DR error measures

This is a technical or mathematical page. While the subject may be of some relevance to music, the page treats the subject in technical language.

Least-squares linear error (here linear means "in frequency space, not pitch space") is a proposed error measure for approximations to delta-rational chords. It has the advantage of not fixing a particular interval in the chord when constructing the chord of best fit. However, like any other numerical measure of concordance or error, you should take it with a grain of salt.

Fully DR

The idea motivating least-squares linear error on a chord as an approximation to a given delta signature is the following (for simplicity, let’s talk about the fully DR case first):

Say we want the error of a chord 1:r₁:r₂:...:r_n (in increasing order), with n > 1, in the linear domain as an approximation to a fully delta-rational chord with signature +δ₁ +δ₂ ... +δ_n, i.e. a chord

$x:x+{\delta }_1:\cdots :x+\sum _{l=1}^n{\delta }_l.$

Minimizing the least-squares frequency-domain error by varying x gives you the closed-form solution

$x=\frac{\sum _{i=1}^n{D}_i}{-n+\sum _{i=1}^n{f}_i},$

which you plug back into

$\sqrt{\sum _{1=1}^n(1+\frac{D_i}{x}-f_i{)}^2}$

to obtain the least-squares linear error.

Partially DR (one related delta set, one free variable)

Suppose we wish to approximate a target delta signature of the form $+{\delta }_1+?+{\delta }_3$ with the chord $1:f_1:f_2:f_3$ (where the +? is free to vary). By a derivation similar to the above, the least-squares problem is

$\underset{x,y}{\text{minimize}}\sqrt{(\frac{x+{\delta }_1}{x}-f_1{)}^2+(\frac{x+{\delta }_1+y}{x}-f_2{)}^2+(\frac{x+{\delta }_1+y+{\delta }_3}{x}-f_3{)}^2},$

where y represents the free delta +?.

We can set the partial derivatives with respect to x and y of the inner expression equal to zero (since the derivative of sqrt() is never 0) and use SymPy to solve the system:

<syntaxhighlight lang="py"> import sympy x = sympy.Symbol("x", real=True) y = sympy.Symbol("y", real=True) d1 = sympy.Symbol("\\delta_{1}", real=True) d2 = sympy.Symbol("\\delta_{2}", real=True) d3 = sympy.Symbol("\\delta_{3}", real=True) f1 = sympy.Symbol("f_1", real=True) f2 = sympy.Symbol("f_2", real=True) f3 = sympy.Symbol("f_3", real=True) err_squared = ((x + d1) / x - f1) ** 2 + ((x + d1 + y) / x - f2) ** 2 + ((x + d1 + y + d3) / x - f3) ** 2 err_squared.expand() err_squared_x = sympy.diff(err_squared, x) err_squared_y = sympy.diff(err_squared, y) sympy.nonlinsolve([err_squared_x, err_squared_y], [x, y]) </syntaxhighlight>

The unique solution with x > 0 is $(x,y)=(\frac{2{\delta }_1+{\delta }_3+\frac{2(-2{\delta }_1^2{f}_1+{\delta }_1^2{f}_2+{\delta }_1^2{f}_3-{\delta }_1{\delta }_3{f}_1+{\delta }_1{\delta }_3{f}_2-{\delta }_1{\delta }_3{f}_3+{\delta }_1{\delta }_3+{\delta }_3^2{f}_2-{\delta }_3^2)}{2{\delta }_1{f}_1-2{\delta }_1-{\delta }_3{f}_2+{\delta }_3{f}_3}}{f_2+f_3-2},\frac{-2{\delta }_1^2{f}_1+{\delta }_1^2{f}_2+{\delta }_1^2{f}_3-{\delta }_1{\delta }_3{f}_1+{\delta }_1{\delta }_3{f}_2-{\delta }_1{\delta }_3{f}_3+{\delta }_1{\delta }_3+{\delta }_3^2{f}_2-{\delta }_3^2}{2{\delta }_1{f}_1-2{\delta }_1-{\delta }_3{f}_2+{\delta }_3{f}_3}).$

Partially DR (arbitrary)

We similarly include a free variable to be optimized for every additional +?, after coalescing strings of consecutive +?'s and omitting the middle notes, and after trimming leading and trailing +?'s. If two variables are related to each other but not to the integer deltas in the signatures, they have a common variable.

Todo: The L-BFGS-B algorithm is suited for five-variable (base real-valued harmonic + four free deltas; a realistic upper bound on real-world use cases of partial DR) optimization problems with bounds, so let's talk about that

External links

• Inthar's DR chord explorer (includes least-squares linear error calculation)