[book 3][ch 6.1] Image doesn't match the settings

[genmeblog]

When I change samples per pixel from 64 to 100 (as it is in listing 17) my image is still much noisier than Image 3. In my reimplementation 1000 gives the matching result.

[genmeblog]

Rendered examples on my side (possibly buggy):

Image 2 is almost the same as in the book:

Image 3, after changing samples per pixels to 100:

Image 4, imperfect PDF

Image 5, random_in_hemisphere, this is undefined in the book, btw. Anyway the result is strange...

[hollasch]

This work should be accomplished as a result of #863.

[dimitry-ishenko]

When I change samples per pixel from 64 to 100 (as it is in listing 17) my image is still much noisier than Image 3. In my reimplementation 1000 gives the matching result.

Maybe this is an old issue, but Chapter 6.1 now says to increase number of samples per pixel to 1000:

Let’s adjust some parameters for the Cornell box:

 int main() {
     ...
+    cam.samples_per_pixel = 1000;
     ...
 }