AVS Forum banner

1 - 9 of 9 Posts

·
Registered
Joined
·
4,002 Posts
Discussion Starter · #1 ·
Apologies if this isn't the right forum. Warner is doing a 6k mastering of A Star is Born. Sound and Vision magazine debates whether 6k is necessary or 4k is just enough. One thing they do touch in the article is whether or not they should focus on increasing the color bit depth than the resolution when mastering. Of course, the Blu-Ray that results will only be 8-bit and 1080p, but Warner wants to do this (instead of a 2k or 4k) for "future proofing".

http://www.soundandvisionmag.com/fea...-about-6k.html


Here's the IMDB for the film.

http://www.imdb.com/title/tt0047522/


Personally, I'm in the give me more category as typically the greater the master, the greater the Blu-Ray (or HD VMD?/DVD), high def broadcast, D-Theater, etc. transfer tends to be.
 

·
Registered
Joined
·
3,314 Posts
I actually agree with Lowry. The whole reason for scanning 35mm at 6k is primarily because the newer scanners have 6k CCDs. This is mainly because of progress in imaging technology ( you can use a 6k sensor for the price of a 2k one from a few years back) CE manufacturers like to quote ever increasing bitdepth in their DACs for the same reason ( even though it doesn't actually make a difference beyond 10bit).


6k is also handy because it means you can buy one scanner that will handle every available film format.


As per usual the article balks at discussing bitdepth and colorspace.film is usually scanned to 10bit log dpx or cineon formats. Some people scan 12bit log some people scan 16bit linear and some even scan 16bit log. This is not the same as "RGB HD" ( whatever they mean by that).


Even 10bit video is not the same colorspace as a 10bit log film scan. For a start you need something like 18bits to linearly encode the same information in a 10bit log film scan without perceptable banding.


Secondly one is a full negative density encoded range the other is merely gamma corrected video.


1080p video is not "2k". 2k refers to a 2k film scan , if you went around refering to 1080p as 2k in the film industry people would think you were daft.


A 2k fullap film frame is 2048x1556 4k is 4096x3112 ( someone in that article can't do basic maths). Its normal these days to scan the full 35mm frame for every format : previously you used to get the soundtrack area cropped for academy but its been a while since I saw anyone bother to do this.


You do get smaller scan sizes for 3 perf (2048x1152) , I've never actually seen a 2 perf scan .
 

·
Registered
Joined
·
3,118 Posts
It seems Lowry and the article agree with James Cameron:

According to Lowry... we should clamor for higher dynamic range. “Instead of 10 bit, why not 16 bit?," he asks. "Wouldn’t it be nice to have high dynamic range?” This will reveal more details in the shadows and the blacks, with higher frame rates for film


The article says:
Quote:
How about shooting at 48 fps to double the capture rate? VISION3 film stock — with very low grain and stunning sharpness — would blow viewers away.

- the same frame rate James Cameron wanted.


Also, if there really was 4K of resolution in a film, is it correct that you would really need to scan at at least 8K to get that 4K resolution worth (due to Nyquist)? You could still store the data at 4k if that's all the resolution there was in the film, but the scanner would scan at 8K and downsample to 4K?
 

·
AVS Forum Special Member
Joined
·
11,139 Posts

Quote:
Originally Posted by Joe Bloggs /forum/post/14323390


It seems Lowry and the article agree with Also, if there really was 4K of resolution in a film, is it correct that you would really need to scan at at least 8K to get that 4K resolution worth (due to Nyquist)? You could still store the data at 4k if that's all the resolution there was in the film, but the scanner would scan at 8K and downsample to 4K?

The Arri draft article , which appeared edited in a recent SMPTE journal issue, shows 6k downconverted to 4k (adequate to avoid aliasing) in a table on page 15. Suspect you'd get varying opinions on what's adequate/necessary. P2Fig.12 shows 2k, 4k, and 10k scans of a negative test pattern detail.-- John
 

·
Registered
Joined
·
1,325 Posts
I'm at a loss with the math in this article:


page 27, bottom
Quote:
On the object side this would correspond to 0.2 mm in a distance of 1 m (or 1 minute of arc).

page 28, top
Quote:
Allowing for some amount of tolerance, this would be around 0.3 mm at 1 m distance(= 1.03 minutes of arc).

Common sense would tell me that going from 0.2mm to 0.3mm (50% increase) would increase the arc by the same amount, i.e. 1.5 minutes of arc (instead of 1.03).


Am I missing something?


Diogen.
 

·
AVS Forum Special Member
Joined
·
11,139 Posts

·
Registered
Joined
·
1,325 Posts
The final version has exactly the same text (page 20, right column, from top)
Quote:
Resolution of the human eye


The fovea of the human eye (the part of the retina that is

responsible for sharp central vision) includes about

140 000 sensor-cells per square millimeter. This means

that if two objects are projected with a separation

distance of more than 4 m on the fovea, a human with

a normal visual acuity (20/20) can resolve them. On the

object side, this corresponds to 0.2 mm in a distance of

1 m (or 1 minute of arc).



In practice of course, this depends on whether the viewer

is concentrating only on the center of the viewing fi eld,

whether the object is moving very slowly or not at all, and

whether the object has good contrast to the background.


As discussed previously, the actual resolution limit

will not be used for further discussion, but rather the

detail size that can be clearly seen. Allowing for some

amount of tolerance, this would be around 0.3 mm at 1

m distance (= 1.03 minutes of arc ).
In a certain range,

one can assume a linear relation between distance and

the detail size.

Diogen.
 
1 - 9 of 9 Posts
Top