Byte and Bit Operations
caution
This page describes lower level concepts of working with binary data and is intended for advanced creators.
You can use the Bitcast
method on DataTokens to do value-preserving type changes ("unsafe casts") on primitive data types in Udon.
Some standard C# classes for operating on raw binary data are also available, including BitConverter
and Buffer
.
Example Code
Byte and Bit Operations Example, Basic Serializer
using System;
using System.Text;
using UdonSharp;
using UnityEngine;
using VRC.SDK3.Data;
public class BitConverterExample : UdonSharpBehaviour
{
void Start()
{
//Create test data
int originalInt = 63;
double originalDouble = 734531.433d;
string originalString = "Test string";
float[] originalFloatArray = new float[] { 543, 12.6f, 63.1231f };
//Serialize and then deserialize
byte[] serialized = Serialize(originalInt, originalDouble, originalString, originalFloatArray);
Deserialize(serialized, out int newInt, out double newDouble, out string newString, out float[] newFloatArray);
//Print the results to see if everything matches
Debug.Log($"{originalInt} - {newInt}");
Debug.Log($"{originalDouble} - {newDouble}");
Debug.Log($"{originalString} - {newString}");
Debug.Log($"{originalFloatArray.Length} - {newFloatArray.Length}");
for (int i = 0; i < originalFloatArray.Length && i < newFloatArray.Length; i++)
{
Debug.Log($"{originalFloatArray[i]} - {newFloatArray[i]}");
}
//For individual values we can also use DataToken Bitcasting to get bit access
double doubleValue = 123.456d;
DataToken doubleToken = new DataToken(doubleValue);
//We used ulong because it has the same size as a double (8 bytes)
DataToken ulongToken = doubleToken.Bitcast(TokenType.ULong);
DataToken resultDoubleToken = ulongToken.Bitcast(TokenType.Double);
Debug.Log($"{doubleToken} - 0x{ulongToken:02X} - {resultDoubleToken}");
}
/// <summary>
/// An example function which serializes a pre-determined set of data into a byte array
/// </summary>
/// <param name="intValue">Integer which will be encoded into the output</param>
/// <param name="doubleValue">Double which will be encoded into the output</param>
/// <param name="stringValue">String which will be encoded into the output</param>
/// <param name="floatArrayValues">Float array which will be encoded into the output</param>
/// <returns></returns>
byte[] Serialize(int intValue, double doubleValue, string stringValue, float[] floatArrayValues)
{
int size = 0;
byte[] intBytes = BitConverter.GetBytes(intValue); //Convert int to bytes
size += intBytes.Length;
byte[] doubleBytes = BitConverter.GetBytes(doubleValue); //Convert int to bytes
size += doubleBytes.Length;
byte[] stringBytes = Encoding.UTF8.GetBytes(stringValue); //Convert string to bytes
size += stringBytes.Length;
Debug.Log($"String byte length {stringBytes.Length}");
byte[] stringLengthBytes = BitConverter.GetBytes(stringBytes.Length); //Convert string length to bytes
size += stringLengthBytes.Length;
byte[] floatArrayLengthBytes = BitConverter.GetBytes(Buffer.ByteLength(floatArrayValues));
size += floatArrayLengthBytes.Length;
//It is not necessary to convert the float array into a byte array because we can BlockCopy it directly
size += Buffer.ByteLength(floatArrayValues);
byte[] output = new byte[size]; //Allocate an array of the correct size that should fit all of our items
int offset = 0;
Buffer.BlockCopy(intBytes, 0, output, offset, intBytes.Length); //Write int - this should take up 4 bytes
offset += intBytes.Length; //Increment offset so the next item can write to the correct location
Buffer.BlockCopy(doubleBytes, 0, output, offset, doubleBytes.Length); //Write double - this should take up 8 bytes
offset += doubleBytes.Length;
Buffer.BlockCopy(stringLengthBytes, 0, output, offset, stringLengthBytes.Length); //Write the length of the string so the decoder knows how much to decode - this should take up 4 bytes
offset += stringLengthBytes.Length;
Buffer.BlockCopy(stringBytes, 0, output, offset, stringBytes.Length); //Write string - this is variable, which is why we need to encode the length of the string above
offset += stringBytes.Length;
Buffer.BlockCopy(floatArrayLengthBytes, 0, output, offset, floatArrayLengthBytes.Length); //Write the length of the float array so the decoder knows how much to decode - this should take up 4 bytes
offset += floatArrayLengthBytes.Length;
Buffer.BlockCopy(floatArrayValues, 0, output, offset, Buffer.ByteLength(floatArrayValues)); //Write float array - this can be done directly without a byte array conversion because it's already an Array
offset += Buffer.ByteLength(floatArrayValues);
Debug.Log($"Encoded data in {output.Length} bytes");
return output;
}
/// <summary>
/// An example function which deserializes a pre-determined set of data described in the Serialize function above
/// </summary>
/// <param name="input">Input bytes - must be formatted in the expected manner by the Serialize function above</param>
/// <param name="intValue">Output int value deserialized from the data inside the input</param>
/// <param name="doubleValue">Output double value deserialized from the data inside the input</param>
/// <param name="stringValue">Output string value deserialized from the data inside the input</param>
/// <param name="floatArrayValues">Output float array value deserialized from the data inside the input</param>
/// <returns></returns>
bool Deserialize(byte[] input, out int intValue, out double doubleValue, out string stringValue, out float[] floatArrayValues)
{
int offset = 0;
intValue = BitConverter.ToInt32(input, offset);
offset += 4; //Increment the offset so the next item reads from the correct location. Ints should be 4 bytes long
doubleValue = BitConverter.ToDouble(input, offset);
offset += 8; //Doubles should be 8 bytes long
int stringLength = BitConverter.ToInt32(input, offset);
offset += 4; //String length is an int, which should be 4 bytes long
Debug.Log($"Decoding string length {stringLength} at offset {offset} for buffer length {input.Length}");
stringValue = Encoding.UTF8.GetString(input, offset, stringLength);
offset += stringLength; //Strings are variable length
int floatArrayByteLength = BitConverter.ToInt32(input, offset);
offset += 4; //Float array length is an int, which should be 4 bytes long
floatArrayValues = new float[floatArrayByteLength / 4]; //Create a new float array of the correct size to receive the data
Buffer.BlockCopy(input, offset, floatArrayValues, 0, floatArrayByteLength); //Copy the data from the input into the float array
return true;
}
}