Adam Wilt: F23, RED, EX1

[whachusay]

Interesting read...

Enjoy!

http://provideocoalition.com/index.php/awilt/story/three_three_letter_cameras_ex1_f23_red/

[joe 1008]

Interesting!!!

[Sumfun]

EX1 actually held up pretty well to the much more expensive cameras, IMHO.

[basspig]

Yes, it sure looked a lot better than the Red, and almost as good as the F23.
Frankly, I'm starting to think that the Red, with its one CMOS chip, is but a glorified Canon HV20. :)
I never liked the color rendering of Bayer filtered cameras, now the moire problems the filter presents. The 4K camera really resolves 2.4K at its best.

Now for the F23, I'd almost pay $200K just to be free of CA. CA really bothers me--my eye just goes right to it, every time. When I see it, I think "cheap lens, must be a prosumer camcorder".

But that the EX1 holds up this well against monster cams with presidential prices, speaks volumes for the EX1.

[booth]

Yes, it sure looked a lot better than the Red, and almost as good as the F23.
Frankly, I'm starting to think that the Red, with its one CMOS chip, is but a glorified Canon HV20. :)

While I agree that the EX1 is a great camera for the price I can't help feel that you're getting a little over-excited there BP.

To shed some perspective. The EX1 is almost as good as the 350. The 350 does not match or near the F23, nor the Red, and so nor does the EX1.

Looking at EX1 and 350 footage together they look very close but the 350 just edges it to my eyes. It just looks nicer unless you're shooting in low light. Now have a look at 350 footage alongside Red, it's a different ball-park.

From the outset Adam states that it's an unfair test. There were many things that weren't tested or compared. One of the shortcomings of the EX1 is the color-space of its recording format, which doesn't match the 4:4:4 Red, even despite the Red's blue-channel flaws. The Red is still in beta.

I wish you were right though, because we'd all save a lot of $$$'s :laugh:

[Stevet]

I agree the RED is better, as it should be at it's cost in many ways.
It sure does produce one high rez beautiful image.

I was hoping it would be a bit more sensitive. We had the EX1 side by side. The RED ONE is rated at 320 ASA. At this setting the EX1 was around two stops faster.

Yes, I know the math does not work out, but this is what we saw. Of course, it's all about what noise levels you're willing to live with. The RED at 320 was REAL clean. Of course, nothings stopping you from working with higher ISO. Under 3200K lighting, the RED can get a bit noisy when increasing ISO, but the noise is a very fine grain at 4K images.

There's been some discussion of noise levels from the RED at 5600K verses 3200K lighting. I believe Jim Arthurs has looked at this in detail.

[Jim Arthurs]

There's been some discussion of noise levels from the RED at 5600K verses 3200K lighting. I believe Jim Arthurs has looked at this in detail.

Have I ever... :).

Start here...

http://ftp.datausa.com/imageshoppe/outgoing/RED_232/blue_channel_compare.png

And here are some .dpx files of the above if anyone wishes to experiment on whole 10bit frames...

http://ftp.datausa.com/imageshoppe/outgoing/RED_232/COLOR_TEMP/BLUE_GREEN_3200_VRS_5400K

The bottom line is that the next firmware build (expected before NAB) may have some improvements in this area, but my examples are representative of the current state of the entire system.

I'm only personally bothered by this noise when doing key work, else I couldn't care less. But, if you're keying, try for daylight balanced sources for both blue and green screen.

Regards,

[SteveBagley]

Now for the F23, I'd almost pay $200K just to be free of CA. CA really bothers me--my eye just goes right to it, every time. When I see it, I think "cheap lens, must be a prosumer camcorder".

It's the lens not the F23 that isn't giving Chromatic Abberation in those tests -- but a cheap lens on the F23 and you'll have Chromatic Abberation just as the HDW750P would give with the same lens, but an expensive lens on the 750 and that won't give you any CA.

Interstingly though, I suspect if you took a look at the dual-link HDSDI output of F23 with a good lens pointing at test charts that go up past 960cpw then you'd start to see colour aliases present in the image from the pixel offset on the sensors (although they are full 1920x1080 sensors, they use precision offset to increase the resolution past HD so that they can then filter the signals properly). As I understand it, it is certainly visible on the output of the F23's relative the HDC-1500.

Steven

[Stevet]

The thing is, there's no proof that the EX1 has an blanket CA issue over all existing cameras. Mine looks great under conditions that normally show CA, High light level illuminating high contrast edges, wide open apeture, tele end of lens.

Also, previous reviews have mentioned that CA was real low on the EX1. I'm editing multicamera concert footage shot with JVC HD100 cameras. There's CA is some of the shots for sure.

[craigbowman]

I notice that Adam comes into the comment section later on and mentions the super speeds he put on Red were old and soft. Duh! Why even bother with a chart test at all?

What a total waste. So soft glass doesn't resolve very well. Who would have guessed?

I'm sure glad David Stump head of the ASC technical group did chart tests on Red and I got to see them and read his comments. I would love to see some side by side tests comparing cameras run by him.

[matthew77]

Last time I spoke to Mr. Stump, he was very impressed with the EX1, BTW.

[craigbowman]

The EX is a great camera. So are a lot of other cameras. Tests should be conducted properly and who better than David.

[TheMusician]

I have always been a big fan of the colors produced by the HVX200 and Panasonic in general. By comparison, I have always felt that Sony Cameras(and TVs) are a little "colder" out of the box. The thing that I really liked about this article was the color comparisons on page 5. After color correcting using an eye-dropper on the card, you can hardly see an appreciable difference in colors produced by these cameras(except for a red tinge to the blacks in the guys shirt on the EX1). Even better, these clips demonstrate how much more you can see into the shadows on the EX1. I'm really liking this camera.

[harddrive]

From the outset Adam states that it's an unfair test. There were many things that weren't tested or compared.
I don't feel "unfair" to be the right word (even though it's what Adam himself uses) - "limited" or "incomplete" may be more appropiate, but that doesn't mean there aren't many valuable things to be learnt. Obviously there's no tests with motion, and he concentrates on the front ends of the cameras, not recording options. The real conclusion seems to be the most expensive camera (by far) outperforms the cheapest, with the middle priced camera somewhere in between. No great surprises, really?
One of the shortcomings of the EX1 is the color-space of its recording format, which doesn't match the 4:4:4 Red, even despite the Red's blue-channel flaws. The Red is still in beta.
I've felt for a long time that the whole science of deriving images from imagers is strewn with misconceptions, and there's some bits of Adams report that should be blown up large and pinned in public places. For example (from page 3):
"“But wait!”, you cry, “isn’t RED a 4k camera?” Well, yes… and no. Yes, it has 4k pixels across the sensor, but it also uses a Bayer mask for decoding color. Pulling fine detail out of a Bayer mask image without triggering a variety of luma and chroma artifacts is a devilish task; RED is actually doing pretty well in getting 2.5k out of a 4k imager, especially given its progressive-scan nature .........

Even if you use the full 4k imager’s width, the resulting 2.5k image isn’t markedly sharper than the 1.9k image captured by a three-chip 1920x1080 camera. In HD terms, it’s about 30% more resolvable detail (put the other way, full-raster HD resolves about 76% as much detail compared to a 4k Bayer mask image). "
A 4k single sensor will deliver a 4k image - if you work in monochrome. Beyer filtering is a very clever means of deriving colour without the use of three sensors, but it mut be realised it comes at a price, most notably the de-Beyering process reducing the resolution below the native monochrome resolution of the chip. HOW MUCH below is a bit open to debate, and depends on algorithms used and even subject matter. Adams figures seem a little pessimistic, but "about 2/3" is a reasonable ball-park figure to agree upon.

Also important to realise is that these figures are for luminance. A consequence of the Beyer pattern is that the chrominance resolution of the de-Beyered image is less than for luminance. None of this means that this sort of chip is not capable of giving an extremely high quality output, just that it is inaccurate to describe it as giving a "4k 4:4:4" output, see what Adam then goes on to say:
"And compared to a 2k film scan or full-raster 2k digital cinema masters, it’s still 25% finer detail. This is nothing to be sneezed at: it allows you to push into the image in post by 25% to crop out intruding mics and light stands and the like, or just to better compose a tight close-up, with no resolution loss in the final image. "
No point having a 4:4:4 recording system, if the front end can't deliver a 4:4:4 signal.