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 :)

Dragon 6k not eligible for UHD "Tier 1" camera?

Regarding the Epic Dragon test I wonder how "Recording were made onto CFast card in the camera", like they say?
 
BRANDON JAMESON said:
Sony F65 is 20 megapixels.

http://www.sony.co.uk/pro/product/b...-motion-picture-camera/f65/overview/#overview
Click to expand...

On it's best day Sony lists the F65 total available pixels as 8192x2160, or..... 17.7 megapixels.

I've enjoyed working with and shooting with the F65. I have not enjoyed the rather cloudy marketing of the camera right down to not listing sensor specifications in the official manual.

It's 17.7 megapixels on a 45 degree angle. Decent for 4K, interpolated to 8K.
 
Noah Yuan-Vogel said:
So you don't think the pixel softness we see in RED's cameras is from the OLPF? It does seem odd that the EBU paper on Epic Dragon says there is still aliasing, which is surprising considering how soft the debayered images are at the pixel level, something I always assumed was due to REDs OLPFs being much more aggressive than other manufacturers'. What else would it be?
Click to expand...

I have not empirically tested all the OLPFs out there, but I wouldn't be surprised if the RED OLPFs are stronger than most. That said, the effect of any camera's OLPF is known and measurable - which means that the processing algorithm can account for that in decode - meaning that some of the de-focus effect can be "inverted".

More to the point, an R3D file is a data set used to "construct" a RGB image. How exactly you "develop" that image has a lot to do with how it looks (and measures). Do we know which scaling filter EBU testers used? Gamma curve? Etc...

Cheers - #19
 
Would be interesting to see what the measurements would have been using the RedCast module to pump out the 2160P signal - especially if that signal was dialed in to calibrated chart values.

I should probably stop posting on this topic, but it drives me nuts when testing methodology has more impact on results than camera capabilities. At least in the beginning, RED was NOT trying to make a better TV camera - they wanted to make a worthy film replacement. The goal was more film camera video tap than a hero live feed.

Once you move past measuring the "tap" and construct your RGB signal from RAW Bayer data, then processing methodology comes into play. I firmly believe that with a critically focused high resolving lens, I could develop a RGB picture of typical charts that would measure far better than whatever settings the EBU used - as could most colorists and professional DITs.

Cheers - #19
 
Blair S. Paulsen said:
Would be interesting to see what the measurements would have been using the RedCast module to pump out the 2160P signal - especially if that signal was dialed in to calibrated chart values.

I should probably stop posting on this topic, but it drives me nuts when testing methodology has more impact on results than camera capabilities. At least in the beginning, RED was NOT trying to make a better TV camera - they wanted to make a worthy film replacement. The goal was more film camera video tap than a hero live feed.

Once you move past measuring the "tap" and construct your RGB signal from RAW Bayer data, then processing methodology comes into play. I firmly believe that with a critically focused high resolving lens, I could develop a RGB picture of typical charts that would measure far better than whatever settings the EBU used - as could most colorists and professional DITs.

Cheers - #19
Click to expand...

It seems also a little rich when just a few months ago we had the Arri brigade telling everyone that their 3.2K cameras etc were fine for 4K. At the end of the day I hate to say it but it is all video and these are all digital video cameras. But they have distinct approaches and looks and one form of test or measurement seems to overlook the subjective aspects. What sucks as has been mentioned is that people read all this and perceive these things to be the impartial truth. Which sticks us as Redusers with always being the ones having to explain things - which is a crap position.
 
Industry and politics...

Anyway, I figured any of you guys who have been in the biz for a while should already know that ARRI pixels are 10X better and more detailed than the other guys' pixels. :crazy:

After all, isn't that why Sony counts every other horizontal pixel twice on the F65? Gotta compete somehow, even if it takes marketing smoke and mirrors to do it.
 
The sad reality is that someone who really knows their stuff can make any of the top cameras seem mediocre, or fantastic, depending on choices in testing methodology. The core issue is that every other top camera does far more digital signal processing before writing data to media than RED.

Cheers - #19
 
Blair S. Paulsen said:
The core issue is that every other top camera does far more digital signal processing before writing data to media than RED.
Click to expand...

Could you clarify what you mean or cite a source for this? In what sense do arriraw, Sony raw, or BM dng raw represent more processed recordings than red code raw?
 
You know what's funny.... Just this morning measuring MTF for Weapon and at 5KHD and 6KHD (never mind full width 5k and 6k) out-resolves a 4K zone plate.....

Graeme
 
Graeme Nattress said:
You know what's funny.... Just this morning measuring MTF for Weapon and at 5KHD and 6KHD (never mind full width 5k and 6k) out-resolves a 4K zone plate.....
Click to expand...

It's an interesting thing really in regards to who is doing testing and setting "standards" these days. Stateside in most camps from what I've seen most cameras that shoot 3840x2160 through 4096x2160 and above contain the minimum pixel count requirements for UHD 4K production (also DCI spec for those that use that). Then there's the process of diving into the resolution tests, color tests, on board codecs, capture/codec/post artifacts, etc....

All of these leads to unique Approved Camera Lists for UHD or DCI 4K production. Total Pixel Resolution gets you in the door to be considered, the more technical tests are the bouncer that lets you into the club or not. Specific main concerns of color reproduction, total captured resolution, total captured dynamic range, codec flexibility/artifacts, and visual artifacts tend to be the main categories I've seen cameras halted at. This more or less is what's developed in the last 3-5-ish years as 4K is taking a firmer hold.

Fast forwarding/rewinding to January of 2016 when the UHD Alliance announced their specifications for UHD Premium, we now how a clearer scope in regards to what to aim for when it comes to display tech, which effects our decisions regarding source image acquisition as they are made up of nearly everybody who's anybody.

I've been involved with more than a handful of the companies that make up the UHD Alliance and all of those companies mostly aim for cameras and content that hits a generally similar target. For the ones that make lists, the Approved Camera List at most of those entities often overlaps, which is good.
 
More control, more complexity

More control, more complexity

Noah Yuan-Vogel said:
Could you clarify what you mean or cite a source for this? In what sense do arriraw, Sony raw, or BM dng raw represent more processed recordings than red code raw?
Click to expand...

Starting with when the photons hit the sensor>create voltages, and ending either at a live spigot OR writing data to media, there are digital signal processing steps along the way. Going to a live output, the amount of processing is significantly greater and, with a Bayer CFA, includes construction of an RGB image using an algorithm to "read" the RAW data set. Even when writing "RAW" Bayer data to media, like a RedMag, there is still some upstream processing to yield a superior data set. My point is that the amount of processing done in the front end of the camera, before any data values are written, varies from camera system to camera system.

Let's use noise reduction as an example. AFAIK, all of the top CMOS Bayer CFA motion cameras do at least some NR early in the signal chain to improve signal to noise ratio. Out of all of them, RED does the least noise reduction in camera. Why? For some material, it may suit creative intent to have a noisier, "grainier", grittier texture. Moreover, less NR also means less attenuation of high frequency detail and greater fidelity to scene referred values. Good noise reduction techniques can really improve SNR, which is why most camera systems use NR more aggressively than RED. The downside is that once you toss data that the DSP decided was noise, it's gone. With an R3D file, you can apply NR with an eye to the particular material at hand, being more or less aggressive depending on the effect it has on the image in relationship to the intended look.

Hope that helps.

Cheers - #19
 
Tarquin Cardona said:
Clearly Graeme you must have done something wrong:)
Click to expand...

Again, one of the funny things with measuring resolution is that when you do something wrong, you normally get a lower figure than that which the system is capable of, not a higher one.... :-)

Graeme
 
Graeme Nattress said:
You know what's funny.... Just this morning measuring MTF for Weapon and at 5KHD and 6KHD (never mind full width 5k and 6k) out-resolves a 4K zone plate.....

Graeme
Click to expand...

I'd love to see those shots, any special sauce or processing involved in preparing your zone plate shots or just straight out of RC-X?
 
Nothing too special, other than I wrote some rather clever analysis software to turn the zone plate image into an MTF curve and give resolution read outs.

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