comparison with a DLSR please?

[dvpixl]

wonderful images Jim. The first one is just pure eye candy for details.

[Brian Harbauer]

Jim, any chance you could upload a dpx or a tif? The tone that is held is incredible!

Great eye for composition too!

Cheers

[Elizabeth Lowrey]

Last one... I'll stop now. ISO 800.

Jim

Jim, that is a freakin gorgeous shot!

[Häakon]

Alrighty.... so I went out this afternoon and took some shots with my RED (original version, build 12), a Canon 5D (12 megapixel full-frame professional dSLR), and a Canon XTi (10 megapixel 1.6x crop prosumer dSLR). The intent was to get some comparison images between the cameras per the original poster's request. This is not a scientific test or a be-all, end-all set of conclusions; rather it is somewhat informal and merely one isolated set of shots that may provide insight to those who view them. Hopefully it doesn't start off a firestorm.

Carrying around three camera bodies, a tripod, my RED and all of its gear, and two lenses from place to place was not an easy task. The first thing that's clearly obvious from the outset is that RED is a lot more inconvenient as a still camera than a dSLR is. Jim has mentioned that he no longer uses his still bodies anymore because of how great the image quality is from RED - and while it is great, there is no way I can see RED being an SLR replacement. The size, weight, and boot up time prevent it from being a reasonable choice. That being said, it's the quality of the images that matter most, so let's delve into the pictures.

For the RED, I used the 18-50mm RED zoom on all of the shots. For both of the Canon bodies, I used the Canon 24-105L zoom. The XTi has an APS-C sized sensor, which is a fancy way of saying that it's 1.6x smaller than the size of a standard 35mm frame of film. Fortunately, the sensor size is very close to that of RED's "Mysterium" chip, so focal lengths behave pretty consistently between these two bodies. The 5D, on the other hand, is a "full frame" camera - or one whose sensor is the same size as a 35mm frame of film. When putting a lens on the 5D, the focal length behaves precisely as it would on a standard 35mm still photography camera. This made setting up some of the shots a little tricky as I had to roughly "find" the frame for each camera to make the shots similar, but for this informal test it wasn't a problem. There are a couple inconsistencies with framing between shots, but nothing that had an impact on the test.

Normally when shooting motion picture, our shutter speed is pretty well set - so the aperture and sensitivity of the camera are the only things that we can use to affect the light level. Traditionally, we can use ND filters to cut down excess light when needed. When shooting stills, however, the shutter speed can be changed to anything we like - so I took this approach while doing these tests. Because it was a sunny day, using a fast shutter helped not only to ensure clear, steady shots, but to help control exposure as well.

This first shot was taken at an f/4 on all cameras. The RED and the 5D were both set to ISO320 (as this is RED's "native" ISO rating), and the XTi was set to ISO400 because no ISO320 is available. The shutter speed was 1/1000 for the RED and 1/1600 on both Canon bodies. The RED and XTi used a 50mm focal length while the 5D was set to 80mm to compensate for the crop factor. All images were shot in the RAW mode.

Here are the images:

RED:


5D:


XTi:


For these tests, I set all of the bodies to a daylight (5600 degree) color temperature, though it's important to note that each renders color slightly differently. Also, due to the RAW conversion process for both the Canon images and the RED shot, it is difficult to get absolutely matched color "on-the-fly." The Canon software is much more robust than the RED offerings, but both REDCINE and REDAlert! are in their initial stages of development and should see marked improvement over time. As you can tell, there is still plenty of information available in both the shadow and highlight areas of the frame - and this was taken on a bright, sunny day, so that kind of performance is very strong. As we will see, the dynamic range of the Mysterium chip seems to be on par with the dSLR offerings; a great thing indeed.

[Häakon]

Shrunk down to those sizes, of course, the images all appear more than acceptable. RED has already proven that for the web, it can take great still frames. How does it hold up detail-wise next to the other dSLRs, however? Here are some 100% crops from each of the original images above:

RED (100% crop):


5D (100% crop):


XTi (100% crop):


Here, the detail and resolution of the dSLRs outperforms the RED - especially the full-frame camera - but the RED still holds up well. Unless your intention is to create 20"x30" prints from your RED camera, it is a very usable option as far as image quality is concerned. I have long maintained that the 4K ability of the RED is it's superior strength over other HD cameras - because even if your final output is "only" 1080, the oversampling ability that the RED provides will give you a far stellar picture than any other "standard" HD camera out there.

[Häakon]

Moving on, here are some other shots from the day:

I chose this shot specifically to test the range of each camera from deep shade (the sun was shining from the opposite side of the building) to bright sky (which would never hold up on a lesser video camera). I have raised the shadow level of these images slightly to show how much detail remains in the dark areas of each shot - while they all easily hold the sky in the upper part of the frame. Again, I was handily impressed by what each camera was able to deliver.

RED:


5D:


XTi:

[Häakon]

Here are some 100% crops of the images above:

RED (100% crop):


5D (100% crop):


XTi (100% crop):


These images were all shot at an f/8.

[Häakon]

RED (100% crop):


5D (100% crop):


XTi (100% crop):


Here we can see more noticable noise in the XTi shot, due to it's raised ISO level (the camera's "base" is ISO100). It does appear that the RED frame is slightly underexposed compared to the other two, but I have found that at ISO320 it is cleaner overall. However, this is slightly offset by the fact that the camera is not as sensitive.

[Häakon]

RED (100% crop):


5D (100% crop):


XTi (100% crop):


I do see consistently more detail in the dSLR shots, as noted by these 100% crops - but remember that you're comparing a frame of video footage to a 12 megapixel dSLR image! No other camera in this price range would even come close.

[Häakon]

Here's a nice shot of the Mormon Temple down here in San Diego. Because they patrol this place like it's Fort Knox, I was only able to shoot it (through a fence) with the RED and the 5D. Here we see some lovely color rendition from the RED, and a cleaner sky as well:

RED:


5D:


RED (100% crop):


5D (100% crop):