Red - it's grotesquely incompetent

[Adrian Sierkowski]

Dunno; I think a lot of it has to do with people feeling marketed to. A lot of it has to do with the moon-promises which were made, and the misunderstandings. A lot of it has to do with people. I'm not one to name names, or point fingers, or anything of the sort. Not my place. People are just people and they will always disagree, on everything. In fact, agreeing that we always will disagree may just be the one real agreement there is! (thank you post-modernity)

[Radoslav Karapetkov]

Dunno; I think a lot of it has to do with people feeling marketed to. A lot of it has to do with the moon-promises which were made, and the misunderstandings. A lot of it has to do with people. I'm not one to name names, or point fingers, or anything of the sort. Not my place. People are just people and they will always disagree, on everything. In fact, agreeing that we always will disagree may just be the one real agreement there is! (thank you post-modernity)

 

M-hm.

 

Pople are people... (and so on).

[Walter Graff]

If RED is such a plague, people can just ignore it, no hard feelings are necessary.

 

Because people take life personally. Imagine if you didn't like a model of a car such as the new Chevy Malibu? Most folks would simply look for a car that did fit their needs. Could you imagine people spending their free time constantly putting it down? It serves no purpose except selfish needs. But as I said in my earlier posts about marketing of RED, when you open yourself up to the public, some take what you say personally, can not realize that life is not fact, nor are their feelings correct, just their own, and the focus is lost. Recently Hyundai ran a TV campaign touting a new model as half the price of a BMW or Mercedes, but equal in style, ride, and comfort. And some folks tried it and realized it was true. And others thought it was good, but not in that caliber. And others simply said it couldn't be. And a few spent a lot of time on websites writing nasty emails taking it very personally and being very hateful. All different ways of reacting.

[Brad Grimmett]

This is not the place to discuss religion, Jesus or the Bible.

Actually, I think quite a few people would like to compare Red to Jesus. But then again, a lot of people don't believe Jesus is the savior....

[Tom Lowe]

I've always thought RED would basically be a 2K camera, because pixel for pixel it will only hold up once it's been downsampled. This is true of CMOS DLSR cameras with similar-sized chips.

[Keith Walters]

I think this proves the technical point rather elegantly:

 

http://gunshotfx.com/red/

 

There is only the tiniest, most microscopic and insignificant resolution increase in the direct "4K" over the resampled version. Ergo, it's not a 4K camera - it's not even a 3K camera. It's right there for everyone to look at.

 

Needless to say, marketeers are rarely dissuaded by mere facts.

 

Phil

I hate to rain on your parade, Phil, and I would not want anyone to think I've drunk the Red soda pop, but are you SURE that picture was focussed correctly?

The thing is, if you look at most of the girl's face, it's true you don't see too much difference.

But have a look at the reflection in her eyes.

To me it looks like you can clearly see the iris blades in the original 4K image, but they're distinctly degraded in the down-rezzed/up-rezzed version.

Because they're reflecting off a curved surface, highlights like that aren't subject to the same focussing rules as everything else, behaving like they are imaged by a very short focal length lens and so they tend to remain in focus regardless of the setting of the main lens.

[Sam Wells]

I've always thought RED would basically be a 2K camera, because pixel for pixel it will only hold up once it's been downsampled. This is true of CMOS DLSR cameras with similar-sized chips.

 

The Dalsa Origin footage I saw projected at 4K, similar sized chip, I'd say it held up.

 

The motion work I'm doing with the Nikon D3 - larger chip but similar pixel (photosite) count - is getting downsampled to 1920x1080 but that's 'cause I ain't got the computer horsepower to work in 4K - but I do hope to do a short test and see what happens --- I don't have a 4K projector at hand either, I'll have to find some co-conspirators ---

 

-Sam

[John Sprung]

--- I don't have a 4K projector at hand either, I'll have to find some co-conspirators ---

One work around for testing only is to crop out of your 4K image a chunk that matches the grid of your projector. Maybe divide it up into, say, four pieces that you could look at one at a time. As long as the pixels are mapped one to one, you're seeing the resolution you really have, just not the whole picture at once.

 

 

 

 

 

-- J.S.

[Phil Rhodes]

> are you SURE that picture was focussed correctly

 

No.

 

As soon as Red let someone perform some objective technical evaluation of the thing, I'll be happy to look at it. Until then, they're going to have to get used to people drawing what conclusions they can from the information that's available.

 

P

[Keith Walters]

> are you SURE that picture was focussed correctly

 

No.

 

As soon as Red let someone perform some objective technical evaluation of the thing, I'll be happy to look at it. Until then, they're going to have to get used to people drawing what conclusions they can from the information that's available.

 

P

 

Yes, past experience with geting real information from other manufacturers was a bit like the demolition notice for Arthur Dent's house in The Hitchhikers Guide to the Galaxy. It was on display for everyone to see, kept on display in a locked cupboard in the Council basement with the stairs removed and sign on the door: "Beware of the Leopard"! :lol:

 

Always delightful was the splendid and seemingly endless array of inventive excuses that fanboys and other apologists come up with, as to why they were so reluctant to allow unsupervised testing, reminiscent of the Monty Python Cheese Shop Sketch. Occam's Razor always suggested that the most likely explanation was that there was a damned good reason why they were reluctant to give out too much information!

 

Hence my current project for a battery-operated 15-stop LED array. I have placed the order for matched brightness LEDS, now I am working on getting my $50 lux meter calibrated to match a $2,000 one. Strangely, owners of various cameras do not seem overly interested in participating in any tests...

[Daniel Sheehy]

..Strangely, owners of various cameras do not seem overly interested in participating in any tests...

Pity, as your little project has interesting potential I think.

[Sam Wells]

One work around for testing only is to crop out of your 4K image a chunk that matches the grid of your projector. Maybe divide it up into, say, four pieces that you could look at one at a time. As long as the pixels are mapped one to one, you're seeing the resolution you really have, just not the whole picture at once.

-- J.S.

 

Interesting idea, John !

 

(Be interesting to directly compare with Red actually. Do they have their Nikon mount available yet ?)

 

-Sam

[Jake Kerber]

Keith, if you're in the L.A. area I might be interested in getting your LED fixture involved with some RED testing I'm planning to do.

 

Also, for what it's worth, it sounds like Phil played with the camera without recording and 'processing' any of his own images. If he was just looking at the 720P output (that I think uses REC 709) then I'm sure some of the colors were off and overall the image wasn't truly representative.

 

Regarding the footage online: Much of it is shown without details as to the lenses used and how it was 'processed.' I'm still learning and testing this camera, so I don't have all the answers, but it appears that because of the many ways people are cooking the 'RED RAW' in post, people are seeing varying degrees of quality posted online (which, of course, isn't the most ideal place to view any camera's images). And everyone here knows how much difference good glass can make, so I won't even go there.

 

I'm hoping to do some more conclusive testing to see which of the various artifacts I've seen are related to the camera and its sensor versus the post-production processes used on the images--something anyone considering using the camera should do themselves. Don't rely too much on second hand info.

 

-Kerber

D.P

L.A.

[Keith Walters]

Keith, if you're in the L.A. area I might be interested in getting your LED fixture involved with some RED testing I'm planning to do.

Sorry, I'm located in the Sydney, Australia area!

 

Rumour has it that there are somewhere between 50-70 REDs either in Sydney, or on order. However nobody I know has access to one!

 

Also, for what it's worth, it sounds like Phil played with the camera without recording and 'processing' any of his own images. If he was just looking at the 720P output (that I think uses REC 709) then I'm sure some of the colors were off and overall the image wasn't truly representative.

 

Regarding the footage online: Much of it is shown without details as to the lenses used and how it was 'processed.' I'm still learning and testing this camera, so I don't have all the answers, but it appears that because of the many ways people are cooking the 'RED RAW' in post, people are seeing varying degrees of quality posted online (which, of course, isn't the most ideal place to view any camera's images). And everyone here knows how much difference good glass can make, so I won't even go there.

While post-processing is indeed important at the end ot the day, I am not particularly interested in that; all I really care about is what useable data is actually captured by the CMOS sensor. Arguments about Post-Production "massaging" have more validity with film capture, because with that, the point where part of an image is considered unworkable due to overexposure is very much determined by what is actually overexposed, how large the area is, what surrounds it and so on. Overexposure on somebody's face will be less acceptable than the same level of overexposure on an overcast sky for example.

 

With an electronic camera, the transition from useable to useless is much more abrupt, and there is absolutely nothing post-production can do to restore data from overloaded photosites. (Apart from drawing them by hand :lol: )

 

I'm hoping to do some more conclusive testing to see which of the various artifacts I've seen are related to the camera and its sensor versus the post-production processes used on the images--something anyone considering using the camera should do themselves. Don't rely too much on second hand info.

 

-Kerber

D.P

L.A.

 

The most foolproof test would be as follows:

Make up a wooden box that can hold two or three cameras, with a glass window on one side. Paint the inside black and fit a device constructed like an oversize mattebox that can hold wooden frames covered with sheets of ND gel. Make an assortment of such frames with one stop, two stop and three stop ND gels

 

If there is a film camera involved, load it with your choice of film, and set any video cameras to the same effective ASA rating. Put the same type of lens on whatever number of cameras you have, take you box outside in the daylight and with a light meter find out what the iris should be set to on the contrasty scene of your choice. (Unless fast film is used, it will need to a bright sunny day).

 

If you fit all three gels and open up the iris six stops the cameras should be correctly exposed. After that you start them all running and progressively pull out the gels.

 

the combinations would be

1+2+3 = 6 Stops

2+3 =5 Stops

1+3 = 4 Stops

3 =3 Stops

2= 2 Stops

1=1 Stop

0=0 Stops

 

Process and telecine the film, do your Red Posting etc, and compare on three identical monitors!

[Guest Josh Gannon]

Rumour has it that there are somewhere between 50-70 REDs either in Sydney, or on order. However nobody I know has access to one!

 

 

Keith, if you would like to see a Red on a production demonstration basis just call Camera Collective. Red's have been available in Sydney for about 3 months now.

 

 

 

Josh.

[John Sprung]

While post-processing is indeed important at the end ot the day, I am not particularly interested in that; all I really care about is what useable data is actually captured by the CMOS sensor. Arguments about Post-Production "massaging" have more validity with film capture, because with that, the point where part of an image is considered unworkable due to overexposure is very much determined by what is actually overexposed, how large the area is, what surrounds it and so on.

 

Unfortunately, post processing is necessary to know what useable data you have. Much like film has to be run through chemical processes before you can see anything, the Red data has to be run through offboard computer processes before you can see what you really have at its full final quality.

 

The output you have for on-set viewing is much like a film camera's video tap (perhaps a deceptively high quality video tap). Bottom line, Red is not a TV camera. Misunderstanding that point may be the source of a lot of disappointment.

 

 

 

 

-- J.S.

[Walter Graff]

Bottom line, Red is not a TV camera. Misunderstanding that point may be the source of a lot of disappointment.

 

So where does the film go then?

[Jake Kerber]

Ultimately, for us Cinematographers, we care about what the final image looks like, right.

 

So, just like you would do a film stock test--applying any photochemical or digital treatment that is being discussed as part of the look and taking the material through the whole workflow--before shooting a feature on 35mm film, one would/should do the same before using an electronic camera. Especially the RED, as it's designed to record the RAW (yes, I know wavelet compression is applied) information from the sensor, allowing one more freedom to process the data as they wish to achieve the desired look.

 

Any D.P. who wants to retain as much control as possible over the final look of their photography should, IMHO, be knowledgeable and concerned about the post-production treatment of the material (whether the material originates on film or a digital medium). That's part of the D.P.'s job--especially in this day and age when data flows and image manipulation

seems at the fingertips of many throughout the process. Of course, hopefully all involved are on the same page regarding the look and there's no need to be concerned.

 

-Kerber

D.P.

L.A.

[John Sprung]

So where does the film go then?

 

Um, OK....

 

Bottom line, Red is neither a TV camera nor a film camera.

 

It's something new and different for which there isn't a generally accepted and practically brief name. Compressed raw digital data camera? Nah -- way too many words. Just plain "digital camera"? Nah -- there are already a lot of TV cameras that use that name.

 

 

 

 

-- J.S.

[Robert Houllahan]

Um, OK....

 

Bottom line, Red is neither a TV camera nor a film camera.

 

-- J.S.

 

 

Data Camera? Like Origin and SI or Phantom where there is no immediate way to get "useable" video out?

 

-Rob-

[Walter Graff]

Bottom line, Red is neither a TV camera nor a film camera.

 

Sort of Like the candy Razzles? Razzles sold themselves asking the question, is it a gum or a candy. Turned out both, first it was a candy then a few chews later it turned into a gum...brilliant!! :)

 

 

http://candy-crate.stores.yahoo.net/razzles1.html

[John Sprung]

Data Camera? Like Origin and SI or Phantom where there is no immediate way to get "useable" video out?

 

-Rob-

I think you've got it. Let's start calling all of them Data Cameras, and see if we can make the name stick.

 

 

 

 

-- J.S.

[Keith Walters]

Keith, if you would like to see a Red on a production demonstration basis just call Camera Collective. Red's have been available in Sydney for about 3 months now.

Josh.

Hi Josh,

 

I know of people in Sydney who have REDs, and other HD video cameras as well, but in my experience they aren't too keen to allow unsupervised testing, presumably in case I tell them something they don't want to hear!

 

For what it's worth, I don't expect the RED to be as bad as some people have claimed, and I would really like these people to explain how they arrived at the conclusions they did. Accurately measuring dynamic range is not something that can be accurately done without specialized equipment.

 

On the other hand if the RED really can handle 11 stops, my tests should show this unequivocally.

 

So far all I have done is a rough "proof of principle" test based on a simple setup with two brightness-matched ultra-bright white LEDs. One (the "O stop" reference), has its maximum rated 20 milliamps flowing, and this is what the iris is set from.

 

The second LED then had its brightness progressively reduced until its image became unuseable. Since LEDs become non-linear at very low current levels, to dim the second one I used a mixture of NDs to drop it down six stops (for the "coarse" setting) and then reduced the current from its initial 20 milliamps, in steps down to 10, 5, 2.5, 1.25 and 0.625 milliamps.

 

This gives a maximum 11-stop difference between the two LEDs.

 

All I am prepared to say at this point is that no video camera I have tested (HD or otherwise) could display both LEDs at the same time with that attenuation, or go anywhere near it. As far as I can determine, on 11 stops, the dimmest LED really was 1/2048th as bright as the reference one, but I will soon be doing more sophisticated tests to confirm this.

 

Nobody else appears to do the tests this way, but I think the reason is simply that white LEDs are a relatively new device, and people don't appreciate that they are not like ordinary light bulbs, in that their light output can be very precisely controlled, without any change to their colour temperature.

[Keith Walters]

Unfortunately, post processing is necessary to know what useable data you have. Much like film has to be run through chemical processes before you can see anything, the Red data has to be run through offboard computer processes before you can see what you really have at its full final quality.

 

The output you have for on-set viewing is much like a film camera's video tap (perhaps a deceptively high quality video tap). Bottom line, Red is not a TV camera. Misunderstanding that point may be the source of a lot of disappointment.

-- J.S.

Are you saying that if something is white-clipped on the live monitor output, Post is somehow going to magically bring that back?

 

It there's white clipping on the video tap of a film camera, an experienced operator should be able to judge how bad it will be on the film, and unless it's really severe, Post production will as like as not be able to pull something out of it.

 

The same thing most definitely does not apply with a video camera, as many a newbie operator has found to his cost.

[John Brawley]

So far all I have done is a rough "proof of principle" test based on a simple setup with two brightness-matched ultra-bright white LEDs. One (the "O stop" reference), has its maximum rated 20 milliamps flowing, and this is what the iris is set from.

 

The second LED then had its brightness progressively reduced until its image became unuseable. Since LEDs become non-linear at very low current levels, to dim the second one I used a mixture of NDs to drop it down six stops (for the "coarse" setting) and then reduced the current from its initial 20 milliamps, in steps down to 10, 5, 2.5, 1.25 and 0.625 milliamps.

 

 

Have you thought about using pulse width power for finer control of the brightness and linarity issues ?

 

Also, what kind of ND's are you using ? I doubt they'd be super duper accurate for this kind of test...especially if it was lighting gel.

 

jb