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 :)

Camera testing and comparisons

Status
Not open for further replies.
Peter Moretti said:
So it does not matter what rating you give the camera b/c you're killing the image on both directions and counting how many half stops it took to do that. All rating the camera at 800 would have done is add a stop to the top and remove a stop from the bottom. But the total number of stops would have stayed the same, AFAICT.
Click to expand...

I think that this is exactly the point. ISO in a test like this is somewhat irrelevant. It is a starting point.
 
geoffboyle said:
I knew I shouldn't publish these, or at least not share them here.
Click to expand...

Come on, please, don't lump us all into that bucket.

Stay long enough and you will get an accurate reading on how eager the greater community is to receive you --and defend you when need be.

We will of course ask every question conceivable specific to Red, but that will only complete the picture of yours and of this community's character. I am convinced that fairness prevails in both.

sander kamp said:
Geoff, from what I understand it is a fair test, personally I can't see anything wrong with it.

People here might be disappointed that the Red One MX doesn't fare better against the Alexa. I was under the impression that the latitude difference would be 1/2 to 1 stop but this shows an entire different picture. So don't take it personally if people are not happy with the test, it's not about you.
Click to expand...

This is exactly what I was speaking about.

From what I gather this was a very tedious and dreadful process, but the comparison is in fact unfinished; and it being geared to a broad audience also means that a loud [*and mistaken] message has been sent.

Your word carries a lot of weight, and I am glad to see you post here. Please don't run away, and do continue to teach those of us who are looking for lessons in your posts.

*It seems there are missing frames on the top end, as explained by Peter.
 
Peter Moretti said:
Sander,

The test is perfectly fair, FWICT. But your impression illustrates why it's current presentation is probably not.

You are seeing a huge difference in terms of DR between the two cameras. Probably FOUR stops, b/c the the Alexa Log has eight more half-stop images than the Red MX Log has.
Click to expand...

No, I wasn't looking at how many images were posted. I understand that is not a measure of things. I was looking at the log images and how they compare and it seems that Alexa has about 2 to 2.5 stops more dynamic range, which is a huge enough.

Now PLEASE somebody tell me I am seeing things wrong and there is more room in the RED images. I have a MX camera myself and I need to rent it out. But unless somebody points out a serious flaw in the test or can do something to the R3D files that suddenly shows 2 more stops of information you have to be realistic and admit to what you see.
 
Guys.......Jim already posted (months ago) a Dynamic Range test results chart
comparing the Alexa and MX and the differnce was about a 1/2 stop in favor
of the Alexa. So this DR question has already been answered. THE END. ;)


:smash::smash::smash:
 
Roberto Lequeux said:
Come on, please, don't lump us all into that bucket.

Stay long enough and you will get an accurate reading on how eager the greater community is to receive you --and defend you when need be.

We will of course ask every question conceivable specific to Red, but that will only complete the picture of yours and of this community's character. I am convinced that fairness prevails in both.
Click to expand...

I agree with Roberto ... yes Redusers are passionate and fiercely loyal (I watched a video recently where Vincent LaForet and Philip Bloom joked that we Red folks are "Scientology crazy" in our dedication, LOL) ...

With that said, I believe most are very fair and open to whatever truth comes out of a test. I think some of the feedback you've gotten is just that when Red is involved, everyone wants to be as accurate as possible -- otherwise it could skew people's perception unfairly.

But trust me Geoff, we all know how much work goes into this without any monetary reward, it's really just out of curiosity, desire to share and be of service. So thanks for that! :)

Sincerely,

Anthony
 
Peter Moretti said:
Why not converted from ProRes 4:4:4:4, since that's the highest quality the camera can currently record to?
Click to expand...

4:4:4 VS 4:2:2 won't change dynamic range. Or will it?

I'ts an YUV encoding VS RGB encoding... or is it? With prores you never know...


RED 400 asa or 800 asa doesn't matter. I see more noise in the RED dpx @ -6 than in Alexa's dpx @ -6.

But "color rendition" and overall contrast in Alexa Log files @ -6 is flatter than RED ones. You got less noise in the low lights but it's perhaps due to how Arri encodes the shadows in the log file? Is there noise reduction applied as this is not raw?

A lot of question marks???


Different cameras with diffenrent post production path. I'm also interested in Alexa now. When is the 4k version available? ;-)

Pat
 
I love Geoff's work and I thank him for the effort put on the test but I must say that it seems a wee biased probably involuntarily.

I'm comparing DPX's and the Alexa log DPX's should really stop at + 5 and a half stops since it shows the same loss of detail as the RED log +4.

I would rate the difference at the high end in 1 stop and a half.

On the low end of exposure I would agree that pics looks fairly similar for RED MX log -6 stops and Alexa log -7 stops.

So, yes, it may be two good stops and a half in log.

In the REC709 test, Geoff favors again the high end of the Alexa. To my eye is less the one stop here.
 
Peter Moretti said:
Tony,

If I've understood Geoff's methodology correctly, that range of pictures accurately captures 14 stops dynamic range of the sensor.

So it does not matter what rating you give the camera b/c you're killing the image on both directions and counting how many half stops it took to do that. All rating the camera at 800 would have done is add a stop to the top and remove a stop from the bottom. But the total number of stops would have stayed the same, AFAICT.
Click to expand...


Thank you! that is precisely my approach.

The reason I do this, as I've already said is to compare the output of lots of different cameras in a way that will show relative differences, absolute sensitivity was not an issue.

Everyone uses a different setting to get what the feel is "right"
 
OK, I can download the R3D's by right clicking and using "save link as" however once I've downloaded them I can't get RedCineX to open them.

I used the button on the player to save them.

I really haven't got time to do all this again. I'll try and add the files that I didn't because I couldn't read them but I have the Canon, SI and Panasonic files to do and I also have a living to earn.

I do these tests for ME so that I can see what I need to know, the things that can cause me problems whilst I'm working.

They're very simple tests to do and anyone here can duplicate them.

They are totally non-biased because as I've said before I really don't give a damn what I shoot with as long as it will give me the images I need. I also need to know what is going to happen to a bunch of cameras that I may have to use.

If it's useful to you then great, if you don't like it then do your own tests.

The one thing you can rely on is that I'm fiercely independent and will not bend results any way for anyone.
 
I've now added the Canon 7D

The SI will have to wait becasue it seems that Cineform may have changed again!

Panasonic is waiting until I can get their software to work to allow me to play with the files, I can't install the drivers because I don't have a camera attached!! also when I did get the player to work on my laptop it would only show me files that were on P2 cards and not files that had been copied to hard drives.

Life is too short.

If some one will explain how to reliably get single frame R3D's out of REDCineX then I'll have another go!
 
Hi Mr. Boyle,

It was a pleasure meeting you at the ASC clubhouse opening.

My recommendation would be to archive all your current redcine x settings (which are in username/library/red)

Then trash the folder.

Then install redcine x v356.

Apply redcolor 2 / redlogfilm to your clips, then click r3d snapshot, or output your dpx stills.

Also, even though I'm not a canon guy - did you use 5dtorgb for your dpx?

Also would love to throw gh2 and af100 in the mix.

I'm still a red guy but I like my gh2 for
Pocket camerawork.
 
Dan Kanes said:
....
Then trash the folder.

Then install redcine x v356.

Apply redcolor 2 / redlogfilm to your clips, then click r3d snapshot, or output your dpx stills.

Also, even though I'm not a canon guy - did you use 5dtorgb for your dpx?
Click to expand...


Thanks, I think I'll just remove everything then run a reg cleaner then reinstall.

The Canon shots were made by loading the original QT files into Iridas Framecycler and then rendering out single DPX files without adding any corrections.
 
geoffboyle said:
Thank you! that is precisely my approach
Click to expand...

I understand your approach now; I must have misunderstood your method of doing this test. My bad, Geoff. I do appreciate you making the test public, but I hope you're able to address any issue there might be if RED thinks something has been done wrong in the post-side of things. :beer:
 
Tony Lorentzen said:
I understand your approach now; I must have misunderstood your method of doing this test. My bad, Geoff. I do appreciate you making the test public, but I hope you're able to address any issue there might be if RED thinks something has been done wrong in the post-side of things. :beer:
Click to expand...

Thanks, I've now removed and cleaned the system and reinstalled.

R3D snaps now work, I'll get them up in about 14 hours or so.
 
Thank you once again for your time and un-paid effort. I apologize for any technical misconceptions I may have. Someone please point them out to me if I have them.

Please understand Geoff that I speaking with utmost respect, as I always do when talking to artists. I hope everyone can forgo any preconceptions of "fan-boy" tone, if there are any.

geoffboyle said:
I do these tests for ME so that I can see what I need to know, the things that can cause me problems whilst I'm working.
Click to expand...

Saying this doesn't delete the weight your words carry Geoff. CML is a powerful megaphone; and having started it it is inconceivable you wouldn't have an idea of how many, and how much, people listen to you. Many with little technical know-how, as you pointed out and included in your target audience.

I haven't looked at the files yet. I am waiting for the R3Ds. But with different people agreeing that at least a few more Red One M-X frames should have been posted by the standards with which Alexa frames were posted, I can't see why you wouldn't kindly take a quick look and comment. I am inclined to believe them because those claims are closer to earlier DR measurements. The frames from your comparison which were posted on this thread seem to make it clear the same standards were not utilized, which only suggest they were not compared when choosing to stop at the top end --and do not suggest any conspiracy theories. I believe you are trying to be un-biased, and I am certainly not questioning your intentions. I think you most likely haven't found the time between uploading all other camera tests and a very busy work schedule to take a look at this and comment.

I apologize for my lack of patience if it feels that way, but your subsequent posts seem to indicate you don't intend to go back to upload the missing frames.

I want to get to the bottom of the DR difference. I thought it was settled, but I am shocked now by what your test suggests through the absence of so many frames. I more than willing to accept any DR ratio between Red One M-X and Alexa, so I hope you can find the time to let us know if there is something people are missing that supports the absence of those frames, then I will be able to look at your comparison and accept it as complete and the most current.
 
Did you preset your 7D with an Picture Style custom profile?

I've done a feature with the 5D mkII where the custom made Picture Style gives more natural colors and DR rendition. We have to treat the 5D mkII or 7D like in the old days the F900... with different setups as the factory preset to optimize the camera.

Only a few take the time and hussle to do so, wich is not in favor of these DSLR.

Patrick

PS: I'm not talking here of miracles with those presets.
 
They can work pretty well, I even wrote an article on this process I like to call "pre-grading". Follow these steps:
– shoot the scene in photo-mode as RAW
– open it in Canon's "Picture Style Editor" (free)
– grade as close as possible to the look you'd like for your scene
– save as preset and load into camera
– shoot video

The article is in German, but if you'd like a short summary and some screen shots, give me a PM.

Not as flexible as the RED workflow, but making much better use of the sensor than standardized presets.
 
Status
Not open for further replies.
Back
Top