Apple Final Cut Pro 5 - Manual (Page 1337)

550 Part III Output
About Video Bit Depth
Each pixel uses a certain number of bits to store the intensity of a color channel. For
example, a typical RGB media file may use 8 bits for each of the red, green, and blue
channels. Using 8 bits allows for 256 intensity values per color channel.
The more bits used per color sample, the more subtle variations of color are possible.
For example, if you use 10 bits per color sample instead of 8, there are 1024 possible
intensity values per color channel instead of 256.
When using the Y´C
B
C
R
color space, Final Cut Pro supports either 8- or 10-bit video
media files for capture and output, as well as for rendering video filters and transitions.
To determine which color space and bit depth your video interface supports, see the
documentation that came with the interface. For more information on which filters
support 10-bit resolution, see “
Video Filters Available in Final Cut Pro ” on page 203. For
more information on which transitions support 10-bit resolution, see Volume II,
Chapter 22, “Refining Transitions Using the Transition Editor.”
Most DV video formats use 8-bit color samples, while some professional video formats
such as D-5 use 10 bits per color. Many third-party video interfaces and codecs support
10-bit video capture, processing, and output, which improves color quality significantly.
This is especially true for subtle color gradations.
When rendering effects such as transitions or filters, 8-bit precision is faster, but yields
lower quality when using clips captured at 10-bit resolution. Rendering effects at full
10-bit precision results in higher quality video, but slower rendering times and slightly
larger render files.
32-Bit Floating Point Processing
Most effects and compositing are the result of adding, subtracting, multiplying, or
dividing pixel intensity values. Final Cut Pro supports high-resolution video processing
of Y´C
B
C
R
sequences by performing calculations in 32-bit floating point number space.
Compared to 8- and 10-bit integer calculations, 32-bit floating point numbers have an
extremely high level of precision, which helps to avoid rounding errors that can
accumulate as you add more layers to a composite, or add multiple filters to a clip.
In most cases, you should choose to render your sequence using 32-bit floating point
space (high-precision YUV) for final rendering before output or export.