Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

ISO Pics analyized

Status
Not open for further replies.
C

CCIRandall

New member
Joined
Jul 19, 2007
Messages
2
Reaction score
0
Points
0
Hey everyone,

Over at CamcorderInfo's blog we ran the RED pics through some testing software we use on all of our cameras, both camcorders and digital still cams. We realize (as we say in the post) that the images they put out weren't in lab-like conditions as most of the shots are underexposed and not evenly lit, so the color readings we got weren't any sort of scientific, but we did find some crazy amazing noise levels.

Check out the post here:
http://www.camcorderinfo.com/content/RED-camera-pics-out-our-analysis-Smooth-but-bad-32968.htm
 
Cool charts, but you don't seem to have taking into account that the camera was operating gained-up and yes, every single shot was deliberately underexposed...
 
Either I don't understand something or something went wrong when testing. When I look at the images in Safari (which shows color profiles) the RED images on this forum from look totally different than on camcorderinfo. Did you take in account that the test images were Adobe RGB on not sRGB?
 
That makes total sense as what you're seeing is the quality of the sensor giving an image with as much lattitude as possible for post colour correction. Because it's raw, you can decode it however you like though.

Graeme
 
Also, the colours are very accurate, but they're not in the same colour space that the test software was expecting, hence they look very under-saturated.

Graeme
 
Running tests on posted JPEGs?

Running tests on posted JPEGs?

For noise and color accuracy, no less?

What a royal waste of time, IMHO.

:w00t:
 
read more carefully

read more carefully

this is exactly the reason why jim and team red are sometimes hesitant to post day to day results that aren't just pretty pictures. now others who don't read carefully too might pick up this thread from google and wrongly assume the iso stills were intended to be properly exposed.

thanks for honesty and engineering excellence team red. keep it up.
 
I hope people realize when we say these are preliminary images, we mean it. I mean the point is that the noise levels are CRAZY GOOD, and that's all we really wanted to point out. The color stuff is interesting, but yeah, it's just the run-of-the mill "We don't really know too much yet."

I think we're being fair when we say, "Since the shots were under exposed, the noise readings would likely score better than our current analysis. Noise free filmmaking in a digital world? Perhaps, although we're still waiting to see it in our labs. We talked with our friends at DigitalCameraInfo.com for a comparison to high-end still cams, and they said noise wise it blows away every camera they've tested. They told us while these images might have some color shifting going on, it's really impossible to gauge any sort of useful color data because all of the images are underexposed and we can't really give it a fair comparison to another camera till it's properly exposed."

Don't worry RED heads, we're not saying it's bad.
 
They've said it can't be any kind of fair comparison for color but that the low noise blows anything they've seen away. If nothing else, it certainly vindicates Red's demonstration of low noise.

doh! Beat me to it above.
 
As I said, if you test for sRGB for an image not in sRGB, you get funny results :-) Nice to get independent confirmation on the noise though.

Graeme
 
Hi Graeme,

What is the real world difference between RGB and sRGB ?
 
CCIRandall said:
I hope people realize when we say these are preliminary images, we mean it. I mean the point is that the noise levels are CRAZY GOOD, and that's all we really wanted to point out. The color stuff is interesting, but yeah, it's just the run-of-the mill "We don't really know too much yet."

I think we're being fair when we say, "Since the shots were under exposed, the noise readings would likely score better than our current analysis. Noise free filmmaking in a digital world? Perhaps, although we're still waiting to see it in our labs. We talked with our friends at DigitalCameraInfo.com for a comparison to high-end still cams, and they said noise wise it blows away every camera they've tested. They told us while these images might have some color shifting going on, it's really impossible to gauge any sort of useful color data because all of the images are underexposed and we can't really give it a fair comparison to another camera till it's properly exposed."

Don't worry RED heads, we're not saying it's bad.
Click to expand...

It would be worthwhile to provide full data about the images on the site with the test results, including info on why the pictures were underexposed and the fact that the color space of the images was different than that of the testing software, so people don't get an incomplete impression.

Could save time playing catch-up with people who think you've identified flaws that don't really exist.

The point on the low noise is the real news here, and it seems that you feel this point is valid, even startling, notwithstanding the deliberate underexposure of the image. Why not clarify the confusions that muddy up the findings so people can get an accurate sense of what the picture represents?
 
Well, the camera has it's own RGB space, unique to RED. sRGB is defined by specific numbers. If you know one and the other, you can construct a matrix that transforms from one to the other. Being video, we don't generally go to sRGB, but to REC709 instead. If you know the numbers, you can transform between any colour space you want.

Graeme
 
I think I would wait until the LART team gets some testing done.
You cannot perform any true test with the JPEG images that were
posted. I remember the Shoot out awhile back between the HDV
cameras and the HVX. They did not at all have the cameras tweaked
to perform at the best possible levels. They were knocking the HVX
but I think the HVX sells many more units than the comparable
Canon or JVC cameras. They should re-run those tests and compare
images with the Canon HV20. And the LART team should throw a Canon HV20
in the mix as well.
 
Red Guy said:
They were knocking the HVX
but I think the HVX sells many more units than the comparable
Canon or JVC cameras.
Click to expand...

Just because I happened to go through this myself I don't think that's accurate. I believe JVC said they sold 12000 HD-100s before the HVX hit the market. I'm not sure how market share shakes out now though.

I ran my own tests for a month and tried every HVX setting there was. The HVX has it's points but the roundup tests by the pros were largely the same results I saw.

In the end we've all got to test for ourselves. I hold all outside tests with a grain of salt. Though I did see Jackson's short in Vegas and I pretty much trust that.
 
joelnet said:
Just because I happened to go through this myself I don't think that's accurate. I believe JVC said they sold 12000 HD-100s before the HVX hit the market. I'm not sure how market share shakes out now though.

I ran my own tests for a month and tried every HVX setting there was. The HVX has it's points but the roundup tests by the pros were largely the same results I saw.

In the end we've all got to test for ourselves. I hold all outside tests with a grain of salt. Though I did see Jackson's short in Vegas and I pretty much trust that.
Click to expand...

I see footage from all of these cameras. In the hands of a skiiled op they
can all produce great results. I have also seen some crap HVX footage.
I like the Canon but the HDV tape compression is a bummer.
I also shoot timelapse with a Nikon D200 and I hope the RED will
be able to produce moving (and timelapse) images that look as
good as what I am getting out of the Nikon.
 
While the display of our low noise levels is nice, what is missing form this article is that it is a "cartoon shot" that we pulled off just to show the rubberband-like nature of our image. We certainly have not done any color matrix calibrations for that radical of under-exposure. And that fact seems to be missing in the article. If someone is going to take and use one of our images without permission, I think the least they could do is talk to us before publishing so we can make sure all the facts are there. I realize that Randall had good intentions, but this tells us we just shouldn't post anything more until we are done.

Jim
 
CCIRandall said:
Hey everyone,

Over at CamcorderInfo's blog we ran the RED pics through some testing software we use on all of our cameras, both camcorders and digital still cams. We realize (as we say in the post) that the images they put out weren't in lab-like conditions as most of the shots are underexposed and not evenly lit, so the color readings we got weren't any sort of scientific, but we did find some crazy amazing noise levels.

Check out the post here:
http://www.camcorderinfo.com/content/RED-camera-pics-out-our-analysis-Smooth-but-bad-32968.htm
Click to expand...

Just yesterday I was on CCI checking out your review of the Canon HV20, and wondering why we haven't heard from you guys over here on Reduser! Welcome, and thanks for the test results:biggrin:
 
Red Guy said:
In the hands of a skiiled op they
can all produce great results.
Click to expand...

Thanks for adding that. I meant to say that. They are all fine and if you are good with any of them you'll get the good stuff.

RED will still kick their asses thorougly into submission soon. But they are all fine.

(Ok, I'm positive talented people will be using these and lesser cameras to tell better stories than some REDs will ever tell... but I'd still rather be shooting RED)
 
Status
Not open for further replies.
Back
Top