Overview The Ternary serializer uses base-3 (ternary) encoding instead of binary, representing each DMX channel as 6 ternary digits. Each digit can have one of three values (0, 1, or 2), encoded as different intensity levels.
Class: Ternary Namespace: GlobalImplements: IDMXSerializerConfiguration Constants Size in pixels of each ternary digit block (4×4 pixels)
Number of digit blocks per column (8 × 6 = 48 blocks)
How It Works Ternary Representation Each byte value (0-255) is converted to 6 ternary digits:
// Example: Value 170 in different bases Decimal : 170 Binary : 10101010 ( 8 digits , base - 2 ) Ternary : 20022 ( 6 digits , base - 3 , padded ) // Ternary breakdown for 170: 170 ÷ 3 = 56 remainder 2 → digit [ 5 ] = 2 56 ÷ 3 = 18 remainder 2 → digit [ 4 ] = 2 18 ÷ 3 = 6 remainder 0 → digit [ 3 ] = 0 6 ÷ 3 = 2 remainder 0 → digit [ 2 ] = 0 2 ÷ 3 = 0 remainder 2 → digit [ 1 ] = 2 0 ÷ 3 = 0 remainder 0 → digit [ 0 ] = 0 Result : [ 0 , 2 , 0 , 0 , 2 , 2 ]
Intensity Mapping Each ternary digit is encoded as a grayscale intensity:
Ternary Value Intensity RGB Value 0 0% (0, 0, 0) 1 50% (127, 127, 127) 2 100% (255, 255, 255)
Capacity Base-3 can represent larger values with fewer digits:
Binary : 2^8 = 256 values with 8 bitsTernary : 3^6 = 729 values with 6 digits (only 0-255 used)
This means 25% fewer blocks per channel compared to binary encoding.
Methods SerializeChannel Encodes a DMX channel value as 6 ternary digit blocks.
public void SerializeChannel ( ref Color32 [] pixels , byte channelValue , int channel , int textureWidth , int textureHeight )
The pixel array to write the encoded digits to
The DMX channel value (0-255) to encode
Width of the output texture
Height of the output texture
Process:
Converts the channel value to 6 ternary digits
For each digit:
Maps ternary value (0, 1, 2) to intensity (0%, 50%, 100%)
Calculates position based on (channel * 6) + digitIndex
Creates a 4×4 grayscale block
Writes the block to the pixel array
Skips blocks that exceed texture bounds
convertToTernary Converts a byte value to a 6-digit ternary representation.
public byte [] convertToTernary ( byte N )
The decimal value to convert (0-255)
Returns: byte[] - Array of 6 bytes, each containing 0, 1, or 2Algorithm: byte [] bytes = new byte [ 6 ]; int index = 5 ; // Start from least significant digit for ( int i = 0 ; i < 6 ; i ++ ) { bytes [ index -- ] = ( byte )( N % 3 ); // Get remainder N /= 3 ; // Divide by 3 } return bytes ; // [most significant ... least significant]
DeserializeChannel Decodes a DMX channel value from ternary digit blocks.
public void DeserializeChannel ( Texture2D tex , ref byte channelValue , int channel , int textureWidth , int textureHeight )
Not Properly Implemented : The deserialization currently uses binary decoding logic and does not correctly convert ternary digits back to byte values. This is a known limitation.
Usage Example // Create ternary serializer var ternary = new Ternary (); // Initialize pixel array int textureWidth = 256 ; int textureHeight = 256 ; Color32 [] pixels = new Color32 [ textureWidth * textureHeight ]; // Test conversion byte testValue = 170 ; byte [] ternaryDigits = ternary . convertToTernary ( testValue ); Debug . Log ( $"Ternary representation of { testValue } : [ { string . Join ( ", " , ternaryDigits )} ]" ); // Output: "Ternary representation of 170: [0, 2, 0, 0, 2, 2]" // Encode channel ternary . SerializeChannel ( ref pixels , 170 , 0 , textureWidth , textureHeight ); // Visual result for channel 0, value 170: // Block 0: Black (ternary 0) // Block 1: White (ternary 2) // Block 2: Black (ternary 0) // Block 3: Black (ternary 0) // Block 4: White (ternary 2) // Block 5: White (ternary 2)
Visual Pattern Ternary encoding produces distinct three-level patterns:
Value: 0 (ternary: [0,0,0,0,0,0]) ████████ All black Value: 255 (ternary: [0,2,2,1,0,0]) ████▓▓▓░ Mixed pattern ^ ^ ^ | | └── 0 = black | └──── 1 = gray └────── 2 = white Value: 121 (ternary: [0,1,1,1,1,1]) ██▓▓▓▓▓▓ Mostly gray
Channel Capacity Since each channel uses 6 blocks instead of 8:
int maxBlocksPerColumn = blocksPerCol ; // 48 int maxChannelsPerColumn = maxBlocksPerColumn / 6 ; // 8 channels int columns = textureWidth / blockSize ; int totalChannels = maxChannelsPerColumn * columns ; // Example: 256×256 texture // Columns: 256 / 4 = 64 // Channels: 8 × 64 = 512 channels
Capacity by Resolution Resolution Channels vs Binary 128×128 128 +23% 256×256 512 +23% 512×512 2,048 +23% 1024×1024 8,192 +23%
Advantages
Higher Density 25% more channels than binary (6 vs 8 blocks per channel)
Three Levels Intermediate gray level provides smoother gradients
Visual Distinction Three-level patterns are visually distinct from binary
Efficient Encoding 3^6 = 729 possible values in 6 digits
Limitations Incomplete Implementation : Deserialization does not properly convert ternary back to binary. This serializer is currently encode-only .
No Working Decoder : Cannot reliably receive ternary-encoded dataCompression Sensitivity : Middle gray level (127) may be distorted by video compressionThreshold Complexity : Requires two thresholds (0.33 and 0.66) instead of oneExperimental : Not widely tested or used in production
Decoding Considerations To properly implement ternary decoding:
public byte DecodeTernary ( byte [] ternaryDigits ) { // Convert ternary digits back to decimal int result = 0 ; int multiplier = 1 ; for ( int i = 5 ; i >= 0 ; i -- ) // Start from least significant { result += ternaryDigits [ i ] * multiplier ; multiplier *= 3 ; } return ( byte ) result ; } // Read ternary digit from pixel intensity public byte ReadTernaryDigit ( Color color ) { float intensity = color . r ; // Assuming grayscale if ( intensity < 0.33f ) return 0 ; if ( intensity < 0.67f ) return 1 ; return 2 ; }
Best Practices 1. Video Quality Ternary encoding is more sensitive to compression than binary:
Required Settings: - Bitrate: >= 30 Mbps (higher than binary) - Compression: Near-lossless (CRF 12 or lower) - Color Space: RGB - Chroma Subsampling: 4:4:4
2. Testing Test the full range of values:
// Test critical values byte [] testValues = { 0 , 1 , 2 , 85 , 127 , 170 , 254 , 255 }; foreach ( byte value in testValues ) { byte [] ternary = ternary . convertToTernary ( value ); Debug . Log ( $" { value } → [ { string . Join ( ", " , ternary )} ]" ); }
3. Visualization Create debug visualization to verify encoding:
public void VisualizeChannel ( byte channelValue ) { byte [] digits = convertToTernary ( channelValue ); string visual = "" ; foreach ( byte digit in digits ) { visual += digit switch { 0 => "█" , // Black 1 => "▓" , // Gray 2 => "░" , // White _ => "?" }; } Debug . Log ( $"Channel { channelValue } : { visual } " ); }
Mathematical Properties Ternary Number System Positional values (base-3): [d5] [d4] [d3] [d2] [d1] [d0] 243 81 27 9 3 1 Example: 170 in ternary = 0×243 + 2×81 + 0×27 + 0×9 + 2×3 + 2×1 = 0 + 162 + 0 + 0 + 6 + 2 = 170 ✓
Value Ranges Digits Min Max Range 1 0 2 3 2 0 8 9 3 0 26 27 4 0 80 81 5 0 242 243 6 0 728 729
Since DMX needs 0-255 (256 values), 6 ternary digits provide more than enough range.
Future Improvements This serializer is experimental. Contributions to implement proper deserialization are welcome!
Potential enhancements:
Complete the DeserializeChannel implementation
Add two-threshold decoding (0.33 and 0.67)
Implement error correction for the middle gray level
Add compression resistance testing
Compare performance vs binary in real-world conditions
When to Use Ternary encoding is best suited for:
Research and experimentation with alternative encoding schemesHigher channel density requirements where decoding is not neededVisualization purposes where three levels provide better visual feedbackOne-way transmission scenarios (encode-only)
Do not use Ternary serializer for production applications requiring bidirectional communication until deserialization is properly implemented.