1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
|
<?php
// Mirrored from https://github.com/Danack/HexFloat
require_once __DIR__ . "/FloatInfo.php";
require_once __DIR__ . "/Float32Info.php";
use HexFloat\FloatInfo;
use HexFloat\Float32Info;
/**
* Returns a string containing a hexadecimal representation of the given float,
* using 64 bits of info
*
* @param float $number
* @return string
*/
function floathex(float $number)
{
return strrev(unpack('h*', pack('d', $number))[1]);
}
/**
* Returns a string containing a hexadecimal representation of the given float,
* using 32 bits of info
*
* @param float $number
* @return string
*/
function floathex32(float $num)
{
return strrev(unpack('h*', pack('f', $num))[1]);
}
/**
* Convert a floating point number to a FloatInfo object,
* which contains string representations of the float's sign,
* exponent and mantissa
* @param float $num
* @return FloatInfo
*/
function float_info(float $num)
{
$float64 = floathex($num);
//Sign bit: 1 bit
//Exponent: 11 bits
//Significand precision: 53 bits (52 explicitly stored)
$chars = str_split($float64);
// 3 bits from this
$byte1 = hexdec($chars[0]);
// 4 bits from this
$byte2 = hexdec($chars[1]);
// 1 bit from this
$byte3 = hexdec($chars[2]);
$sign = '0';
if ($byte1 >= 8) {
$sign = '1';
}
$exponentString = substr($float64, 0, 3);
$exponentValue = hexdec($exponentString) & 0x7ff;
$exponent = sprintf("%b", $exponentValue);
$exponent = str_pad($exponent, 11, '0', STR_PAD_LEFT);
$mantissa = substr($float64, 2);
$mantissa = hexdec($mantissa) & 0xfffffffffffff;
$mantissa = sprintf("%b", $mantissa);
$mantissa = str_pad($mantissa, 52, '0', STR_PAD_LEFT);
return new FloatInfo(
$sign,
$exponent,
$mantissa
);
}
/**
* Convert a floating point number to a Float32Info object,
* which contains string representations of the float's sign,
* exponent and mantissa
*
* @param float $num
* @return Float32Info
*/
function float_info_32(float $num)
{
$float32 = floathex32($num);
$chars = str_split($float32);
// 3 bits from this
$byte1 = hexdec($chars[0]);
// 4 bits from this
$byte2 = hexdec($chars[1]);
// 1 bit from this
$byte3 = hexdec($chars[2]);
$sign = '0';
if ($byte1 >= 8) {
$sign = '1';
}
$exponent3Bits = ($byte1 & 0x7);
$exponent4Bits = $byte2;
$exponent1Bit = ($byte3 & 0x8) >> 3;
$exponent = ($exponent3Bits << 5) | ($exponent4Bits << 1) | $exponent1Bit;
$exponent = sprintf("%b", $exponent);
$exponent = str_pad($exponent, 8, '0', STR_PAD_LEFT);
$mantissa = substr($float32, 2, 6);
$mantissa = hexdec($mantissa) & 0x7fffff;
$mantissa = sprintf("%b", $mantissa);
$mantissa = str_pad($mantissa, 23, '0', STR_PAD_LEFT);
return new Float32Info(
$sign,
$exponent,
$mantissa
);
}
/**
* Produce a debug string that shows the Sign, Exponent and Mantissa for
* two floating point numbers, using 64bit precision
*
*
* @param float $value1
* @param float $value2
* @return string
*
* Example result
* ┌──────┬─────────────┬──────────────────────────────────────────────────────┐
* │ Sign │ Exponent │ Mantissa │
* │ 0 │ 01111111011 │ 1001100110011001100110011001100110011001100110011010 │
* │ 0 │ 10000011001 │ 0111110101111000010000000100000000000000000000000000 │
* └──────┴─────────────┴──────────────────────────────────────────────────────┘
*
*/
function float_compare(float $value1, float $value2)
{
$float_info_1 = float_info($value1);
$float_info_2 = float_info($value2);
//Sign bit: 1 bit
//Exponent: 11 bits
//Significand precision: 53 bits (52 explicitly stored)
$output = "┌──────┬─────────────┬──────────────────────────────────────────────────────┐\n";
$output .= "│ Sign │ Exponent │ Mantissa │\n";
$format_string = "│ %s │ %s │ %s │\n";
$output .= sprintf($format_string, $float_info_1->getSign(), $float_info_1->getExponent(), $float_info_1->getMantissa());
$output .= sprintf($format_string, $float_info_2->getSign(), $float_info_2->getExponent(), $float_info_2->getMantissa());
$output .= "└──────┴─────────────┴──────────────────────────────────────────────────────┘\n";
return $output;
}
/**
* Produce a debug string that shows the Sign, Exponent and Mantissa for
* two floating point numbers, using 32bit precision
*
* @param float $value1
* @param float $value2
* @return string
*
* Example result
* ┌──────┬──────────┬─────────────────────────┐
* │ Sign │ Exponent │ Mantissa │
* │ 0 │ 01111011 │ 10011001100110011001101 │
* │ 0 │ 10011001 │ 01111101011110000100000 │
* └──────┴──────────┴─────────────────────────┘
*
*/
function float_compare_32(float $value1, float $value2)
{
$float_info_1 = float_info_32($value1);
$float_info_2 = float_info_32($value2);
$output = "┌──────┬──────────┬─────────────────────────┐\n";
$output .= "│ Sign │ Exponent │ Mantissa │\n";
$format_string = "│ %s │ %s │ %s │\n";
$output .= sprintf($format_string, $float_info_1->getSign(), $float_info_1->getExponent(), $float_info_1->getMantissa());
$output .= sprintf($format_string, $float_info_2->getSign(), $float_info_2->getExponent(), $float_info_2->getMantissa());
$output .= "└──────┴──────────┴─────────────────────────┘\n";
return $output;
}
/**
* So. One of the disadvantages of non-HDRI compiled Image Magick
* is that it can't accurately represent a '50%' color accurately.
*
* For example, if ImageMagick is compiled with 16bit color depth
* then the two closest colors to midpoint are:
* 32767 / 65535 = 0.5 - (1 / (2 ^ 17)) = 0.499992370...
* or
* 32768 / 65535 = 0.5 + (1 / (2 ^ 17)) = 0.50000762951
*
* Either way there is going to be 'error' of
* 0.00000762939453125
*
* The problem is even worse when ImageMagick is compiled with 8-bit
* numbers (though this really shouldn't be used any more) and the
* error would be 0.001953125
*
*/
function get_epsilon_for_off_by_half_errors()
{
// These could be defined better...
$epsilon_values_for_non_hdri = [
'255' => (1 / (pow(2, 8) - 1)) + 0.0000000000001,
'65535' => (1 / (pow(2, 16) - 1)) + 0.0000000000001,
'16777215' => (1 / (pow(2, 24) - 1) ) + 0.0000000000001,
'4294967295' => (1 / (pow(2, 32) - 1)) + 0.0000000000001,
];
// These could definitely be defined better...
$epsilon_values_for_hdri = [
'255' => 0.0000000000001,
'65535' => 0.0000000000001,
'16777215' => 0.0000000000001,
'4294967295' => 0.0000000000001
];
if (Imagick::getHdriEnabled() === false) {
$quantum = (string)Imagick::getQuantum();
if (array_key_exists($quantum, $epsilon_values_for_non_hdri) !== true) {
throw new Exception(
"Quantum values is $quantum which is not any of (2^(8|16|24|32)) - 1. Please report this as a bug."
);
}
return $epsilon_values_for_non_hdri[$quantum];
}
$quantum = Imagick::getQuantum();
if (array_key_exists($quantum, $epsilon_values_for_hdri) !== true) {
throw new Exception(
"Quantum values is $quantum which is not any of (2^(8|16|24|32)) - 1. Please report this as a bug."
);
}
return $epsilon_values_for_hdri[$quantum];
}
|