What is bit depth?
The “bit depth” in a video is “the amount of data allocated per pixel”.
It has a different meaning than “bit depth” in audio.
“Bit depth” in audio…
→ The amount of data allocated per sample
“Bit depth” on video …
→ The amount of data allocated per pixel
Reference – meaning and relationship of sample rate, bit depth, and bit rate
further,
bit depth
color depth
pixel depth
They all have the same meaning.
For example, if the “bit depth” is “24 bits”, it means that 24 bits are assigned to a pixel.
Also, “bit depth” is expressed in the unit of bpp (bits per pixel).
Example: 24 bits per pixel = 24 bpp
File size calculation
30fps image
for example,
Bit depth: 24bpp
Resolution: 1920×1080
for videos
1920 pixels ✕ 1080 pixels ✕ 24 bits = 49766400 bits = 6220800 bytes ≒ 6.2 MB
You can see that it will be a whopping 6.2MB per frame.
(For a 1 minute video at 30 fps, 6.2 MB/frame ✕ 30 frames/s ✕ 60 s ≒ 11 GB)
By the way, this is the so-called “uncompressed video (RGB24)”.
Differences due to color space.
450px-YUV_UV_plane.svg
Quote: YUV – Wikipedia
Even if it says the same “24bpp”, the mapping method differs depending on the color space.
for example,
For “24bpp” in RGB:
→ Assign 8 bits to each of R, G, B
For YUV (YCbCr) of “24bpp”:
→ Assign 8 bits to Y, U, V respectively
It is so.
In the case of YUV, the number of bits allocated depends on the “sample ratio”.
The correct answer is to express YUV as YCbCr, but for some reason it’s commonly called YUV in the PC world, so I’ll call it YUV.
In the case of YUV, it becomes “effective bit”
In the case of YUV, the “sample ratio” differs depending on the format, so the method for taking the brightness and color difference will be different.
for example,
With “YUV444”, 8 bits each are taken from Y, U and V.
With “YUV420”, Y takes 8 bits each, but U and V share 8 bits.
It will be like this.
So in the case of “YUV420”, you can see that the actual number of bits times one is “12 bits”.
The “effective number of bits per pixel” at this time is called the “effective bit”. (“12bit” for “YUV420”)
From this, there is a double difference in file size between “YUV444” and “YUV420”. . (It’s actually compressed with a codec, so it doesn’t make much of a difference)
