Jump to content

Talk:Gouraud shading

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

The Apple

[edit]

piece about apple is a bit difficult to read. Needs to explain example a bit more clearly without getting too in depth as to what causes a specular highlight. Also, we need a page on specular highlights.


The Rotating Sphere

[edit]

the rotating sphere is very distracting - what about linking to the image instead of displaying it? —Preceding unsigned comment added by 84.58.130.196 (talk) 15:07, 1 September 2007 (UTC)[reply]

the rotating sphere is great illustration of gouraud shading distinct features, so it surely must stay 217.76.195.16 (talk) —Preceding comment was added at 09:08, 27 November 2007 (UTC)[reply]

This is really distracting, we should either slow down the model or have it stationary. Awally88 (talk) 04:12, 20 June 2011 (UTC)[reply]

I agree. Maybe we should upload an new image of the sphere in which it is not rotating. --Kri (talk) 11:59, 22 June 2011 (UTC)[reply]
Said and done. --Kri (talk) 12:20, 22 June 2011 (UTC)[reply]
Undone. Animated images were much more useful when I was trying to understand the concept when I first saw this page. Surely that's more important. --JelloB (talk) 05:42, 15 July 2013 (UTC)[reply]
Without the animation it is less clear what what the article refers to with "... continuously fading in and out across neighboring portions of the model, peaking in intensity when the intended specular highlight passes over a vertex of the model ..." I'm not sure whether the text needs to be changed to take the missing animation into account. --Martin Kraus (talk) 12:39, 22 June 2011 (UTC)[reply]
I agree that the animated image is in this case more or less necessary; if you think it's distracting, scroll it out of view. I do notice another issue though: Wikipedia's thumbnailing algorithm does a very poor job in this case! Notice how the wireframe is not rotating with the sphere. I think someone with the know-how should generate a custom thumbnail that we can include full-size. (And maybe it could spin a bit slower, say half as fast?)

Pronunciation

[edit]

Someone fluent in IPA could add pronunciation info 77.253.68.171 (talk) 11:13, 29 May 2008 (UTC)[reply]

I think it's "Guh row".146.115.34.7 (talk)

All I've heard is "GORE-odd." Psychlohexane (talk) 03:04, 7 May 2009 (UTC)[reply]

The best approximation is "Goo-row" (the final "d" should not be pronounced). —Preceding unsigned comment added by Shading83 (talkcontribs) 20:43, 22 October 2010 (UTC)[reply]

I've added IPA spelling and respelling. Coulomb1 (talk) 16:57, 11 February 2024 (UTC)[reply]

Something else

[edit]

Gouraud shading is independent of the shading model used at the vertexes. It need not be Phong. It could for example be Lambertian or something totally different, and Gouraud would work just fine, interpolating between the vertices.146.115.34.7 (talk) —Preceding undated comment was added at 11:58, 30 September 2008 (UTC).[reply]

Yep, this guy is right. Gouraud shading is simply the interpolation of colors across the face of a triangle based on the shaded colors computed at the vertices. The actual reflectance model used does not enter into it at all. Xot (talk) 09:44, 22 June 2009 (UTC)[reply]

The Seapot

[edit]

No reference to "Seapot" found on the Web. I believe this is a confusion with Alan Newell's Utah teapot which became a graphics icon in 1975. —Preceding unsigned comment added by Shading83 (talkcontribs) 20:49, 22 October 2010 (UTC)[reply]

reference to specific reflection model?

[edit]

Since 1) Gouraud shading was proposed before the Phong reflection model was proposed and 2) Gouraud shading can be applied to any reflection model (e.g. Blinn-Phong reflection model etc.), the description should be rephrased not to assume any specific reflection model. --Martin Kraus (talk) 09:51, 9 June 2011 (UTC)[reply]

bilinear interpolation?

[edit]

Why bilinear interpolation? Usually the colors are interpolated linearly in triangles. Polygons are often decomposed into triangles and then linear interpolation is used again. Thus, even for quadrilaterals it is more common to use a linear interpolation. --Martin Kraus (talk) 10:41, 9 June 2011 (UTC)[reply]

Reading the original publication, I realized that Gouraud was suggesting linear interpolation on edges of the polygon along the y axis and then another linear interpolation on each scanline. This corresponds to linear interpolation for triangles and bilinear interpolation for rectangles on the screen (and something rather strange for general quadrilaterals). I think in general Gouraud shading is nowadays assumed to make no suggestion about the specific kind of interpolation. --Martin Kraus (talk) 12:46, 22 June 2011 (UTC)[reply]

low-polygon surface?

[edit]

Why does the introduction mention a "low-polygon surface"? Gouraud shading actually works better with highly tessellated surfaces. Or does the "low" in "low-polygon" refer to the number of vertices? Why not "triangle surface" then? --Martin Kraus (talk) 10:43, 9 June 2011 (UTC)[reply]

Thinking about it, I realized that "low-polygon surface" refers to a surface with (on average) less than one vertex per pixel. Thus, I guess it makes sense, although the average reader will probably have no clue what this is about. --Martin Kraus (talk) 10:48, 9 June 2011 (UTC)[reply]

relation with Mach Bands

[edit]

"Gouraud shading will instead produce a highlight continuously fading in and out across neighboring portions of the model, peaking in intensity when the intended specular highlight passes over a vertex of the model." - I believe this phenomenon is known as "Mach Bands". Computer Graphics with OpenGL / Hearn&Baker, 3rd Edition, section 10-10, pg. 594-595: "Highlights on the surface are sometimes displayed with anomalous shapes, and the linear intensity interpolation can cause bright or dark intensity streaks, called Mach bands, to appear on the surface." — Preceding unsigned comment added by 85.65.164.238 (talk) 09:53, 20 July 2011 (UTC)[reply]

The intensity peak the article talks about does not occur because of a Mach band, which (according to the article about Mach bands) is an optical illusion that occurs because the human visual system performs spatial high-boost filtering on the luminance channel of the image captured by the retina. The article about Gouraud shading only describes what is going on in the computer and on the screen. What the book Computer Graphics with OpenGL talks about also occurs but it occurs along the edges and not on the vertices. Mach bands are not mentioned in the article; go ahead and add a note about them if you feel like it. —Kri (talk) 13:09, 20 July 2011 (UTC)[reply]

What is the common misconception referred to?

[edit]

The section here https://wiki.riteme.site/wiki/Gouraud_shading#Common_misconception explains how Gouraud shading is different from perspective correct shading. The title of the section suggests that there is some misconception about how Gouraud shading works or what it is. It is not clear what the misconception is, however. Perhaps is the misconception that people believe that Gouraud shading is perspective correct? If that is the case, it should be made more clear. — Preceding unsigned comment added by Prof dr finkelstein (talkcontribs) 18:25, 11 December 2018 (UTC)[reply]

The misconseption is that 100s of 3d engines claim they are using Gouraud shading. The original paper on Gouraud shading says Gouraud shading = linear interpolation between points but these engines are not using linear interpolation between points therefore they are not using Gouraud shading. The image linked shows the difference in results between using Gouraud shading and what most 3d engines render which is not Gouraud shading.

a comparison between Gouraud shading (linear interpolation) and non Gouraud shading (perspective correct interpolation)

— Preceding unsigned comment added by Greggman (talkcontribs) 02:03, 16 June 2019 (UTC)[reply]

Too much present tense

[edit]

The article uses the present tense a lot(‘is an interpolation method used in computer graphics’, ‘is most often used’) but hardware support for better shading models have made Gouraud shading obsolete for approximately twenty years now.

I agree. Nobody uses Gouraud shading anymore. I'll add a sentence in the introduction paragraph explaining this. Coulomb1 (talk) 17:03, 11 February 2024 (UTC)[reply]

misleading/duplicated seashell example?

[edit]

no prior knowledge on the topic, just passing by - looks like the gouraud-shaded seashell on this page is identical to the phong-shaded seashell on Phong shading. pretty sure in both cases it is actually flat shaded, going by the comparison directly underneath the seashell on this page. should these images look the same as flat shading for both of these pages? (maybe "graphics complex" is technical terminology where this diagram would be accurate?) Ralcore (talk) 12:12, 5 November 2022 (UTC)[reply]

References

[edit]

Should we add this: https://www.youtube.com/watch?v=jPqiUpYm0Ho ?

Henri Gouraud talks about the development of Gouraud Shading and others. 5.12.117.207 (talk) 20:58, 11 May 2024 (UTC)[reply]