Jump to content


Photo

Scanning: 8, 10 or 16 bits?


  • Please log in to reply
7 replies to this topic

#1 Mr. Macgregor

Mr. Macgregor
  • Basic Members
  • PipPipPip
  • 108 posts
  • Cinematographer
  • Los Angeles

Posted 09 August 2006 - 08:49 AM

Hello.

I shot a shortfilm in s35mm 3 perf. Only one roll. Nothing more. So aproximately 90% of the footage will be in the final cut.

Now i have to scan the negative in order to edit and color correct the film.

I asked the lab for prices and options.

And here comes the problem. This is Spain, not the US. They have the tools, but this kind of work is not what they usually do. So i must give them all the details.

They usually give the frames in TGA. I wonder if TIFF files would give any advantage here. Will both file systems offer more than 8bits? Will the scanner be able to deliver 10bit? Even 16bits per channel? Will this make sense?

Tiff files can handle 16 bits per channel. However i am not sure if there are displays capable of showing it.

Will actually FCP or After Effects be able to work with such color resolution?

Also, since the final negative must be projected in scope 2.35, what do you think, should i render the secuence in the end in anamorphic myself, or just deliver to the lab the 2k frame sequence and let them do the job?
  • 0

#2 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 09 August 2006 - 10:03 AM

Hello.

I shot a shortfilm in s35mm 3 perf. Only one roll. Nothing more. So aproximately 90% of the footage will be in the final cut.

Now i have to scan the negative in order to edit and color correct the film.

I asked the lab for prices and options.

And here comes the problem. This is Spain, not the US. They have the tools, but this kind of work is not what they usually do. So i must give them all the details.

They usually give the frames in TGA. I wonder if TIFF files would give any advantage here. Will both file systems offer more than 8bits? Will the scanner be able to deliver 10bit? Even 16bits per channel? Will this make sense?

Tiff files can handle 16 bits per channel. However i am not sure if there are displays capable of showing it.

Will actually FCP or After Effects be able to work with such color resolution?

Also, since the final negative must be projected in scope 2.35, what do you think, should i render the secuence in the end in anamorphic myself, or just deliver to the lab the 2k frame sequence and let them do the job?


Hi,

I have scanned on a Spirit to 10 bit DPX, the facility then converted to both 8 & 16 bit TIFF in their Flame and both were printed back to film. The 16 bit was clearly better. You can't see a difference on the Flame Monitor!

After Effects can work at upto 32 bit floating point, not sure about FCP.

I would deliver the files in their native resoloution.

Stephen
  • 0

#3 Mr. Macgregor

Mr. Macgregor
  • Basic Members
  • PipPipPip
  • 108 posts
  • Cinematographer
  • Los Angeles

Posted 09 August 2006 - 10:17 AM

the facility then converted to both 8 & 16 bit TIFF in their Flame and both were printed back to film. The 16 bit was clearly better. You can't see a difference on the Flame Monitor!



Why did they converted to 8 bits then?
You say that 16 is clearly better but you could not see it on the flame monitor. Then i guess you could only tell the difference in the theater. Is this correct?

Thanks.
  • 0

#4 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 09 August 2006 - 12:43 PM

Why did they converted to 8 bits then?
You say that 16 is clearly better but you could not see it on the flame monitor. Then i guess you could only tell the difference in the theater. Is this correct?

Thanks.


Hi,

The facility believed their own BS saying there would be no difference! I asked them to prove it.

You are correct, only the 35MM print looked better in the theater.

Stephen
  • 0

#5 Gavin Greenwalt

Gavin Greenwalt
  • Basic Members
  • PipPipPip
  • 225 posts
  • Other

Posted 07 October 2006 - 04:37 PM

There is only one prototype monitor on earth that I know of that can display 16 bit (half float) images in its native format. You can think of an HDR (16 bit) image like a negative, it has far more information than you'll see on the final projected print and you have to define which part of the data you want displayed.
  • 0

#6 Mark Allen

Mark Allen
  • Basic Members
  • PipPipPipPip
  • 591 posts
  • Director
  • Los Angeles

Posted 07 October 2006 - 04:48 PM

In the US if you were doing this like an effects situation, you would scan the film to cineon files which After Effects handles fine. For editorial, you'd then make movies of those files in a corrected colorspace and of an editable size. Then conform them back in AE based on your cut.

In theory you could use a program like automatic duck to do the program to program translation and then swap out your footage (treat the editorial versions as proxies). But with a project as short as yours and eye conform might make sense.

Then, you work on the project in AE with an adjustment layer converting the cineon to rgb for viewing while you work (or your own adjustment thereof). Then, when you're happy, check your cineon files to see if they also look good, render out and deliver to the lab cineons.

The theory of using cineons vs. tga 8 bit is that you simply have more range to work with and less data loss in transfers.
  • 0

#7 Dirk DeJonghe

Dirk DeJonghe
  • Basic Members
  • PipPipPipPip
  • 605 posts
  • Industry Rep
  • Kortrijk,Belgium

Posted 08 October 2006 - 04:45 AM

If you work with 8bit linear images, you will very quickly end up with banding because there are too few steps in the darker parts of the picture. 12 bit is really the minimum for linear images if you want to do any color correction or compositing at all.

Log 10 bit images such as Cineon or DPX solve this problem because they have finer steps in the darker parts and coarser steps in the lighter parts of the picture (unscientific explanation). The human eye is most sensitive in the darker parts and less in brighter parts.

I did a digital grading on HDCAM material shot 8bit linear. Banding was unavoidable in night shots.

Film is logarithmic, the human eye is too.
  • 0

#8 Mr. Macgregor

Mr. Macgregor
  • Basic Members
  • PipPipPip
  • 108 posts
  • Cinematographer
  • Los Angeles

Posted 08 October 2006 - 12:29 PM

Wow, Old topic i see. The truth is that i am actually working with cineon files in AE. So problem solved.
  • 0


Opal

CineLab

CineTape

Media Blackout - Custom Cables and AKS

Aerial Filmworks

Broadcast Solutions Inc

Paralinx LLC

The Slider

Rig Wheels Passport

Gamma Ray Digital Inc

Metropolis Post

FJS International, LLC

Tai Audio

Technodolly

Visual Products

Glidecam

Abel Cine

Ritter Battery

rebotnix Technologies

Willys Widgets

Wooden Camera

Metropolis Post

CineTape

Glidecam

Broadcast Solutions Inc

Gamma Ray Digital Inc

Willys Widgets

rebotnix Technologies

Wooden Camera

Paralinx LLC

Visual Products

Technodolly

Media Blackout - Custom Cables and AKS

The Slider

Opal

FJS International, LLC

Tai Audio

Abel Cine

Aerial Filmworks

CineLab

Ritter Battery

Rig Wheels Passport