C#•2w ago

Get 16 bit floating point precision as a ushort

I saw on SO that using Convert.ToUInt16(2.2) would give me the 16 bit representation of the float, but instead it just gives me 2, what is the proper way to do what i want?

31 Replies

Angius•2w ago

Not 100% sure what you want to achieve here

prxxssOP•2w ago

i want 2.2 to be represented as 0x4066 that is the IEEE754 representation of a 16 bit floating point number

MODiX•2w ago

Angius

REPL Result: Success

var bytes = BitConverter.GetBytes(2.2f);
var s = "";

foreach (var b in bytes)
  s += b.ToString("X2");

s

var bytes = BitConverter.GetBytes(2.2f);
var s = "";

foreach (var b in bytes)
  s += b.ToString("X2");

s

Result: string

CDCC0C40

CDCC0C40

Compile: 495.085ms | Execution: 40.475ms | React with ❌ to remove this embed.

Angius•2w ago

Huh, not quite the expected result

prxxssOP•2w ago

😭

Angius•2w ago

That said... not sure how can you represent a float as an ushort One has decimal points... another doesn't

prxxssOP•2w ago

maybe i can use a span :p

Angius•2w ago

Unless you do float -> binary -> ushort, to get an ushort value of the binary representation of the float Which would be... largely pointless, but possible

prxxssOP•2w ago

well the goal is float -> binary string but a 16 bit binary string

Angius•2w ago

Floats are 32-bit

prxxssOP•2w ago

are there 16 bit floats then? i mean like a "half" type floats are singles

Angius•2w ago

Not in C#, no float is 32 bits, double is 64, decimal is 128 Ah, wait https://devblogs.microsoft.com/dotnet/introducing-the-half-type/ https://learn.microsoft.com/en-us/dotnet/api/system.half?view=net-8.0 Available since .NET 5

prxxssOP•2w ago

now i need to figure out a way to convert a float to a half?

Angius•2w ago

You will lose precision doing so

prxxssOP•2w ago

yeah im fine with that this doesnt seem to work

Convert.ToUInt16((Half)2.2)

Convert.ToUInt16((Half)2.2)

Unhandled exception. System.InvalidCastException: Unable to cast object of type 'System.Half' to type 'System.IConvertible'.
   at System.Convert.ToUInt16(Object value)
   at MapEncoder.EncodePosDword(Single x, Single y) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\MapEncoder.cs:line 18
   at MapEncoder.EncodeNotes(List`1 notes) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\MapEncoder.cs:line 59
   at MapEncoder.Encode(Map map) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\MapEncoder.cs:line 90
   at Program.<Main>$(String[] args) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\Program.cs:line 30

Unhandled exception. System.InvalidCastException: Unable to cast object of type 'System.Half' to type 'System.IConvertible'.
   at System.Convert.ToUInt16(Object value)
   at MapEncoder.EncodePosDword(Single x, Single y) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\MapEncoder.cs:line 18
   at MapEncoder.EncodeNotes(List`1 notes) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\MapEncoder.cs:line 59
   at MapEncoder.Encode(Map map) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\MapEncoder.cs:line 90
   at Program.<Main>$(String[] args) in C:\Users\antho\Documents\Code\CSHARP\BitWriter\Program.cs:line 30

Angius•2w ago

Well then, look into the Create methods You can't represent a floating point number as an integer Best case scenario it gets rounded, worst case it gets truncated

prxxssOP•2w ago

oh wait, i think the issue is that ToUInt16 takes in a generic of IConvertable but half doesnt support it

blueberriesiftheywerecats•2w ago

just do Half h = (Half)2.2f

prxxssOP•2w ago

yeah that works, its the fact that i cant convert it to a ushort

blueberriesiftheywerecats•2w ago

Its 16 bit now, why convert it

prxxssOP•2w ago

i need to write the binary version of it to a file well not to a file to a string i need to convert a half to a binary string

Angius•2w ago

Do that, then Where does the ushort come in?

prxxssOP•2w ago

because i cant directly convert a half to a binary string

blueberriesiftheywerecats•2w ago

Convert.ToString(half, 2);

prxxssOP•2w ago

uh huh

Angius•2w ago

theHalf.ToString("X2")

blueberriesiftheywerecats•2w ago

bool?

prxxssOP•2w ago

thats hex right? bool is the first overload for ToString, it means there isnt an overload supporting Half

Angius•2w ago

You wanted a hexadecimal string, no?

blueberriesiftheywerecats•2w ago

Half implements IBinaryNumber Maybe casting it to that will help you

prxxssOP•2w ago

huh?

Unhandled exception. System.FormatException: Format specifier was invalid.

Unhandled exception. System.FormatException: Format specifier was invalid.

string rep = v.ToString("X2").PadLeft(size, '0');

string rep = v.ToString("X2").PadLeft(size, '0');

i dont think x2 can be used on a non-integer type i found

BitConverter.HalfToUInt16Bits

BitConverter.HalfToUInt16Bits

it works, thanks guys for the brainstorming

Gaming

Programming

Get 16 bit floating point precision as a ushort