[DiscordArchive] <@250706991515828224> what do you mean by multiplying by 2-n?
[DiscordArchive] <@250706991515828224> what do you mean by multiplying by 2-n?
Original source
Looks weird to me, decimal part is not used
rektbyfaith
06-04-2018, 08:30 AM #11Archived author: Skarn • Posted: 2018-06-04T08:30:24.161000+00:00
Original source
Looks weird to me, decimal part is not used
Original source
There's only one part, really (unless they're doing something super abnormal)
rektbyfaith
06-04-2018, 08:30 AM #12Archived author: Quantam • Posted: 2018-06-04T08:30:42.598000+00:00
Original source
There's only one part, really (unless they're doing something super abnormal)
Original source
You're just shifting the binary decimal point
rektbyfaith
06-04-2018, 08:30 AM #13Archived author: Quantam • Posted: 2018-06-04T08:30:51.969000+00:00
Original source
You're just shifting the binary decimal point
Original source
<@197724887387734016> if you look closely it is only off by one due to rounding
rektbyfaith
06-04-2018, 08:31 AM #14Archived author: schlumpf • Posted: 2018-06-04T08:31:19.689000+00:00
Original source
<@197724887387734016> if you look closely it is only off by one due to rounding
Original source
```py
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 * float(integral_part) * 2**-self.dec_bits
```
rektbyfaith
06-04-2018, 08:31 AM #15Archived author: Skarn • Posted: 2018-06-04T08:31:47.890000+00:00
Original source
```py
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 * float(integral_part) * 2**-self.dec_bits
```
Original source
is that correct for reading?
rektbyfaith
06-04-2018, 08:31 AM #16Archived author: Skarn • Posted: 2018-06-04T08:31:55.279000+00:00
Original source
is that correct for reading?
Original source
I use my bitshift stuff and it seems to work good enough
rektbyfaith
06-04-2018, 08:32 AM #17Archived author: Deleted User • Posted: 2018-06-04T08:32:33.045000+00:00
Original source
I use my bitshift stuff and it seems to work good enough
Original source
My suspicion is that there is no sign bit either (it's two's complement), but without values to check I have no proof of that
rektbyfaith
06-04-2018, 08:32 AM #18Archived author: Quantam • Posted: 2018-06-04T08:32:36.049000+00:00
Original source
My suspicion is that there is no sign bit either (it's two's complement), but without values to check I have no proof of that
Original source
32767
rektbyfaith
06-04-2018, 08:33 AM #19Archived author: Skarn • Posted: 2018-06-04T08:33:31.511000+00:00
Original source
32767
Original source
is the default value for transparency
rektbyfaith
06-04-2018, 08:33 AM #20Archived author: Skarn • Posted: 2018-06-04T08:33:39.550000+00:00
Original source
is the default value for transparency
1 Guest(s)
1 Guest(s)