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 :)

How did Claudio Miranda's test go?

Status
Not open for further replies.
Finner said:
Ruairi

There is an easy way to deal with it. Just click on Sanjin's Avatar and then click on the add him to your ignore list. Then bingo all of his posts disapear. A lot of us wish reduser would just kick him off for good but for now I have found putting him on my ignore list as the best option.
Click to expand...

Ah ! Thank you, thank you ! I did not know we could do that !
When I think of all wasted time I had with graphomaniacs !
 
Finner said:
Ruairi

There is an easy way to deal with it. Just click on Sanjin's Avatar and then click on the add him to your ignore list. Then bingo all of his posts disapear. A lot of us wish reduser would just kick him off for good but for now I have found putting him on my ignore list as the best option.
Click to expand...

Wow, you're right that works !

That's like magic,

Thank you Finner !
:biggrin:
 
Redrum, you forget though that all cameras with all kinds of sensors need low pass filters to avoid aliasing. It's especially important with bayer sensors as they will chroma alias if you don't. That means no camera should measure it's true resolution, but should always produce a lower figure.

Graeme
 
Back on topic?

Back on topic?

Hopefully this post doesn't get lost in the flame war. I notice that nobody's discussing the following:

Rolling Shutter. On the footage I have shot. When the camera hit some rough ground (heavy vibration), the buildings appeared to expand and collapse. This could be a problem when shooting effects with fast moving objects.
Click to expand...

I've seen rolling-shutter-related artifacts with photo flashes first hand, but I've yet to witness what claudio describes above. This sounds like a potentially serious issue, assuming "heavy vibration" means something one might experience by just driving around in a car. Would a similar artifact occur during extremely fast pans, etc?

Apologies if this has been covered elsewhere.

mitch

p.s. - don't feed the trolls.
 
Graeme Nattress said:
Redrum, you forget though that all cameras with all kinds of sensors need low pass filters to avoid aliasing. It's especially important with bayer sensors as they will chroma alias if you don't. That means no camera should measure it's true resolution, but should always produce a lower figure.
Click to expand...
Right, got caught up in the demosaic to the neglect of the other factors (my bad) but yes every system needs to have some manner of low-pass to cut off the highest frequencies, and that means they all register lower than their rated pixel counts.

I still say Red's going to be sharper than anyone else currently out because you start out with the most information so even though things like demosaicing are going to chop into your overall total, you're still going to have more left over at the end. Has Red stated any actual performance results for sharpness? My guess is 3K. Maybe if you someday make a super-Red with three Mysterium sensors and a beamsplitter, then you could have raw 4K @ 4:4:4, which might even approach the limit of human vision to resolve detail depending on how large your projection is?
 
meeotch said:
This sounds like a potentially serious issue, assuming "heavy vibration" means something one might experience by just driving around in a car.
Click to expand...
It's called "wobble" and it plagues little consumer-cams, might not be such a big issue with Red because it's so heavy and isn't likely going to be put in the same scenarios.

Would a similar artifact occur during extremely fast pans, etc?
Click to expand...
No, fast pans will make images lean. The wobbly jello effect happens in vibration. No idea how much the Red might be affected by it.
 
Graeme Nattress said:
Thanks David :-)


Making the best possible image is not easy. It's never been easy, and if it were, where would the art be then? And you can't just rely on art, you've got to experiment, and apply science and mathematics (yes, ouch, you can't ignore them) and learn and keep learning, and keep shooting. That's why I never stop thinking of new ways to get better images from our sensor.

Graeme
Click to expand...

Dear Graeme, even with those bad comparing conditions, one thing only bother me a little. Have you seen the night shots with high buildings after sunset ?
Well, it's a detail but I found the red lights at the back of the cars were (on the shot made with RED)
very "white", dont you think so ?
I do not have this trouble with CANON 20D.
Do you ?
 
Issue with any kind of prism is getting lenses for the camera - especially for S35 sized sensor.

I don't think it's easy to see the full resolution a RED can produce now, unless you're sitting very near the screen. That's based upon what I saw at our IBC screenings and back then the demosaic was not as advanced as it is now, and getting better all the time.

I think we've said we're about 75% measured resolution, or just over 3k. That can improve to an extent, and the perception of it can improve too, but we're getting very close to the amount of detail the OLPF is letting through, and I'm very wary of the effects of aliasing on motion....

Graeme
 
RedRum Thanks a lot.

I was waiting for these kind of input: raw 4K @ 4:4:4.

P.S. glad to see that all those X, Y,...are grouping again.:)
 
Jean Luc Leon,

place-charles-de-gaulle.jpg


hope to meet you here soon and "discuss" all about the RED.:ninja: :clown2:


RED = Nouvelle Vague
 
Graeme Nattress said:
Issue with any kind of prism is getting lenses for the camera - especially for S35 sized sensor.
Click to expand...
For mortals, maybe. But for you guys? Surely some genius solution is underway... :shiftyph34r:

I think we've said we're about 75% measured resolution, or just over 3k.
Click to expand...
Okay, so about what I thought... that makes sense, and it's still leaps beyond 2K.

That can improve to an extent, and the perception of it can improve too
Click to expand...
Which is the really cool part of recording the raw data is that you wizards can even improve the look of already recorded footage!

Thanks for the corrections and the answers. Good stuff.
 
The cameras lattitude comparisons turned out as I had thought.

Sony advertises over 12 stops but most rate it at 11 or so useable stop range.

Every high level DP that has shot red has placed it in the 9.5 or so stop range.

From the compareable stills on Claudios site I would judge the f23 about 1.5-2 stops better. The color of the f23 seems to hold better also. I am interested to see how the new shipping reds test out.

Problems I see.


1. The red dot flares are a problem and though some people that have never even touched a camera say that it has been fixed I have never actually seen any red employee post that the camera no longer does it (I could be wrong on this as they may have). A comment from a red employee on if the camera no longer does this would be good.

2. The black line vertical flair I do believe that I read somewhere that a red employee said that new cameras no longer do this. It would be great to get confirmation on that.

3.
Rolling Shutter. On the footage I have shot. When the camera hit some rough ground (heavy vibration), the buildings appeared to expand and collapse. This could be a problem when shooting effects with fast moving objects.
Click to expand...
This is a quote from Claudios test page and I have not seen this before so it would be good to get an example of the problem and information on if the new cameras no longer do it?
 
Why don't we just wait a few days and test the new cameras...? Then it won't matter what anyone says.

Jim
 
A new test would be great to see Jim.

Any chance you could answer the 3 questions I posted above?

Thanks
 
I'm a little concerned about the clipping of the taillights too. This makes the image somewhat cold and unnatural to me. I've seen this more in red footage, white clipping of strongly coloured lights.
Strangely Red does display strongly coloured LED light quite well, even when it clips, like the orange LEDS on the back of the bus, image 14 on http://www.claudiomiranda.com/redvsf23/index.html

This red-turning-into-white happens as soon as the red channel clips.
It seems something weird happens in the sensor or the debayering that makes green and blue channel clip along with the red, making it white.

Strangely the blue channel of the F23 is very black in the taillights, making it a very warm orange colour. This must be due to some DSP I guess, but I like the result of it.

Is this an issue that has been looked into or maybe will be in the future?
Maybe something that can be done in Redcine while debayering?
 
I'm just hoping that at some point we start talking about the cameras that are in peoples hands. If we are constantly comparing the real cameras to the theoretical future cameras than we can theoretically delude ourselves eternally. At some point we have to start learning the boundaries of what we have, that's what I'm hungry for, that's why I'm here.

IBloom
 
Finner said:
A new test would be great to see Jim.

Any chance you could answer the 3 questions I posted above?

Thanks
Click to expand...

As soon as you change your avatar...

Jim
 
I've noticed the red channel clip 'early' in lots and lots of shots. I think it just goes with the territory of having a green biased sensor. Often the data is still 'discoverable' in the red channel but you have to throw the white balance completely out of wack to find it.

If you have to boost the red channel by a stop to balance the light loss at the filter you're losing a stop of latitude.

(The biggest reason why I want a REDStream color mode)
 
Status
Not open for further replies.
Back
Top