Welcome to our community

Be a part of something great, join today!

Register 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 :)

Red vs Genesis vs Alexa

Just so we all know, F35/Genesis is a single CCD with a colour filter array. It's a different pattern to Bayer, but it's still a colour pattern. It's not 3 chip and it doesn't read RGB direct as implied above. The image I linked to above shows that quite clearly as you'd not get the rainbow moire if the samples were co-sited. So to say 30bpp is bollocks. The output file format is 10bits per channel. And talking bits is actually irrelevant - what matters is measured noise or dynamic range.

I think the reason that "article" doesn't get much publicity today is that whatever they said about RED applies even more so to Alexa (and the D21 and Dalsa was ignored) because to bring them in didn't fit in with the anti-RED message because D21, Alexa and Dalsa are all Bayer pattern and the Alexa typically records to a compressed codec.

And John Galt from Panavision made big things of "marketing pixels" in his criticism of RED, because we got a measured 3.2k resolution from ~9mp array. But when Sony claims 8k resolution from a 20mp rotated Bayer array it gets no criticism or mention....

Graeme
 
Thanks Graeme

It is an old and outdated argument, as is the article, but as a newbie to this forum, and indeed to RED, it was what I had been told by other camera technicians, without having ever had the opportunity, or knowledge, to debate, or contest the claim.

I had the good fortune to play with the EPIC today, and from first impressions, what an amazing piece of technology.
 
Tim, unfortunately there's a lot of mis-information about a lot of different camera systems, not just ours. Some of it is from a lack of understanding on how imaging systems work. To address that issue I recommend Alan Roberts' book "Circles of Confusion" published by the EBU. It goes into more detail than you need, but covers most everything.

As Jim says, when we started RED we knew nothing. Since then we've learned a lot, and so much of that knowledge is empirical, and science backed. The camera systems we make are proof enough that we know what we're talking about and we put our money where our mouth is. So ask away, and learn along with us. There's a lot of people on this forum that will tell you how much they've learned about cameras in general from reading the posts here, not just fro me, but from other RED people and our users too. We're pretty open about our biases, but that doesn't stop us giving the right answers.

Graeme
 
I often hear that the color processing of the Alexa is better than the RED camera... More film like, organic, better skin tones, etc. now when I checked this page on ARRI... http://www.arri.com/camera/tutorials/alexa_color_processing.html#

I found out as I long knew that this color processing would be mainly used in REC709 mode. The thing is that most features shot on the ARRI Alexa have been shot not in REC709 processing but linear with ARRIRAW. ARRI themselves state this:

"The Rec 709 output of the camera is either used for on set preview or when the program is edited for television without extensive color correction. Those images are displayed without any further transformation. While this eases the post production it reduces the possibilities in color correction. The images have been tone-mapped and transformed into the target color space."

Now if people can understand basic color processing and what it does to their feature they will understand that this mode is underisable. Now I can see if your going straight to television and don't have a colorist or can't afford ARRIRAW capturing which is very expensive. I just get so irritated when people say stuff like Roger Deakins is shooting Bond on an ARRI Alexa without realizing that the movie is shooting in ARRIRAW with no color processing applied. So the color science of ARRI does not come into play at all.
 
Andrae Palmer said:
I often hear that the color processing of the Alexa is better than the RED camera... More film like, organic, better skin tones, etc. now when I checked this page on ARRI... http://www.arri.com/camera/tutorials/alexa_color_processing.html#

I found out as I long knew that this color processing would be mainly used in REC709 mode. The thing is that most features shot on the ARRI Alexa have been shot not in REC709 processing but linear with ARRIRAW. ARRI themselves state this:

"The Rec 709 output of the camera is either used for on set preview or when the program is edited for television without extensive color correction. Those images are displayed without any further transformation. While this eases the post production it reduces the possibilities in color correction. The images have been tone-mapped and transformed into the target color space."

Now if people can understand basic color processing and what it does to their feature they will understand that this mode is underisable. Now I can see if your going straight to television and don't have a colorist or can't afford ARRIRAW capturing which is very expensive. I just get so irritated when people say stuff like Roger Deakins is shooting Bond on an ARRI Alexa without realizing that the movie is shooting in ARRIRAW with no color processing applied. So the color science of ARRi does not come into play at all.
Click to expand...

Strange conclusions.

;ost of what I know from the Alexa is shot in prores 4444 LogC and is definitely colorprocessed....
I am not too fond of those prores files, though, but that is another story.

Higher budgeted things are shot in ArriRAW, but the same processing does apply there, pretty much like when you go from RedRAW to RGB RedlogFilm.

The processing can happen in linear, but that doesn't mean that colorscience isn't applied.

Then, to your or others points on skintone...
I think as RED allways captures RAW, that may sometimes have the downside that people are on their own. I think skintones on the Epic is vastly better than on the R1 MX, mostly due to more data, less compression, better signalchain and processing capacity overhead....

And I really like the skintones on the R1 M, too...

And I think that matching Alexa prores and R1 MX is quite doable.

But I think it is strange to use ArriRAW against Alexa. That's where I think the camera really starts to be very good...
 
If what you are saying is true then how comes these charts are so way different: http://archiv.arri.de/AlexaColorProcessing-23Feb2011/html/alexamodes.html

LogC both with film matrix on and off looks like the color science on any current RED MX or Epic. Rec709 on the other hand is vastly different with more agressive color processing. I'm betting that most shooters with the Alexa are shooting in REC709 mode.
 
Around me, I have not seen any, but that may be a local issue.

rec709 on the Alexa is very similar to RedGamma on R1 MX in top and bottom, and you are right, it is very aggressive and pushes away close to 4 stops of latitude.

The FilmMatrix a stupid and evil thing that should allmost never be used. very bad for you and your friends.

LogC is very similar to RedLogFilm for the simple reason that they are both variations over the CineonLog description. That makes a lot of sense. But it's not unmatrixed. It's just log.

But back to RED... We're at reduser... :)
 
Well as RED owners and operators we often have to defend our camera system so to speak from people that don't know what they are talking about. I think it's good to be armed with some information. Especially when this argument is the main one being used against the Epic or RED One.

From what I'm researching here... ARRIRAW is "untouched" Bayer pattern image data:
http://forums.creativecow.net/thread/294/358
 
AFAIK, Log C was unmatrixed in original release, but later got a matrix.

According to Arri, this is the equation for reversing LogC: { (t > 0.1496582 ? pow(10.0, (t -0.385537) / 0.2471896) : t / 0.9661776 - 0.04378604) * 0.18 - 0.00937677 } (not sure for which ISO?) And that's not pure log either....

But if you look at the spec for cineon, this is the reversal formula: (pow(10,(1023*x-685)/300)-.0108)/(1-.0108) , and that's also what Nuke uses to linearize a cineon file to linear light. We use the same because, well, that's standard, and works very well, and it utterly compatible with Nuke and other software that operates in linear light.

As for compression on ArriRAW: Arri claim 14bit AtoD and 16bit processing, but somehow that fits over a 12bit pipe. Methinks there's some compression going on - what do you think?

Graeme
 
Last edited:
"In order to create images for a DI workflow, the image data from the sensor is debayered and then given a Log C gradation (= Cineon-style logarithmic gamma), which is very similar to the gradation of scans from a film negative. No color matrix is applied, so the images stay in the native color space of the ALEXA sensor. The combination of the Log C gradation and the native ALEXA color space make the images look flat and desaturated to the naked eye. They need a Look Up Table (LUT) applied so the director can view them on the set or in post without experiencing a visual crisis. Luckily, for viewing on the set, ALEXA is capable of outputting a Log C (no matrix) signal on the REC OUT, while at the same time outputting a Rec 709 image for viewing on the MON OUT output. A Log C image with no matrix leaves the images as unprocessed as a full color HD image can be, offering the widest range of color correction choice to the colorist in post."

Source: http://www.abelcine.com/articles/in...=52:digital-cinema&Itemid=100003&limitstart=8

Of course the images will look better when you are getting a heavily processed REC709 preview while recording in LogC.
 
Graeme Nattress said:
AFAIK, Log C was unmatrixed in original release, but later got a matrix.

According to Arri, this is the equation for reversing LogC: { (t > 0.1496582 ? pow(10.0, (t -0.385537) / 0.2471896) : t / 0.9661776 - 0.04378604) * 0.18 - 0.00937677 } (not sure for which ISO?) And that's not pure log either....

But if you look at the spec for cineon, this is the reversal formula: (pow(10,(1023*x-685)/300)-.0108)/(1-.0108) , and that's also what Nuke uses to linearize a cineon file to linear light. We use the same because, well, that's standard, and works very well, and it utterly compatible with Nuke and other software that operates in linear light.

As for compression on ArriRAW: Arri claim 14bit AtoD and 16bit processing, but somehow that fits over a 12bit pipe. Methinks there's some compression going on - what do you think?

Graeme
Click to expand...

I was wondering about the 12 bit issue too. In general though i think the color charts speak for themselves. REC709, DCI-P3 looks filmic but is heavily processed throwing away of a lot of latitude... ARRI basically color grades it for you not knowing what your intention is. LogC looks like footage that we are all used to on the RED. The question Graeme now is this... does RED want to take the route of ARRI and have a heavily processed REC709 color space?
 
We believe that RAW gives the end user the best image and the most flexibility to post the way they want to post. It also gives us the flexibility to innovate.

Graeme
 
Graeme Nattress said:
We believe that RAW gives the end user the best image and the most flexibility to post the way they want to post. It also gives us the flexibility to innovate.

Graeme
Click to expand...


Hear, hear! But from a marketing point of view I see how RED can take a few hits from the Apple ProRes REC709 instant color grading crowd. Only God knows why someone would want to pay more for a Apple ProRes camera. Especially when the Epic with affordable third party devices can do the same PLUS native RAW. Having a more aggressive REC709 color space could erase some of the arguments.
 
Andrae, for a more aggressive space, set contrast to 0.5, saturation to 1.2 and be done. That's how I leave my Epic set for my use and it works well for monitoring.

Graeme
 
Graeme Nattress said:
Andrae, for a more aggressive space, set contrast to 0.5, saturation to 1.2 and be done. That's how I leave my Epic set for my use and it works well for monitoring.

Graeme
Click to expand...

You see Graeme... I personally know how to process RED footage to get a great look. I'm referring to people that don't read and like to press one button to get the same default "filmic" look.
 
Sorry, we make you press 2 buttons.

Graeme
 
Graeme Nattress;683664 I wasn't impressed on the F35 from the available charts from CML. With better charts I could be more confident on that. Arri basic image seems to be very under-saturated though said:
I'll try and fix that in January/February.

Intending an Alexa, Epic, F65, C300, F3, F35 testing session.
Click to expand...
 
Geoff, I seem to remember it being hard to acquire F35 charts at the time, so you could only publish what you could get. There was a pattern noise issue on them, and severely clipped blacks if I remember right - could easily have been an incorrect camera setting or something.

Graeme
 
You must log in or register to reply here.
Back
Top