Let's talk sensor...

[David Mullen ASC]

I still don't know what you are referring to in regards to this 1.5 crop factor. When does this 66% resizing take place?

Obviously oversampling is better. If you need to end up with a 4K file, starting out with a 6K sensor is going to look better than if you had a 4K sensor, but I don't get all this talk about a 1.5 crop factor and 66% resizing. Are you talking about a 4K finish? A 2K finish? An HD finish?

And why is 4x3 "full frame"? There really isn't such a term in cine technology. There is "Full Aperture" but that can be any aspect ratio. 4x3 is traditional but it is not particularly "fuller" -- for example, in 16mm, the 4x3 standard 16mm gate is physically smaller than the 1.68 Super-16 gate.

Again, it comes down to what purpose does a 4x3 sensor serve you if your delivery formats are widescreen in some form or another? You can't take advantage of more pixels vertically if you just end up cropping them to achieve a widescreen shape, so there is no oversampling advantage - unless you use anamorphic lenses. Otherwise, the only advantages are: ability to reposition the image vertically, or deliver a 4x3 commercial (but then, most 4x3 is for standard def video, so cropping a 4K image to achieve 4x3 720 x 480 is no big deal, resolution-wise. You are already way oversampled.)

"Full frame" is only a term in DSLR still photography to differentiate sensors that match the size of the 8-perf 35mm still camera frame from those that are half size, i.e. match the size of the 35mm cine formats. Those are the only people who factor in conversions for FOV from what they are used to in 35mm still photography. It's meaningless for cine shooters where the RED sensor already matches the size of 35mm cine formats, so there is no conversion factor. And most cine formats use some form of cropping to achieve aspect ratios like 1.85 or 2.40.

So please explain where this 1.5 crop factor comes from and when this 66% resizing happens.

 

[amrrahmy]

the 4.5K would be 2983 give or take few pixels.(and it's still depends on the recorded material, maybe a bit less in some cases).

 

[amrrahmy]

there is no 4K finish, it's not a 4K file, that's the whole thing it's a 3K file to start with, not a 4k file(it's a pretend 4K file).

 

[amrrahmy]

i'm not usually very good at explaining stuff, and english is my fourth language.

if u google a bit u'll find all the info u need to know about diff kinds of sensors, and how they work and store data, u can take dslr cameras for example, look at a footage from a d2x or any other camera with crop factor, and look at footage from a d3 with and without down scaling. that is pretty easy.
i dont have links, but it's not hard to find, it wont take u 10-15mins to find all necessary info.

 

[David Mullen ASC]

there is no 4K finish, it's not a 4K file, that's the whole thing it's a 3K file to start with, not a 4k file(it's a pretend 4K file).

Where did you get this idea???

A 4K Bayer RAW image that is debayered to 4K RGB may only have an effective resolution of 3K, but that doesn't make it a 3K file size. The size of the files is the size of the files regardless of line chart resolution measured. I could make a 4K file out of a DV image.

It records 4K RAW Bayer files that are converted to whatever you want, usually 4K RGB files. The only time it is a 3K file is when you shoot in 3K mode or create 3K files yourself.

And this still doesn't explain where the 1.5 crop takes place.

Again, you are mixing up DSLR terms with cine technology terms.

Yes, obviously you will get a better 4K RGB finish if you started out with a RAW file with higher pixel resolution. Everyone knows that.

 

[amrrahmy]

full frame means 4:3 1:1 pixel recording, that's universal and used for cine, photo, animation, math, chemistry, physics among other stuff.

 

[David Mullen ASC]

full frame means 4:3 1:1 pixel recording, that's universal and used for cine, photo, animation, math, chemistry, physics among other stuff.

Just provide an outside link or reference that says that is true in cinema technology because I've never run across one. And I read a lot of cinema technology books and magazines...

 

[amrrahmy]

may only have an effective resolution of 3K

there u go.

I could make a 4K file out of a DV image.

it would still be considered a dv image and not a 4k image.
if a dv image would save the image as a 4k image, it doesn't mean it's a 4K image.
if i put a merc badge on my mitsubishi, it would still be a mitsubishi.

 

[amrrahmy]

also the 4K file out of the dv image would loose some info(quality) in the process.

 

[David Mullen ASC]

You're missing the point -- it is still a 4K file size. Effective resolution is a different issue than file size. The RED shoots 4K RAW files and it generates 4K RGB files. If you simply look at the file sizes, you can see that. It doesn't create 3K RGB files and then uprez them to 4K RGB files. That's not how a debayering algorithm works -- it doesn't just take the green pixels and uprez them from 2K to 4K, and the red and blue pixels and uprez them from 1K to 4K. It doesn't take a 4K RAW recording and create a 3K RGB file and then uprez that to a 4K RGB file. It creates 4K RGB files from 4K RAW files that have an effective resolution of 3K -- but they are still 4K file sizes.

 

[amrrahmy]

i think u can look in the arri d21 site.
and u can look for manuals and tutorials of full frame sensors.

for full frame to be 4:3 u can find that in books easily.
for a cinema camera to have a full frame sensor, it would have to find a book made more recently.but a full frame sensor is always 4:3 1:1 and it doesn't have a crop factor. that's why it's 4:3, and not a cheap small sensor with crop factor.

 

[amrrahmy]

it's a smaller sensor than it records(the crop factor happens at the start before the image is recorded),it records a 4K file that is 3K more or less.

 

[amrrahmy]

sorry i don't see ur posts when i'm writing.so i post later, hope that's not too confusing.

 

[Brook Willard]

Regarding crop factors and still cameras.

No part of my diagram or discussion has anything to do with still cameras and their 8-perf "full frame" size. Any 1.5x or 1.6x crop factors that people are getting caught up on are unrelated to the sensor size issues that I'm discussing. If people are trying to visualize the still vs. motion film issues, RED has my chart on the subject up on their website to help clear things up.

As David said, "full-frame" and 4:3 are two different things from two different industries of film.

 

[amrrahmy]

u cant use the recorded file as a 4K file(u can but that wont be the best quality of the image, or usable actually), u can however resize the file, do whatever u want in post and resize again to whatever u want and that would get u the best results.
and u can use a nice dv cam with a (16 or 35mm adapter) and make a usable 2K or 4K image for theater projection.(but not for bluray or internet - u will have to go with the native, u cant upsize for monitors or really expensive tv(that are actually monitors - they are also pretend tvs, but in a good way this time)).

 

[David Mullen ASC]

The RED camera does not crop the Mysterium sensor to 3K and then uprez it to 4K (unless you select 3K RAW mode but convert it to 4K RGB later.)

It records 1:1, pixel for pixel. In 4K mode, it records 4096 pixels across, the same pixels on the sensor, 4096 pixels across (it doesn't use the area outside of 4096 yet). 1:1.

But because it is a RAW Bayer image, it cannot create a 1:1 RGB image without an algorithm -- a 1:1 conversion from RAW to RGB is not possible no matter what the sensor is, not unless you drop the Bayer design. A 1:1 conversion of a 4K RAW recording would give you 2K green pixels and 1K red and blue pixels each. To truly get 4K RGB 1:1 from a sensor, you need a sensor that had 4K each of red, green, and blue photosites, not a Bayer design.

 

[amrrahmy]

the charts also are for the recorded size with the crop factor(not the usable size).not to mention that the compressed files(no matter how good the compression is - i think it's really good by the way) or not as good as uncompressed footage.
so the chart is not really that accurate if ur trying to compare quality.

 

[Peter Majtan]

EPIC is said to be 5K with the ability to derive pristine 4K+ from the Bayer data. RED One is 4K~4.5K with the ability to deliver pristine 3K or a bit over.

For EPIC to resolve pristine 4K it would have to have at least 5252 (which is slightly more then 5K - 5120). I am confident it's going to have even more - close to 6K...

I don't think RED will be producing the Mysterium X in multiple densities. So assuming the same photosite density across all Mysterium X sensors, we could assume that EPIC has a larger sensor.

Now this will not work as Scarlet (based on Mysterium-X) is 2/3" with 3K resolution. Maybe EPIC will have the same density as Scarlet, but in full S35 frame...

The NAB literature and announcements said "full frame" sensor for EPIC. It's a vague description, but I also took it then as a larger size than the RED One.

Doesn't it say "full frame S35"? (Same size as Red One...)

 

[Dustin Cross]

Is there a way to hide posts from amrrahmy? He (or she) doesn't understand anything about Red and post nothing but crap.

 

[amrrahmy]

1:1 is the crop factor, it's 1.5:1(or somewhere between 1.2-2.0 depending on the sensor), u keep confusing the #s.
that happens before the info gets recorded on the pixels.(the recorded being 1:1 after that is the doing of the codec) but that happens to the 1.5:1 size, that's way u need to resize to a usable 3K or so image.

if u just look for crop factor info and possible see a couple of videos and tutorials, u would get a far better idea.

also i'm out for today, i will take a look at the thread tomorrow.
i'll try to help out with any info needed, if i can.