What compares to Mysterium?

[Christian Berg]

Hi Jim, are you considering making a stillcamera? A professional one with full SLR sensor at a better price than the high end Canon and Nikon would be great. I´m in the market but not up for paying that much since i just oaid my Red One.
/Christian

 

[Jeff Deveraux]

Hi Jim, are you considering making a stillcamera? A professional one with full SLR sensor at a better price than the high end Canon and Nikon would be great. I´m in the market but not up for paying that much since i just oaid my Red One.
/Christian

Or how about comparable price but with aluminum body and PL mount?

 

[mradeck]

Are you talking about a RAW image, or JPEG? Have you done all of the color correction and sharpening to the Red image that the Nikon does in-camera to JPEG, or in software in the RAW converter? ...actually I'm not sure any of the DLSRs "raw" output is truly un-sharpened, etc.

red-raw well developed in redcine, nikon compared with jpg(fine-nikon-compression) and with nikon raw -photoshop cs2 - then saved to jpg.

tried different sharpening values in redcine and cs2-raw-developer (also without!).

nikon 3008x2000 scaled to more than 4096x2048 rotation corrected cropped to match better to red. red only rotated a little. panasonic also scaled rotated and cropped.

1. Nikon D70, Zoomlens = best
2. Panasonic DMC TZ3 = ok
3. RED = unsharp,

also when all scaled down to 1080p HD the RED-picture is very unsharp in this comparision

Tested also with DVS-Clipster in 4k, viewed on 4k Mitshubishi LCD-Monitor.

So if i use Nikon D70 for timelaps and want to mix with red 4k2:1 then i have to blur the nikon heavy for matching the ´"RED-look".


http://www.hdforum.de/red/resolution/compare/scrsht/Nikon-D70-DSC_0390-raw-cs2-scrsht01.JPG
http://www.hdforum.de/red/resolution/compare/scrsht/DMZ-TZ3-panasonic-P1010521-cs2-scrsht01.JPG
http://www.hdforum.de/red/resolutio...aprime-A002_C001_000711_0571-cs2-scrsht01.jpg

I think the problem must be a low pass filter glas and the huge compression in RED-Raw

 

[mradeck]

when you shoot in 4k 16 by 9 then red uses only 4096x2304 pixel not more!! so it's 9 Megapixel.
with 4k2:1 it's 8 Megapixel 4096x2048. Red don't use all active pixel at this moment, with these formats. So the size of the used pixel-area is also lower, must be around 20x10mm.

 

[jbeale]

Thank you for doing those tests and posting them. Yes, I agree with you the Red example is the least sharp of the three.

 

[Michel Hafner]

The Red picture looks like mush in the grass. Are you sure that is a good as it gets? That does not look like 3.2K resolution at all.

 

[Emery Wells]

WRONG. A full-frame digital SLR will have a sensor significantly larger than the RED's sensor. Full-frame on still cameras is NOT the same as super 35mm in motion pictures.

I don't mean to be harsh, but people need to stop making this mistake.

Yep, you are correct.

Although my sentiment is that trying to match two disparate systems for the purpose of critical comparison doesn't make sense. If I were in the market for a camera I'd rather take into account the FOV difference and have a full frame sensor but then my application may be different than others. I didn't address the question appropriately.

 

[mradeck]

The Red picture looks like mush in the grass. Are you sure that is a good as it gets? That does not look like 3.2K resolution at all.

yes i compared best results. more complex pictures with many details, have more problems with the red compression. typical for jpeg2000 especially in shadows the details of grass f.e. are often complete missing. On large displays or projections it looks sometimes like hdv because looks simular in loosing much details...
but if you use testcharts you resolve much more details, because the compressor has nothing to do...
but the all pictures of red look to me that there must be a optical ?? (can someone confirm this?) lowpassfilter before compression... because of it's overall unsharpness.


cheers MR

 

[Gunleik Groven]

Untill 16, RED seems to have a problem with dense variations and details in greens, and as they get darker and less saturated, this gets worse.

One of my build 16 hopes...

This comparison was also a very good argument for a IR filter...

 

[Dj Joofa]

but the all pictures of red look to me that there must be a optical ?? (can someone confirm this?) lowpassfilter before compression... because of it's overall unsharpness.

The coarser quantization of higher frequency components in compression acts as a "low pass" filter in the sense that it provides less attenuation of lower frequencies.

 

[Jannard]

These results are very different than I get with the Canon 1D MKIII. But I am shooting build 16.

1. Are you shooting REDCODE 36?

2. Are you checking HIGH on the OLPF box in RED Alert!?

3. Are you shooting at F5.6-f8? You need to take the lens out of the equation.

4. Is this the 1st frame of the sequence from RED? Needs to be from the middle. The 1st frame or two will be softest from camera shake.

BTW, charts are the way to measure resolution. These shots are not measuring resolution from the camera, they are measuring how to break the old compression. I will add that all DSLRs have in camera sharpening. We don't.

How about we try this again in a week after Build 16 is released in Beta?

Jim

 

[Ken Hendricks]

"in a week after Build 16 is released in Beta"

There it is! You heard it here first!

Where is the fanfare?