[DiscordArchive] They are dividing by 0x7FFF there?
[DiscordArchive] They are dividing by 0x7FFF there?
Original source
You said it fails to read most values
rektbyfaith
06-06-2018, 08:28 AM #11Archived author: Skarn • Posted: 2018-06-06T08:28:19.453000+00:00
Original source
You said it fails to read most values
Original source
It tried to handle fixed16 and fixed_point the same, which is wrong.
The part it mostly failed with was writing iirc, with off by one rounding errors.
rektbyfaith
06-06-2018, 08:30 AM #12Archived author: schlumpf • Posted: 2018-06-06T08:30:28.704000+00:00
Original source
It tried to handle fixed16 and fixed_point the same, which is wrong.
The part it mostly failed with was writing iirc, with off by one rounding errors.
Original source
```py
class fixed_point:
"""A fixed point real number, opposed to a floating point."""
def __init__(self, type_, dec_bits, int_bits):
self.type = type_
self.dec_bits = dec_bits
self.int_bits = int_bits
self.value = 0
def read(self, f):
fixed_point_val = self.type.read(f)
decimal_part = fixed_point_val & ((1 << self.dec_bits) - 1)
integral_part = (fixed_point_val >> self.dec_bits) & (1 << self.int_bits) - 1
sign = -1.0 if (fixed_point_val & (1 << (self.dec_bits + self.int_bits)) != 0) else 1.0
self.value = sign * (integral_part + decimal_part / (((1 << self.dec_bits) - 1) + 1.0))
return self
def write(self, f):
sign = 1 if self.value < 0 else 0
integral_part = int(self.value) & ((1 << self.int_bits) - 1)
decimal_part = int((self.value - int(self.value)) * (1 << self.dec_bits))
fixed_point_val = (sign << (self.int_bits + self.dec_bits)) | (integral_part << self.int_bits) | decimal_part
self.type.write(f, fixed_point_val)
return self
```
rektbyfaith
06-06-2018, 08:30 AM #13Archived author: Skarn • Posted: 2018-06-06T08:30:43.931000+00:00
Original source
```py
class fixed_point:
"""A fixed point real number, opposed to a floating point."""
def __init__(self, type_, dec_bits, int_bits):
self.type = type_
self.dec_bits = dec_bits
self.int_bits = int_bits
self.value = 0
def read(self, f):
fixed_point_val = self.type.read(f)
decimal_part = fixed_point_val & ((1 << self.dec_bits) - 1)
integral_part = (fixed_point_val >> self.dec_bits) & (1 << self.int_bits) - 1
sign = -1.0 if (fixed_point_val & (1 << (self.dec_bits + self.int_bits)) != 0) else 1.0
self.value = sign * (integral_part + decimal_part / (((1 << self.dec_bits) - 1) + 1.0))
return self
def write(self, f):
sign = 1 if self.value < 0 else 0
integral_part = int(self.value) & ((1 << self.int_bits) - 1)
decimal_part = int((self.value - int(self.value)) * (1 << self.dec_bits))
fixed_point_val = (sign << (self.int_bits + self.dec_bits)) | (integral_part << self.int_bits) | decimal_part
self.type.write(f, fixed_point_val)
return self
```
Original source
That's the code I am talking about btw
rektbyfaith
06-06-2018, 08:30 AM #14Archived author: Skarn • Posted: 2018-06-06T08:30:48.697000+00:00
Original source
That's the code I am talking about btw
Original source
Can you propose any solution to rounding though?
rektbyfaith
06-06-2018, 08:31 AM #15Archived author: Skarn • Posted: 2018-06-06T08:31:09.150000+00:00
Original source
Can you propose any solution to rounding though?
Original source
The code tries to implement fp as if it was fixed16 for read, so that will be wrong for any non fixed16. I can’t comment on my write function.
rektbyfaith
06-06-2018, 08:33 AM #16Archived author: schlumpf • Posted: 2018-06-06T08:33:09.905000+00:00
Original source
The code tries to implement fp as if it was fixed16 for read, so that will be wrong for any non fixed16. I can’t comment on my write function.
Original source
TLDR: ignore me and just do what others said above
rektbyfaith
06-06-2018, 08:33 AM #17Archived author: schlumpf • Posted: 2018-06-06T08:33:23.044000+00:00
Original source
TLDR: ignore me and just do what others said above
Original source
<@250706991515828224> is your viewer capable of showing individual models like WMV?
rektbyfaith
06-06-2018, 08:37 AM #18Archived author: Skarn • Posted: 2018-06-06T08:37:28.058000+00:00
Original source
<@250706991515828224> is your viewer capable of showing individual models like WMV?
Original source
<@197724887387734016> yes
rektbyfaith
06-06-2018, 08:37 AM #19Archived author: Deamon • Posted: 2018-06-06T08:37:50.196000+00:00
Original source
<@197724887387734016> yes
Original source
the UI though is not there
rektbyfaith
06-06-2018, 08:38 AM #20Archived author: Deamon • Posted: 2018-06-06T08:38:07.080000+00:00
Original source
the UI though is not there
1 Guest(s)
1 Guest(s)