Corrections for the book Ray Tracing Gems. The pages listed are for the released book itself, not for any preprints or other forms of the articles. A corrected copy of the free PDF version of the book can be found linked from http://raytracinggems.com.
This is the printing currently on sale at the publisher's site, Amazon, and elsewhere.
Chapter 3, page 32: The structure "D3D12_FEATURE_DATA_OPTIONS5" should be "D3D12_FEATURE_DATA_D3D12_OPTIONS5", with a second "D3D12" in there near the end (v1.2)
Chapter 3, page 41: "Shader tables are contiguous blocks of 64-bit aligned GPU memory" to "Shader tables are contiguous blocks of 64-byte aligned GPU memory" - change "bit" to "byte". (v1.6)
Chapter 3, page 41: "Shader identifiers are 32-bit chunks" to "Shader identifiers are 32-byte chunks" - change "bit" to "byte". (v1.6)
Chapter 3, page 43: "Unqiue" to "Unique" in two places, lines 11 and 19, in the code listing. (v1.8)
Chapter 11, page 144: Line 15 did not access the "stack" array; latter half should read "material[stack[prev_same].material_idx]) {". (v1.5)
Chapter 15, page 213: Lines 7 and 8 in the code listing should be (v1.9): 7 return sum_sq / (n - 1) - 8 sum * sum / ((n - 1) * n);
Chapter 16, page 226: "float SampleLinear( float a, float b ) {" should have a "float u" parameter for the random variable, i.e., "float SampleLinear( float u, float a, float b ) {" (upcoming: v1.7)
Chapter 16, page 235: The case "a == 0 && b == 0" is not handled and can result in division by 0; above line 3 add "if (b == 0) b = 1;" (v1.2)
Chapter 16, page 241: In Equation 11 change "s+1" to "s+2"; the Phong-like PDF doesn't integrate to 1 over the hemisphere if you include the projection cosine (v1.2)
Chapter 16, page 242: Line 1 of the code at the top, change "1+s" to "2+s" (v1.2)
Chapter 19, page 302: In Equation 1 the first integral is missing (Latex form) "d\omega_i" at the end (v1.2)
Chapter 25, page 442: code line 54, add to RAY_FLAG_SKIP_CLOSEST_HIT_SHADER the flag RAY_FLAG_ACCEPT_FIRST_HIT_AND_END_SEARCH, i.e., "RAY_FLAG_SKIP_CLOSEST_HIT_SHADER | RAY_FLAG_ACCEPT_FIRST_HIT_AND_END_SEARCH". (v1.7)
Chapter 25, page 442: seventh line from bottom, "we do not require any-hit shader results" to "we do not require hit shader results, and that the first ray-primitive intersection encountered is sufficient to know if we’re in shadow or not." Note the change of "any-hit" to "hit". (v1.7)
Chapter 27, page 502: Listing 27-2's caption is something of a copy of 27-4. It should read, "This short closest-hit shader code snippet implements ambient occlusion (AO) lighting with a limited occlusion distance, permitting AO lighting to be used within confined or fully enclosed spaces that would otherwise result in 100% occlusion and full shadow." (v1.6)
Chapter 3, page 43: Lines 14 and 27 of the code listing use "(pData + 32)" to avoid book formatting problems. The verbose, but more informative, code "(pData + D3D12_SHADER_IDENTIFIER_SIZE_IN_BYTES)" could be used instead. (v1.6)
Chapter 10, page 133: While easy enough to find by searching, the URL for the first reference is http://aggregate.org/MAGIC/ (v1.2)
Chapter 12, page 154: "Dupy" to "Dupuy" (v1.2)
Sampling introduction, page 205: There is a comma before and after in: GPU,”, - remove the second comma (v1.4)
Chapter 16, page 227: "Box-Müller" to "Box-Muller", no umlaut (v1.2)
Chapter 19, page 290: "ILMxLabs" to "ILMxLAB" (v1.2)
Chapter 19, page 318: For reference [2] the names should read "Heitz, E., Dupuy, J., Hill, S., and Neubelt, D." (v1.2)
Chapter 25, page 443: "kernelfootprint" to "kernel footprint" (v1.2)
Chapter 25, page 466: Remove editorial note "// Looks good to me, and I added a comma. /Eric" (v1.3)
A few hundred copies of the book were rushed to print in order to be available for the GDC and GTC 2019 conferences. The erratum listed here is fixed in the commercial hardcover, Kindle, Google Book, and PDF editions.
Foreword, page xvii: The passage "photons travel through the is why its introduction" is missing a line between "the" and "is". The whole passage should be "There is no substitute for this capability when the goal is photorealism, where we need to determine the complicated paths along which photons travel through the virtual world. Ray tracing is a fundamental ingredient of realistic rendering, which is why its introduction to the real-time domain was such a significant step for computer graphics." (v1.2)
Thanks to Stephen Hill, Adam Marrs, Daniel Seibert, Martin Stich, Ebor Folkertsma, Mauricio Vives, and Filip Strugar for reporting these errors.
Page last updated May 6, 2021