Even so, all picture controls were still set at neutral/middle.
"Shooting all three cams in auto mode is the only benchmark that can accurately be used as a "test" "
Using "auto" settings only shows what the camera can do when used in "auto", and I don't believe that idea is a valid test because I don't think anyone buys a $3500 to $5000 camera to use it in "auto" mode.
I think the only test that would likely find acceptance as "authoritative" would be to set up the three cameras and specifically tweak the settings in each to extract the "best" picture each is capable of. Only then would someone be able to say that one was definitively "sharper", or "contrastier", or "more colorful" than the others... if they were set up to try to draw out the sharpest, most colorful, contrastiest picture that one could desire. I would be more than happy to conduct such a test, but I fear that some people will just chant "bias bias bias" if it doesn't come out the way they want, which is why I think it would be most unimpeachable to have three independent operators trying to extract the most that they can from their chosen camera.
"Barry, I don't really know what there is to prove, is a 1/2 promist better than a polariser?"
That's exactly the point. Exactly. If you look at the first part of the review, we pretty much came to the conclusion that we were comparing a wrench against a hammer against a screwdriver -- the cameras are so different, and optimized for different purposes, that the choice between them should actually be *easy*, rather than hard. One's a $4,300 shoulder-mount ENG-style interchangeable lens camera. One's an interlaced-only high-def camera with consumer audio. One's a 4:3 camera with 24P and pro audio. The differences between the body styles and price and target market are what one should base their decision on, not on the footage, because the footage on DVD is basically comparable, whereas the other features/differences are not. Unless you *need* HD, in which case the choice should be obvious.
I'm the one who used the word "bogus" and that was the wrong word to use in my attempt to be succinct. I should have used the word "questionable." Spot has already discussed the "questionable" issues in the comparison, so there's no need for me to duplicate his fine efforts.
The results were questionable at best [is what I should have said].
One of the big issues that I see is everyone is so focussed on the HD thing that a big shift in the SD business down here is being overlooked, 16:9!
Now I don't know much about the US market but down under the govt has mandated a switch to DVB, the analogue transmitters get turned off in a few years (yeah I'd like to see that too, but they already sold the spectrum I believe). Now apart from the way better reception more and more content is going to air 16:9 but also the govt has mandated a certain amount of content to be broadcast in HiDef.
So what's all this got to do with cameras? Well there's still a good market for SD content but it has to be 16:9. If it's half decent 16:9 then the broadcaster will upscale it to HiDef for their HD channels. But there's no way 4:3 SD is going to upscale to 16:9 HiDef.
This is no small issue, last year I ran into two old timers who'd stuck to their BetaSP gear for way too long and in a very short space found no one would buy their material, so the demand for 16:9 is getting bigger by the week.
So that unfortunately kind of leaves the DVX100A a bit out of the race. So the choice then comes down to Z1 and XL2. Now I'll accept that the XL2 might give better images in DV however there is another avenue that deserves serious investigation. Shoot in HDV and downscale to SD 4:2:2 and output to DigiBeta. Now that's not all that cheap but if you've managed to sell a program for broadcast you can easily cover that cost. Now Vegas doesn't exactly do that too easily at the moment but I've a sneaking suspicion that day is drawing very close.
The only winge I've got about the FX1/Z1 is that lack of progressive scan. No, I don't give a rats about 24p, it's totally meaningless in PAL land anyway, what I want it for is for PC delivery and streaming content. Why Sony didn't take this issue more seriously escapes me. And it affects their biggest market, early adopters whose viewing world is all computer based.
Just to do a properly formatted screen capture from one of these cameras, don't you have to drop a lot of verticle pixels? It seems to me that most of the resolution difference between these cameras would be visable only in these dropped pixels. A widescreen TV can squeeze a lot of pixels together vertically. Our computer monitors can not.
Also on the topic of verticle resolution: when I view test FX1 footage on my PC, I usually end up looking at it half sized just so that it's small enough to fit on my monitor. Even cut down like this it is higher resolution than the regular SD footage of the XL or DVX. This half sized footage is also progressive. Most HD displays are only 720 pixels high as well. In essense, what most of us are looking with FX footage is progressive playback that is higher resolution than it's SD competitors.
As I'd mentioned, that can make a huge impact. Recommended setting is 13 on sharpness. Because of how the SuperHADS work, that's the equivalent of 5 on other cameras. How they arrived at that figure is beyond me, but there it is.
I agree, people don't buy these to shoot in auto. But each cam needs to be set up to be it's best, and failing that, the only accurate benchmark is to have the cam "be it's best" on it's own.
We'll make the connection, you, me, and Chris. :-) The thing is, now it seems to be a contest of "which camera is better" vs "how good can this really look?"
I'll add my haypenny here (Half cent), and that is that I am inclined to agree with Farss. That and I think that it's gotten too much about my cam is better than your cam on here, IMHO. Grazie has said it plenty, along with many others. The cam is the tool for making the content, it's not the content.
(well, maybe that's a whole penny's worth of thought)