# Posts TODO

A place to keep track of blog posts I’d like to do:

Chebyshev curve fitting / interpolation – with simple working c++ code. Possibly rational chebyshev too. Chebyshev apparently is optimal for polynomial interpolation.
https://www.embeddedrelated.com/showarticle/152.php

Optimistic concurrency (in databases). Select data and version id per row. Update versionid = versionid+1, blah where blah and version id = version id. If rows affected is 0, that means something else beat you to the punch and you can deal with it however.

Cordic math. Every iteration in a loop gives you another bit of precision, since it’s basically a binary search.

2D SDFs for vector graphics – using modulus for “free repeating”. anti aliasing. use your shadertoy verlet physics game as an example?

Verlet Physics Keep last and current position to get implicit velocity. Simulate. Iterative constraint solving. Things “just work” pretty well.

Minkowski Portal Refinement A nice & simple algorithm for collision detection. Maybe talk about algorithm to get depth. Mention JGK, possibly do that after.

Deterministic Simulations using deterministic sim to eg decrease network traffic.

Quick Math: phi / goden ratio show how golden ratio conjugate is the same number past the decimal point. show how this is the only number that can do that. The main point being “i remember this fact, but don’t remember the number”. This fact lets you calculate the number.

Quick math: eulers constant show how e^x being the derivative (and integral) can only work for e. The main point being “i remember this fact, but don’t remember the number”. This fact lets you calculate the number.

Ear clipping – for turning polygons into triangles. extendable to 3d with tetrahedron clipping, and to higher dimensions as well.

Storageless Shuffle With Weights – this is like if you have 3 red marbles and 5 blue marbles, how would you use FPE to storagelessly shuffle them.

Recurrent neural networks (etc) for “time series” learninghttps://twitter.com/Peter_shirley/status/1066832031043149824?s=03

Markov Chain Monte Carlo – Eg. for decryption maybe try 2nd order or higher chains.
Maybe also try with rendering / numerical integration http://statweb.stanford.edu/~cgates/PERSI/papers/MCMCRev.pdf

Blue Noise AO – It’s common to use white noise for sampling and also sample rotation. Start from there and show how to use blue for sampling and also rotation!
http://john-chapman-graphics.blogspot.com/2013/01/ssao-tutorial.html

Other blue noise usage cases – specific usage cases with easy to follow implementations
* fog shafts
* reflections
* dithering

Data Cache When doing coding experiments, there are often pieces of data that take time to calculate that are based on parameters that don’t often change from run to run. Making a data cache can help. Semi compelling usage cases: 1) next largest prime greater than N. 2) integrate a bilinear function. Compare / contrast to content addressable storage. CAS is the hash of the contents, this is the hash of the params that make the contents. code: https://github.com/Atrix256/ProgressiveProjectiveBlueNoise/blob/master/cache.h

Magic Eye Implementation turn a generic depth buffer into magic eye?
https://steemit.com/steemit/@mynameisbrian/how-to-create-a-steemit-themed-magic-eye-image-using-photoshop

Exposure and fstops
exposure is a multiplication. fstops are 2^(stop). feels linear when using fstops. must do multiplication in linear space, else is wrong (show it’s wrong).

Reaction Diffusion aka Turing patterns
https://en.wikipedia.org/wiki/Turing_pattern
https://www.quantamagazine.org/ancient-turing-pattern-builds-feathers-hair-and-now-shark-skin-20190102/

Kalman Filter
https://home.wlu.edu/~levys/kalman_tutorial/

# Audio Stuff

Biquad – a better frequency filter

Compressor & Limiter – automatic volume adjustment to eg avoid clipping. Include “side chain” stuff.

Statistical Search

Binary search works the way it does because it doesn’t know anything about the sorted list.

If you knew the min and the max in that sorted list, you could take a better guess at where to look first by finding the % that the value you are searching for is between min and max, and start at that % in the list.

The problem with that though is that it assumes an even distribution of numbers. If you have a bunch of small numbers and one huge number, this guess won’t be any good.

So, idea… if you fit the sorted list with some low order monotonic polynomial, you could reverse that to get an initial guess.

You could also update your best guess as to where the number was each time you looked in a place and got something that wasn’t your search value. This maybe using a kalman filter?

Faster 1d blue noise generation

1) brute force
2) binary search
3) linear search from initial guess from fit
4) kalman filter?

Use this stuff in generic data lists, not just in blue noise?

Maybe also a fixed # of buckets to cut search down. basically generate multiple in parallel and then append them together (but neighbors across edges do matter!)

james would…
I’d probably just put put uniform buckects
instead of a sorted array of numbers I’d try keeping the numbers ordered like a heep, then binary search is super faster, since it no longer suffers from non locality of memory access
an implicit binary tree where the childeren of a[i] are at a[i2+1] and a[i2 + 2]

# How To: Data Lookups Via Raytracing

Real time raytracing is seemingly finally here. We have a real time raytracing API in directX, another API in vulkan, hardware support in nvidia cards, and this is only the beginning of this new phase of graphics.

Having hardware and API support for raytracing means we can use that to help with the usual raytraced graphics things – reflection, refraction, shadows and more – but it is also new hardware / API to abuse.

For instance, Inigo Quilez talks about how ray / triangle intersection could be used to do 3×3 linear equation solving: http://www.iquilezles.org/blog/?p=4666

In a similar style of hardware/API abuse (but not related to raytracing), I have shown how to make the linear texture interpolator calculate points on curves and surfaces when storing the control points in texels, and also showed how it can evaluate generic polynomials: https://blog.demofox.org/2016/02/22/gpu-texture-sampler-bezier-curve-evaluation/

In the not too distant future, I think it will be common to use raytracing for data lookups in real time graphics, much like we use textures currently. In fact, from a couple conversations I’ve had with folks on twitter, it seems as though some people are already doing this.

As strange as it sounds, raytracing has a significant advantage over texture lookups. A texture lookup is limited to data stored in a regular grid in pixels. Raytracing data lookups get their data from a mesh, with data stored in vertices (in a vertex buffer).

Storing data in the vertices of a mesh means that you can store data points wherever you want. If you need more detail in one area and less in another, you can put more vertices in the higher detail area, and fewer vertices in the low detail area. You could also store data in a blue noise sampling pattern to help fight aliasing problems you might have with the regular grid of a texture. Furthermore, you could actually have holes in your data for invalid data regions, by having holes in the mesh.

Essentially, a mesh is just a generalization of a texture. You are no longer locked to a grid!

How the data lookups are actually done is not too complex either.

For the 2d case where you have a function f(x,y), you would make a triangle mesh where the (x,y) position of each vertex was the location of a data point, and you would make the z value some constant such as 0.5.

To look up the value of the data for some (x,y) input, you could make a ray that started at (x,y,0) and went in direction (0,0,1). When you did your raytrace, you’d get as a result the triangle index and the barycentric coordinates of that triangle. From there you could look up the data from the 3 vertices of the triangle and use the barycentric coordinates to interpolate between the values. The interpolated value would be your result.

You can see how this process goes in the image below. The purple dot is the query location. Just as there are volume textures to store 3d data in 3d textures, raytraced data lookups can also be extended to 3d.

For the 3d case where you have a function f(x,y,z), you would make a tetrahedral mesh where the (x,y,z) position of each vertex was the location of a data point.

To look up the value of the data for some (x,y,z) input, you need to be able to find what tetrahedron the point is in. To do this, you just shoot a ray starting at that (x,y,z) position in any arbitrary direction. That ray will hit one triangle in the tetrahedron. You then shoot a ray from the (x,y,z) position in the opposite direction to get a different triangle in the tetrahedron.

From these two triangles, you’ll have 6 vertices but only 4 will be unique. Those 4 vertices are the vertices of the tetrahedron. You can read the data from the vertices, calculate the barycentric coordinates of the point inside the tetrahedron, and then use those to interpolate the vertex data to get the result.

You can see how this process goes in the image below. The purple dot is the query location again. The most immediate usage case I can think of for this technique would be for diffuse light probe grids. Whether you had a 2d or 3d light probe grid, you’d be able to make probes as dense as needed, or as sparse as you can get away with, in different sections of the geometry. You could also make holes in the mesh to make sure data didn’t interpolate through walls, leading to light leaking. You would use the techniques described above to interpolate the simplex data and get the result.

As this data is likely going to be relatively simple geometry compared to something like an actual game asset, it seems like it ought to be able to be pretty performant too.

Nathan Reed shared a really good idea on twitter, for doing the 3d lookup with only a single raytrace. The idea is that when you knew what triangle you hit, you could look in a table to get the fourth vertex based on whether you hit the triangle from the front or the back. One way to do this would be to have a buffer that had two vertex indices per triangle. The first index would be if you hit from positive, the second would be if you hit from negative.

That way, the index you’d look at would be [triangleIndex*2 + hitBackSide ? 1 : 0]. That data lookup ought to be a lot cheaper than a second raytrace!