The big Single Chip Camera Evaluation, as scientific as your going to get....

[Duke]

Did you see the SCCE shoot out? 12 cameras (Alexa, film, F35, Red, AF100, Sony F3, Phantom Flex, Weisscam HS-2, Canon 1D MkIV, 7D, 5D Mk II, and Nikon D7000)

They had 140 camera and support crew who were experts in each of the cameras and scientifically tested most of the big chip cameras. They didn't draw conclusions just gathered the results. Zaccuto made a documentary that will be out later on blu-ray. They are just starting to release results.

The first snippet is here:
http://www.zacuto.com/the-great-came...11/episode-one
(Note: you can't be behind a proxy and see it. You have to direct access.)

Alexa was the winner in most people's opinion. Red MX won on resolution of course. Red wouldn't let an Epic participate even after Epic's release. (Which causes me to wonder.) Jim Janard specified how he wanted the Red MX set up so you can't argue the Red was at a disadvantage.

Barry Green, the Panasonic guru, said this "The F3 looked excellent in the SCCE tests, it was frequently mentioned in the same breaths as the Alexa, it blew the F35 into the weeds, and it was all done without s-log." Now beating an F35 is really something. The F35 costs ten times as much.

I thought this was interesting "film actually has phenomenal highlight detail, but isn’t actually that good in the shadow detail."

"The F35 (a $150,000 camera), for example, didn't fare nearly as well as I expected, the F3 (a $14,000 camera) pretty much handed it it's butt on a silver platter. The Red MX did pretty well, definitely besting the F35 and most of the others, and dominated in resolution. The Alexa (an $80,000 camera) stole the show. It was just astounding. It was dominant and to my eyes the clear winner, and everyone I spoke to agreed. Just no comparison. And film? Well, that was one of the disappointments. Film just didn't dominate. Always good, but (except for res) the Alexa was always as good or better. The F3 did extremely well. The F3 was indeed a baby Alexa." Camera prices added.

Still waiting for full results of course but the F3 and Red MX (a $25,000 camera) had about the same DR. With S-Log the F3 will have at least another stop of DR to go with that extremely good signal to noise ratio, -63db.

[Flozem]

Thanks for posting... spent my whole sunday morning reviewing last-year's shootout. Never knew they did something like that.... kewl.

[GeorgeMaier]

They describe the 7D as a pro camera??? Really? It's crop right? Semi-Pro at best?

[Flozem]

I know at least a few pro photographers (yes they make a living off of it) that use the 7D for its qualities and much lower pricepoint. It's the first few inches in behind the seeker glass that make the difference, not the equipment...

They can use a few of these cameras + a bit cheaper lenses as backup.

[GeorgeMaier]

Don't get me wrong, I love the 7D, it's a brilliant camera, it's just a surprise to see it with film cameras worth many many times more. Interestingly it seemed to perform better in the dynamic range tests than the 5D in my opinion...

[Daniel Rutter]

Don't get me wrong, I love the 7D, it's a brilliant camera, it's just a surprise to see it with film cameras worth many many times more.

That's because a lot of people go for these cameras when they're shooting indie films. It's got everything the 5D MkII has, minus the full frame and higher price tag. Of course, I guess there's a few other things missing, but nothing that really "counts" if you're filming certain things.

[Flozem]

Don't get me wrong, I love the 7D, it's a brilliant camera, it's just a surprise to see it with film cameras worth many many times more. Interestingly it seemed to perform better in the dynamic range tests than the 5D in my opinion...

That's the whole point of the test... to compare hdslr cameras to high performance HD videocameras. The 7D performed better in the high-key lighting... low light was on par with the 5D and 1D.

The term pro-camera is just a reference to the product classification Canon gives to these hdslr's... in reality, they prolly could also compare with the even cheaper 550D / 600D. The latter two might even perform at the same as or even better level than the 7D. There's not a lot of difference in the HD functionality between these hdslr's.

I mean: 25k+ HD videocameras vs 1k+ hdslr's.... that's great if you're on a tight budget. Though they have a point that post-processing may as well reduce your savings. A great development nonetheless - quality of these HDSLR's is bound to improve in only a few years time...

[Halsu]

Did you see the SCCE shoot out? 12 cameras (Alexa, film, F35, Red, AF100, Sony F3, Phantom Flex, Weisscam HS-2, Canon 1D MkIV, 7D, 5D Mk II, and Nikon D7000)

Yep, seen that, not entirely happy with the way they did the tests...

Alexa was the winner in most people's opinion. Red MX won on resolution of course. Red wouldn't let an Epic participate even after Epic's release. (Which causes me to wonder.) Jim Janard specified how he wanted the Red MX set up so you can't argue the Red was at a disadvantage.

I'd say the above is simply not true. First of all, Red did not supply an Epic to the test, but that's not the same as "not allowing" it. All the other cameras were supplied by rental facilities too, not the manufacturers, as far as i know.

Second, i wonder where you got the impression Jim Jannard gave setup specifications, i'm virtually certain that's not true. In fact, in the discussion on the Red forum, Red employees made it clear that they had nothing to do with the way the test was performed. They were asked AFAIK, but declined. The reason for this was previous experience - all the manufacturer's representatives can usually do in a test like this is to say they dislike the way the test is conducted, but they can't really affect what's being done. The result is a test that's done in a way they disapprove, but has their "signature" in it. Not good.

Still waiting for full results of course but the F3 and Red MX (a $25,000 camera) had about the same DR. With S-Log the F3 will have at least another stop of DR to go with that extremely good signal to noise ratio, -63db.

AFAIK, S-log shouldn't give additional dynamic range with test charts, if the cameras are measured properly: the test is not about how the cameras look and perform in real life shooting scenario (S-log should help there), but simply what is the difference between the absolute brightest and absolute darkest illumination the cameras can register. This doesn't account at all for how the images look, just that a difference between two brightnesses can be extracted from the file in post.

As far as the other, "real life" dynamic range tests go, the method was in my opinion a bit useless: they set every camera with the same ISO, shutter speed and f-stop, and compared when will the cameras clip. This doesn't really tell us much about the camera's performance, and for most cameras in the test, i would call their exposure "wrong".

With cameras like Red One (and Alexa), where the ISO setting is just metadata that doesn't actually affect exposure, this is actually a measure of sensor sensitivity - the more sensitive, the more the camera will clip highlights (but at the same time, the more there's unused dynamic range "footroom" hidden in the shadows).

The proper way (again, in my opinion) to compare the camera's dynamic range is to expose each camera so that no highlights burn on the sensor level, and then see how much usable detail can be extracted from the shadows with acceptable image quality - either using in-camera gamma curves etc. (for DSLR:s and traditional video cameras), or in post (for RAW and Log cameras).

I'm pretty sure the perceived results would have been very different this way.

[Dr. Benway]

...but nothing that really "counts" if you're filming certain things.

Remember the crop factor Daniel. Focal length can count a great deal. Having to buy new glass can count a great deal too. Because your EF glass on a crop factor body suddenly is 1.6x longer - your 24mm is now a 36mm lens.

[Duke]

Halsu, you're giving your opinion when you don't even know how the tests were conducted or about the F3. You looked at the one "7 stops over key" deliberate over exposure test. The objective test of DR tested both the high end and low end with an illuminated chart (ie. the right way.) Hence the bar chart. There were two other dynamic range tests beyond that.

BTW, several tests now show that S-Log on the F3 adds between 1.5-2 stops on the high end, beating the Red MX and placing between the F35 and Alexa. Epic light with HDRx may beat that, but we don't know yet as it isn't available. Epic-M isn't near the same price point.

Please get your facts straight before you give your opinion.