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 :)

Full-frontal insidious attack on Red, by indie director Rian Johnson (2008)

Status
Not open for further replies.
I'm the one who used the words, "In cahoots with" and I apologise. It came from reading pages of posts and building momentum. Maybe a bit the same as ignorant folks coming to inacurate conclusions, or really not knowing what they're doing. Anyway, the end product is consumed by the senses; to what the human eye sees, is belief.
 
Tom said:
Keep in mind that not everyone agrees that Red One resolves to 3.2K. Many tests have pegged the visible resolution closer to 2K.
Click to expand...

With soft lenses, diffraction-limited apertures, or measuring just the red/blue channel, I'm sure that's true.

The whole idea of measuring sensor resolution is silly, anyway. Every Bayer sensor has the resolution of the number of megapixels. The only thing that causes them to vary is the optical system (including OLPF) and post-processing.

So when you compare two sensors for resolution, you're really just testing the lens, the optical anti-alias filter, and differences in post-processing, including compression such as REDCODE.

That's why it's so funny that all the review web sites and magazines do these "resolution" tests. They say one 12 MP camera has "more resolution" than another 12 MP, when in reality is just has a weaker (too weak!) OLPF.
 
Graeme Nattress said:
Tom, it's quite verifiable that luma resolution is visible out to around 3.2k. If someone can't see that in their measurement, it's their measurement that's just not good enough. You can also look at resolution tests of any DSLR out there. They all routinely measure their resolution to be a very similar percentage of the pixel count that we achieve.
Click to expand...

What percentage is that, Graeme?
 
Yes, anywhere in the 70% to 80% region or thereabouts. I have things measuring around 78% at the moment.

Graeme
 
-

-

No email hiring me to clean up this mess.

More people yelling about suing them for their opinion.

I found this (which was factually misrepresented HERE at least once):

"(All the technical specs and claims in this document have been checked both against RED's published specs and real world inspection of their footage. RED reps, if there are any factual errors in this document we'd love to hear about them, and we'll happily update it.)"

So, they ARE open to correcting any errors. Perhaps someone from Red will actually respond (which should have been sorted out, optimally, before publication).

The other thing is this:

"...you haven't actually evaluated the RED camera's results until you've seen them projected full quality on professional equipment.

We have. And under these conditions we could plainly see huge splotches of compression clumping and artificial-looking colors."

This thread almost completely ignores this claim, far from refuting it. And it drifts off into math-land, somewhere pretty irrelevant to the claim.

Everyone who reads the article is going to wonder if this compression "clumping" is a real problem or not. It is the main claim that must be challenged, substantiated, refuted, explained -- or however it plays out.

If the compression was an issue on a professional shoot, where the filmmaker's money, resources and reputation were on the line, then he has just as much at stake as Red's marketing department. And every right to complain about it, publicly or not.

Fan boy ad nauseum aside, this is a critical issue that cries out for attention.

I'm guessing Rian was on an earlier build where the "noise" was not so "smooth." But, MY approach is still the only reasonable approach out there: contact him with an offer to help him analyze the raw files and clean up the "clumping."

How many other readers will suspect Red's compression quality if this isn't addressed?
 
Well they can back up their claims about splotching and clumping by posting raw frames.
 
That post

That post

"All the technical specs and claims in this document have been checked both against RED's published specs and real world inspection of their footage. RED reps, if there are any factual errors in this document we'd love to hear about them, and we'll happily update it."

This thread...drifts off into math-land, somewhere pretty irrelevant to the claim.
Click to expand...

I have a major problem with that thought, as the original post being discussed is full of RED-ONE specification challenges. So its pretty hard to refute any specific claim in it except by telling it the way it is with real math and not the psuedo-math reported.

As for splotches and clumps.

If you have read this board for any length of time then you will know that we have discussed some that arose from codec errors during Build 16 alpha testing, and we've also discussed some arising from bad CF card readers. But until we see a TIFF from their shoot we are in the dark as to what the potential cause of their issue is.
 
Neither Rian nor his "buddy" are remotely qualified to be spouting anything about RED good or bad but in the Indie world people will listen to (and usually worship) anyone who is even remotely famous so this definitely needs to be addressed. When their professional and technical experience with various camera systems deepens a bit, then I'd consider what they have to say but having gone over their resume's, I can clearly see why they are dumping on RED given how radically different it is from the 35mm cameras they have been using and how that difference (as we have seen here in some other threads) usually translates to disappointment whenever someone who actually knows how to use the camera isn't around to explain the differences to them.
 
Splotches and clumps? Is he talking about film grain? :)

With as much complaining and whining that goes on here, there really isn't much fanboyism any more. Just guys shooting, getting on with it and loving Red for very valid reasons.
 
people who have problems, most likely dont know how to work the camera/or the work around post workflow.

his dp or operator obviously didnt know how to work the camera.. no big deal.
should have hired me:)
 
compression and Hollywood

compression and Hollywood

A question : As I understand, RED One can be used uncompressed as well. Are the movies out there based on REDCODE or have they been captured uncompressed? "Knowing", for instance?


While some of the statements of Rian Johnson sound naïve, I do sympathize with the view that "full 4K" is misleading, given that each frame weighs 1MB. Anybody out there claims that the stills you can grab out of REDCODE footage compete with 12MPix DSLR stills?

That does not contradict REDCODE to be sufficient for its purpose, which is not to produce 24 stills per second, but a movie.
 
There is no uncompressed output. Everything you've seen, including Knowing has gone through the visually lossless REDCODE RAW compression scheme.

You may sympathize, but I'd not bother. Fact is, you record a zone plate designed for measuring 4k resolution on the RED, you extract a still, pop that into ImageJ and measure off the resolution. The answer is, as we'd expect with the OLPF (to avoid aliasing) and the demosaicing process, just like in a DSLR, 3.2k resolution, or thereabouts.

The frame size is not 1MB. It's nearer 1.5MB. REDCODE RAW is an advanced wavelet compression scheme. It's not an old, block based DCT scheme.

I know Jim has printed stills from the RED on his wide format printers and they looked great. No reason they should not!

Graeme
 
Graeme Nattress said:
REDCODE RAW is an advanced wavelet compression scheme.
Click to expand...

Thank you, I understand better the magic now, exiting indeed. Cool.

Have I missed something or is there currently no standard compression allowing von 16bit and ICC profiles and all that that uses wavelet compression?

Darn, I'll have to wait for your DSMC announcement announcement. Too bad for Canon.
 
A lot of words, and no proof of anything. And still a big bundle of factual errors, especially about raw. Like "Red didn't build it on-board and they can't do it in real time." - like duh, how do you think you see a viewfinder image, or how do you see the HD-SDI output?

The fact is, and has been proven, and can be shown that there are irriversable transforms applied to the image in it's journey from raw sensor data to a "nice" image, that cannot be undone.

Then we have this amazing flip. Previous the claim was that the "other cameras" have more photosites than RED. Now this claim, that they have less.... "On top of all of this, it is important to note that fewer total photosites (which F23 and Genesis have compared to Red) can sometimes be an ADVANTAGE in reducing noise and increasing dynamic range and sensitivity to light. ". Even though I did the math before, we can do it again...

F23 is 1920x1080x3 = 6,220,800
RED (in 4k16:9) = 4096x2304 = 9,437,184
Genesis is 1920x1080x6 = 12,441,600

So, before it was better to have more photosites when he thought RED had less. Now he thinks RED has more, so it's better to have less photosites. Which is it? Now we know the Genesis has more photosites than RED, so is it's "spec" worse of better....

I think there's only one spec that matters - what the name badge says. Everything else is bullshit.

Graeme
 
I now notice Arri lacks a mention. But Arri shoots raw like RED. Arri uses a Bayer pattern, like RED, Dalsa, SI, every high end DSLR out there (including Sony's).

Graeme
 
Status
Not open for further replies.
Back
Top