I have two 16bit integer raw data.
For example:
High Word = 17142 (dec) or 0100001011110110 (binary)
Low Word = 59759 (dec) or 1110100101111001 (binary)
If you treat two word together as one 32bit float data, it will be "123.456"
Binary --> 01000010111101101110100101111001
How to convert integer array [59759 , 17142] to float 123.456 in Javascript?
Note: [ X (16bit low word) , Y (16 bit high word) ] ==> Z (32bit float)
You can do this with typed arrays and ArrayBuffer, which allow you to interpret the same bits in different ways (but endianness is platform-specific). It's also possible using a DataView on the buffer, which lets you conrol endianness.
Here's the typed array approach which works with the endianness of my platform, see comments:
// Create a buffer
var buf = new ArrayBuffer(4);
// Create a 16-bit int view of it
var ints = new Uint16Array(buf);
// Fill in the values
ints[0] = 59759;
ints[1] = 17142;
// Create a 32-bit float view of it
var floats = new Float32Array(buf);
// Read the bits as a float; note that by doing this, we're implicitly
// converting it from a 32-bit float into JavaScript's native 64-bit double
var num = floats[0];
// Done
console.log(num);Here's the DataView approach, note writing the ints in the opposite order:
// Create a buffer
var buf = new ArrayBuffer(4);
// Create a data view of it
var view = new DataView(buf);
// Write the ints to it
view.setUint16(0, 17142);
view.setUint16(2, 59759);
// Read the bits as a float; note that by doing this, we're implicitly
// converting it from a 32-bit float into JavaScript's native 64-bit double
var num = view.getFloat32(0);
// Done
console.log(num);
Can shorter / make more efficient @T.J. Crowder answer by using array buffer directly :
var data = [59759, 17142];
// Create a buffer
var buf = new Uint16Array(data).buffer
// Create a data view of it
var view = new DataView(buf);
var num = view.getFloat32(0, true);
// Done
console.log(num);
