TGA

Longstanding, common industry format for uncompressed images.  Generally encoded at 8-bit per channel RGB. (24-bit/pixel).

VARIANTS:

[1920 * 1080] px * 30 fps * [8+8+8] b * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 178 MB/s

32-bit
(Adds alpha transparency channel support at 33% additional data.)

[1920 * 1080] px * 30 fps * [8+8+8+8] b * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 238 MB/s

RLE
(Applies lossless run-length encoding for variable but often significant file size savings at the expense of significant and variable realtime playback overhead to "unzip" frames on the fly.  Compatible with 24-bit or 32-bit TGAs.)

Variable

16-bit
(Lower quality at 5-bit per color and optional 1-bit alpha, at a 33% overall data reduction)

[1920 * 1080] px * 30 fps * [5+5+5+1] b * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 119 MB/s

7th 4:4:4

Uncompressed RGB/4:4:4 image data format, most commonly at 8-bit per channel, equivalent in quality and data rate to TGA 24-bit

VARIANTS:

[1920 * 1080] px * 30 fps * [8+8+8] b * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 178 MB/s

10-bit, 12-bit
(more channel value resolution at the expense of equivalent added data rate)

[1920 * 1080] px * 30 fps * [10+10+10] b * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 223 MB/s

(10-bit example.  For 12-bit just replace 10+10+10 by 12 support.+12+12)

4:4:4+A
(adds alpha transparency channel support at 33% additional data.  )

[1920 * 1080] px * 30 fps * [8+8+8+8] b * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 238 MB/s

RGB-Z
(lossless compression, comparable in relative filesize and performance tradeoffs to TGA RLE)

Variable

7th 4:2:2

33% data rate reduction in exchange for 4:2:2 chroma sub-sampling (RGB values are converted to YCbCr representation, full resolution luma/intensity is maintained but essentially chroma/hue component is shared between 2px pairs.  4:2:2 sub-sampling is generally considered the least intrusive form of lossy data compression, because the human eye is generally agreed to be more sensitive to differences in image intensity than hue.)

VARIANTS:

[1920 * 1080] px * 30 fps * [8*4+8*2+8*2] b / 4 px * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 119 MB/s

10-bit

[1920 * 1080] px * 30 fps * [10*4+10*2+10*2] b / 4 px * [1 B / 8 b]* [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 149 MB/s

4:2:2-Z

Variable

12/10-bit 4:2:2+A
 
(Y12CbCrA10.  A uniquely optimized format combining 12-bit Luma [Y], 10-bit 2px shared chroma values [CbCr], plus a 10-bit alpha channel.  Effective total 32-bit/pixel. Supported as of Delta 2.8.11+)

[1920 * 1080] px * 30 fps * [12*4+10*2+10*2+10*4] b / 4 px * [1 B / 8 b]* [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 238 MB/s

7th 4:2:0

Similar to 7th 4:2:2 (8-bit), but a 50% reduction in file size in exchange for deeper chroma value sharing.  Essentially the same chroma value is shared between 4px instead of just 2px.

VARIANTS:

[1920 * 1080] px * 30 fps * [8*4+8*2+8*0] b / 4 px * [1 B / 8 b] * [1 KB / 1024 B] * [1 MB / 1024 KB] ~ 89 MB/s

4:2:0-Z

Variable

7th YCoCg DXT5

A 3:1 flat, lossy data reduction as made popular by the HapQ codec. 8-bit per component.

Exact calculation beyond scope of this guide, but effectively 33% of TGA / 7th 4:4:4 8-bit)

7th NotchLC

A lossy compression algorithm achieving data rate between 5:1 and 8:1 with mild perceived quality loss. Supported as of Delta 2.8+

Note that while NotchLC is currently produced by 3rd party tools exclusively in MOV-wrapped format, this is not recommended for direct playback (as ALL WRAPPED CODECS are not recommended).  There is a tool ‘NotchLCTo7th.exe’, included as of Delta 2.8.7+ which will unwrap NotchLC MOVs, decompress any LZ4 compression, and repackage as a .7th image sequence.  No further image data is changed, so this process maintains original image quality, and is relatively fast in processing versus a full re-render.

Audio track playback and extraction in NotchLC MOVs was not supported until 2.8.12+ by Delta nor NotchLCTo7th.exe.

➢See: MC256 Managing Media for Delta, NotchLC to NotchLC7th utility.

Variable

7th 6to1

LEGACY.  6:1 compression, lower hardware requirements

Exact calculation beyond scope of this guide, but effectively 17% of TGA / 7th 4:4:4 8-bit)

DPX

Uncompressed, most commonly encoded as 10-bit RGB.

Other supported variants:

•8-bit RGB

•8-bit or 10-bit RGB+A

•10-bit YUV/4:4:4

•12-bit 4:4:4

•Little- and Big-endian bit orders

•Filled and packed types

•V2.0HDR is supported from Delta 2.7 build 7.

Exact calculations beyond scope of this guide, but loosely closest in data rate to the 7th 4:4:4 of same bit depth.

Raw YUV 422

Raw, single 4:2:2 frames in 2VUY [UYVY] 8-bit format.

CIN

Uncompressed, 8/10-bit

A7S

8-bit or 10-bit, RGB or 4:2:2 encoding, with optional lossless compression.

SGI

Uncompressed 8-bit