MX and Dragon Comparison

[Gavin Greenwalt]

What are those horizontal stripes on Dragon shots in the "pushed_12800" image? Look at the side handle and the Knopfler box in the foreground.

That's a CMOS artifact: Clipping pixels contaminate all the other pixels in a row. It's seemingly worse on Dragon from all the SxS shots. On the plus side, it's harder to clip with Dragon.

 

[Gavin Greenwalt]

, this test was only made, because so many told me Dragon is much more noisy than MX.
that´s actually untrue. the other stuff, DR, skintones, and so on shouldn´t be judged with a table lamp - that is right ;-)

In your test that might be true, but I've seen quite the opposite in plenty of other tests. Dragon is great but lets not whitewash real concerns.

 

[Brian Boyer]

Overall though, I think everyone is a little disappointed in the "improvement" in the sensitivity of the sensor, a native 640 would have been nice but it looks like we're still around 320, which makes sense since it was stated the photsites are smaller on Dragon.

If I'm understanding correctly, it sounds like the sensitivity of the OLPF is the issue, not necessarily the sensor itself.

True, the smaller photosites don't let in as much light but they seemed to let enough in to most everyone's satisfaction with the more light sensitive old OLPF.

 

[Freya]

i guess i am crazy and blind....to me the mx looks quite a bit better....especially considering that the exposure seems to be different, although i can't figure out how because the small highlight is blown in the dragon but not in mx but the rest of the exposure does not look THAT much

Well just to let you know you are not crazy. I also prefer the MX shot of the lamp although Jan is right that the yellow flowers in the forground clearly show more detail. On my monitor the tractor looks better.

HOWEVER the big difference to me seems to be that the MX footage is just slightly more contrasty and I quite often like more contrasty footage than other people do so....

The MX also appears to have more detail in the green channel, there is a leaf on the left hand side that isn't visible in the Dragon footage at all.

I'm not sure this is a good test for showing the differences tho. Seems to me it shows more about what the two cameras have in common.

Freya

 

[Freya]

The MX also appears to have more detail in the green channel, there is a leaf on the left hand side that isn't visible in the Dragon footage at all.

...and another way of looking at that might be that the MX is a bit more "hot" in the green channel!
Which might not be a good thing when you are looking at a picture from further out or considering skin tones!
Who can say!

Freya

 

[Jan Reiff]

the question IS, why are there so many different results.
i did nothing in this test to look Dragon better ...

 

[Brian Boyer]

I know Jarred says stick with old Dragon if you want, but he is usually saying that when his blood pressure is high after a particularly hot thread of complaints.


Is "not upgrading" really a viable solution? It is swapping one set of problems for a DIFFERENT set of problems. The problem is timing for customers who need the fix.


I`d still like more info as to what is going on with the FIX and where the new 2000 ISO will land. Its a bigger job than they thought, I get that, but assuming red noise is fixed and they get closer to the advertised 2000 than the low number we have just now it will be great times for current owners and future.

When I said people don't have to upgrade, I was talking about those people with an MX camera, like what was being compared in the test. I wasn't talking about old OLPF Dragon to new OLPF Dragon. Although, that's certainly an option for them, too.

Not upgrading from MX to Dragon is as viable an option as it was for people who had Adobe CS2 and didn't upgrade again until CS5. If the upgrade doesn't suit your needs or isn't worth the cost for what you feel you gain in the process, don't do it. It's really that simple.

And by the way, life is pretty much entirely swapping one set of problems for another set of problems.

 

[Brian Boyer]

The MX also appears to have more detail in the green channel, there is a leaf on the left hand side that isn't visible in the Dragon footage at all.

I'm not sure which leaf in which picture you're saying isn't visible but, from what I can tell, it seems what you're calling a leaf is a reflection, which is in a slightly different place in each pic because of the slight difference in framing and the fact that the camera in each pic is in a different orientation with different accessories attached.

 

[Bohdan Klos]

When I said people don't have to upgrade, I was talking about those people with an MX camera, like what was being compared in the test. I wasn't talking about old OLPF Dragon to new OLPF Dragon. Although, that's certainly an option for them, too..

Ah, understand.



I suppose confusion is symptom of what will happen if Old OLPF cameras are left in customer/renters hands. Best that all Old OLPFs are removed from circulation - rental companies offering old OLPF once new OLPF is fixed is bad idea.

 

[Brian Merlen]

I wonder what would rent better today though, without the fix... I have new OLPF, maybe I should go for old until the fix is out? Kind of a pain to deal with shipping though if the FW isn't far off...

 

[Charles Bergquist]

That's a CMOS artifact: Clipping pixels contaminate all the other pixels in a row. It's seemingly worse on Dragon from all the SxS shots. On the plus side, it's harder to clip with Dragon.

From my testing it's not even clipping pixels that cause it, just enough of them in a gathering, like a kino, window, etc. When placed next to others. It's only horizontal as well. This is the hardest thing to swallow on the Dragon sensor compared to MX and every other sensor. It is a CMOS artifact that's present in every camera, but on Dragon it's much brighter and much more saturated ( magenta ). We'll see what the next calibration brings, I know they're aware of it. I feel like this issue is more important than the noise at the moment.

 

[Jan Reiff]

but you don´t see it in a serious grading, pushing up to ISO 4000+ is not really the look for a movie

 

[Charles Bergquist]

but you don´t see it in a serious grading, pushing up to ISO 4000+ is not really the look for a movie

Unfortunately it's at all ISOs. Examples below are 640 ISO.

https://dl.dropboxusercontent.com/u/161718/A003_C016_06202W.0000245.jpg

https://dl.dropboxusercontent.com/u/161718/A003_C016_06202W.0000245F.R3D

https://dl.dropboxusercontent.com/u/161718/A001_C001_0621X6.mov

https://dl.dropboxusercontent.com/u/161718/A003_C041_0620WH.mov

https://dl.dropboxusercontent.com/u/161718/A003_C016_06202W.mov

Or maybe I got a bad sensor if this isn't happening to anyone else at these ISOs. I'd like to wait until the next firmware to see what the Red crew does with this in general.

 

[Niko Pueringer]

Unfortunately it's at all ISOs. Examples below are 640 ISO.

https://dl.dropboxusercontent.com/u/161718/A003_C016_06202W.0000245.jpg

https://dl.dropboxusercontent.com/u/161718/A003_C016_06202W.0000245F.R3D

https://dl.dropboxusercontent.com/u/161718/A001_C001_0621X6.mov

https://dl.dropboxusercontent.com/u/161718/A003_C041_0620WH.mov

https://dl.dropboxusercontent.com/u/161718/A003_C016_06202W.mov

Or maybe I got a bad sensor if this isn't happening to anyone else at these ISOs. I'd like to wait until the next firmware to see what the Red crew does with this in general.

That's pretty rough. Bigger issue than noise, IMO.