Rev 297 | Rev 335 | Go to most recent revision | Details | Compare with Previous | Last modification | View Log | RSS feed
Rev | Author | Line No. | Line |
---|---|---|---|
141 | dmarschall | 1 | |
294 | daniel-mar | 2 | |
221 | daniel-mar | 3 | Implementation detail differences |
4 | ================================= |
||
141 | dmarschall | 5 | |
304 | daniel-mar | 6 | Filter Foundry tries to be as compatible with Filter Factory as possible. |
7 | However, there are some differences that are explained in this documentation. |
||
141 | dmarschall | 8 | |
294 | daniel-mar | 9 | Various implementations |
10 | ----------------------- |
||
141 | dmarschall | 11 | |
294 | daniel-mar | 12 | In the source-code file funcs.c, some functions are implemented twice: |
13 | One instance is the default implementation (older Filter Foundry versions used), |
||
14 | and one instance is a 100% replica of the Filter Factory code, obtained |
||
15 | from the "OPER" resource. |
||
16 | (More information at https://misc.daniel-marschall.de/projects/filter_factory/res_oper.html ) |
||
297 | daniel-mar | 17 | If required, the compiler-definitions `use_filterfactory_implementation_*` |
294 | daniel-mar | 18 | can be set or unset to select the implementation. |
19 | |||
304 | daniel-mar | 20 | In Filter Foundry 1.7.0.8, the following functions have been updated to the Filter Factory replica: |
297 | daniel-mar | 21 | - `rnd(x)` |
22 | - `cos(x)` |
||
23 | - `sin(x)` |
||
24 | - `tan(x)` |
||
25 | - `r2x(d,m)` |
||
26 | - `r2y(d,m)` |
||
27 | - `rad(d,m,z)` |
||
28 | - `c2d(x,y)` |
||
29 | - `c2m(x,y)` |
||
30 | - `sqr(x)` |
||
31 | - `d` |
||
32 | - `m` |
||
33 | - `M` |
||
294 | daniel-mar | 34 | |
35 | |||
36 | sqr(x) |
||
37 | ------ |
||
38 | |||
39 | Filter Factory: |
||
40 | |||
41 | sqr(x)=x for x < 0 |
||
295 | daniel-mar | 42 | |
304 | daniel-mar | 43 | Can be tested with the following expression: |
295 | daniel-mar | 44 | sqr(-20)+21 == 1 |
294 | daniel-mar | 45 | |
46 | Filter Foundry (prior to 1.7.0.8): |
||
47 | |||
48 | sqr(x)=0 for x < 0 |
||
49 | |||
304 | daniel-mar | 50 | Beginning with Filter Foundry 1.7.0.8, the behavior of Filter Factory was implemented. |
297 | daniel-mar | 51 | |
304 | daniel-mar | 52 | |
158 | dmarschall | 53 | i, u, v (Testcase iuv.afs) |
141 | dmarschall | 54 | ------- |
55 | |||
56 | Filter Foundry <1.7 uses the same formulas as in Filter Factory: |
||
57 | |||
221 | daniel-mar | 58 | i=((76*r)+(150*g)+(29*b))/256 // Output range is 0..254 |
59 | u=((-19*r)+(-37*g)+(56*b))/256 // Output range is -55..55 |
||
60 | v=((78*r)+(-65*g)+(-13*b))/256 // Output range is -77..77 |
||
141 | dmarschall | 61 | |
62 | Filter Foundry 1.7 uses more accurate formulas: |
||
63 | |||
221 | daniel-mar | 64 | i=(299*r+587*g+114*b)/1000 // Output range is 0..255 |
65 | u=(-147407*r-289391*g+436798*b)/2000000 // Output range is -55..55 |
||
66 | v=614777*r-514799*g-99978*b)/2000000 // Output range is -78..78 |
||
141 | dmarschall | 67 | |
192 | daniel-mar | 68 | Both formulas follow the same YUV standard but have different accuracy. |
141 | dmarschall | 69 | |
70 | |||
304 | daniel-mar | 71 | I, U, V, imin, umin, vmin (Testcase iuv_minmax.afs) |
72 | ------------------------- |
||
73 | |||
74 | In Filter Foundry 1.7.0.8, the previously undocumented variables I, U, V as well as imin, umin, vmin |
||
75 | have been changed to represent the actual results of the i,u,v variables: |
||
76 | |||
77 | I, imax = 255 (stayed the same) |
||
78 | U, umax = 55 (was 255 in Filter Factory) |
||
79 | V, vmax = 78 (was 255 in Filter Factory) |
||
80 | |||
81 | imin = 0 (stayed the same) |
||
82 | umin = -55 (was 0 in Filter Factory) |
||
83 | vmin = -78 (was 0 in Filter Factory) |
||
84 | |||
85 | |||
141 | dmarschall | 86 | get(i) (Testcase getput.afs) |
87 | ------ |
||
88 | |||
89 | Filter Foundry: |
||
90 | |||
221 | daniel-mar | 91 | get(x)=0 if x>255 or x<0 |
92 | |||
141 | dmarschall | 93 | Filter Factory: |
94 | |||
221 | daniel-mar | 95 | get(x)=x if x>255 or x<0 |
141 | dmarschall | 96 | |
221 | daniel-mar | 97 | Note: The result "x" was most likely not intended but a result of an undefined behavior |
98 | |||
99 | |||
141 | dmarschall | 100 | r, g, b at empty canvas (Testcase emptycanvas.afs) |
101 | ----------------------- |
||
102 | |||
221 | daniel-mar | 103 | In Filter Factory, an empty (transparent) canvas of a new file is initialized as `r=g=b=0` |
141 | dmarschall | 104 | |
221 | daniel-mar | 105 | Filter Foundry initializes it as `r=g=b=255` |
141 | dmarschall | 106 | |
107 | |||
294 | daniel-mar | 108 | rnd(a,b) (Testcases rnd*) |
289 | daniel-mar | 109 | -------- |
110 | |||
111 | Filter Factory uses Donald E. Knuth's subtractive random number generator algorithm, |
||
112 | which has been published in "The Art of Computer Programming, volume 2: Seminumerical Algorithms". |
||
113 | Addison-Wesley, Reading, MA, second edition, 1981. |
||
114 | |||
115 | Beginning with Filter Foundry 1.7.0.8, the same PRNG was implemented, |
||
116 | so that the output of rnd(a,b) is exactly the same now. |
||
117 | |||
118 | |||
294 | daniel-mar | 119 | rst(i) (Testcases rnd*.afs and rst_*.afs) |
120 | ------ |
||
141 | dmarschall | 121 | |
289 | daniel-mar | 122 | Filter Factory contains an undocumented function that sets the seed for the random number generator. |
141 | dmarschall | 123 | |
304 | daniel-mar | 124 | Filter Factory and Filter Foundry beginning with 1.7.0.8 accept a seed between 0 and 32767, inclusively. |
125 | If the argument is not within this range, the operation lowest 15 bits are taken. |
||
141 | dmarschall | 126 | |
304 | daniel-mar | 127 | There are many differences in the implementation between Filter Factory and Filter Foundry in regards rst(i): |
141 | dmarschall | 128 | |
289 | daniel-mar | 129 | **Filter Factory:** |
143 | dmarschall | 130 | |
289 | daniel-mar | 131 | If rst(i) is called in Filter Factory, an internal Seed-Variable is set. |
132 | It does NOT influence any calls of rnd(a,b), because a lookup-table needs to be built first. |
||
133 | The building of the lookup-table is probably done before the processing of the first pixel (x,y,z=0). |
||
134 | It is suspected that the call of rst(i) will take effect on the next calculation. |
||
135 | Due to a bug (or feature?), the random state is not reset to its initial state (0) before the |
||
136 | filter is applied. The preview image processing will modify the random state, and once the filter |
||
137 | is actually applied (pressing "OK"), the random state that was set in the preview picture, will be used. |
||
138 | This could be considered as a bug, but it is probably required, otherwise the call of rst(i) |
||
139 | (inside the preview calculation) won't affect the rnd(a,b) in the real run. |
||
140 | However, in a standalone filter without dialog/preview, there is no preview that could set |
||
141 | the internal seed, so the rnd(a,b) functions will always work using the default seed 0, |
||
142 | and only the subsequent calls will use the rst(i) of the previous call. |
||
221 | daniel-mar | 143 | |
289 | daniel-mar | 144 | **Filter Foundry:** |
143 | dmarschall | 145 | |
289 | daniel-mar | 146 | In Filter Foundry, the function rnd(a,b) retrieves a random number in "realtime"; therefore, if the |
147 | seed is changed via rst(i), there is an immediate effect on the next call of the rnd(a,b) function. |
||
148 | |||
304 | daniel-mar | 149 | For example, the following filter would generate an one-colored picture without any randomness: |
221 | daniel-mar | 150 | R: rst(123), rnd(0,255) |
151 | G: rnd(0,255) |
||
152 | B: rnd(0,255) |
||
144 | dmarschall | 153 | |
289 | daniel-mar | 154 | If you want to generate a random pixel image with a non-default seed, you need to make sure |
155 | that rst(i) is called only once at the beginning (channel 0, coordinate 0|0): |
||
294 | daniel-mar | 156 | R: (x== 0 && y ==0) ? rst(123) : 0, rnd(0,255) |
221 | daniel-mar | 157 | G: rnd(0,255) |
158 | B: rnd(0,255) |
||
144 | dmarschall | 159 | |
289 | daniel-mar | 160 | In Filter Foundry, rst(i) can be called by branches and variables/sliders can |
161 | be used as arguments of rst(i). |
||
144 | dmarschall | 162 | |
163 | |||
143 | dmarschall | 164 | Evaluation of conditional branches |
165 | ---------------------------------- |
||
166 | |||
221 | daniel-mar | 167 | **Filter Foundry:** |
144 | dmarschall | 168 | |
221 | daniel-mar | 169 | Only the branches which will be chosen due to the conditional expression will be evaluated. |
170 | |||
304 | daniel-mar | 171 | This means that the following filter would generate a black canvas: (Testcase conditional_eval_1.afs) |
221 | daniel-mar | 172 | |
173 | R: 1==0 ? put(255,0) : 0 |
||
174 | G: get(0) |
||
175 | B: 0 |
||
176 | |||
177 | In boolean expressions, the evaluation will be aborted if the result is already determined. |
||
178 | |||
179 | So, this will also generate a black canvas: (Testcase conditional_eval_2.afs) |
||
180 | |||
181 | R: 1==0 && put(255,0) ? 0: 0 |
||
182 | G: get(0) |
||
183 | B: 0 |
||
184 | |||
185 | This will also generate a black canvas: (Testcase conditional_eval_3.afs) |
||
186 | |||
187 | R: 1==1 || put(255,0) ? 0 : 0 |
||
188 | G: get(0) |
||
189 | B: 0 |
||
190 | |||
191 | **Filter Factory:** |
||
192 | |||
304 | daniel-mar | 193 | Each branch inside an if-then-else expression will be evaluated. |
194 | This means that the following filter would generate a green canvas: (Testcase conditional_eval_1.afs) |
||
221 | daniel-mar | 195 | |
196 | R: 1==0 ? put(255,0) : 0 |
||
197 | G: get(0) |
||
198 | B: 0 |
||
199 | |||
304 | daniel-mar | 200 | Also, all arguments of a boolean expression will be fully evaluated. |
221 | daniel-mar | 201 | So, this will also generate a green canvas: (Testcase conditional_eval_2.afs) |
202 | |||
203 | R: 1==0 && put(255,0) ? 0: 0 |
||
204 | G: get(0) |
||
205 | B: 0 |
||
206 | |||
207 | This will also generate a green canvas: (Testcase conditional_eval_3.afs) |
||
208 | |||
209 | R: 1==1 || put(255,0) ? 0 : 0 |
||
210 | G: get(0) |
||
211 | B: 0 |