4. Protocol of the metrology testing for HRV programs
For metrology attestation 23 geared-up TRRS (typical RR-settings). Gain-frequency descriptions of these TRRS are presented in a table:

Notes: f is accordion frequency; P is accordion power.

At the choice of the set descriptions of TRRS we came from next terms. Clinically meaningful powers of HRV spectrum are 200-5000 ms2. The top and lower limit of this range is utillized for majority of TRRS. At experimental researches in critical clinic conditions can be registered extremely low powers of frequencies. It was used extremely low power (2 ms2) for the estimation of possibility of making researches with a testable device. Testing protocol included 23 TRRS. The first three TRRS with different power sets of HRV spectrum in the middle of standard frequency ranges are the most models for the estimation HRV-parameters of testable devices and non-linearity of frequency transformations. Selections N4 and N5 with the set powers on the edges of standard frequency ranges (accordingly on the right and on the left frequency ranges) are used for the estimation of exactness of frequency determination. The last eighteen TRRS (6-23) are monofrequency. They and allow to estimate the follow spectral parameters: exactness of of spectral frequency determination and their power in every subrange (VLF, LF, HF), non-linearity coefficient of AFC (amplitude frquency characteristic), level of side accordions (false spectral constituents).

Formalization of testing results was conducted according to the next protocol

1. The error coefficient of calculated spectrum HRV power was calculated by the formula:
K(f)=P(x)/P(0), where
P(x) is the calculated HRV spectrum power;
P(0) is the set power of HRV spectrum.
This coefficient was calculated as for general spectrum power, so separately for every spectral range in all TRRS. The value of this coefficient shows, what part of the set spectrum power is calculated by the testable program. An ideal value of this coefficient is equal to 1.

2. The level of AFC non-linearity. This value of software products was estimated as between ranges (for TRRS N 1-3) and inside the ranges (for TRRS N 6-18)
The non-linearity between ranges was calculated on a formula:
NL(%)=((Pmax-Pmin)/(P_G/3))*100, where:
Pmax and Pmin accordingly maximal and minimum value of spectrum power in three standard spectral ranges;
P_G is the general measured HRV spectrum power. The level of non-linearity was calculated at the equal set powers in three standard ranges. A value NL is shown by a betweenness minimum power-levels in these ranges and value of the middle measured power for all ranges.The level of non-linearity is calculated in ranges by the estimation of degree of distinctions between two calculated power values on different frequencies of one range by equal set powers. For example, for HF set power of TRRS=5000 nms2 index NL_HF_5000 is equal to NL_HF_5000(%)=((Pmax-Pmin)/5000)*100, where Pmax is the calculated as a spectrum power in area of HF for one TRRS, which exceeds an analogical value HF in second TRRS (Pmin) with equal set powers (5000 ms2). An ideal system of HRV AFC calculation must have these values about zero.

3. Noise level (for selections 6-23) which are conditioned appearance of side frequencies is  Nz(%)=(P(x) /P(0))*100, where: P(x) is the calculated power of HRV spectrum in one of two standard ranges where are not setted spectrum powers;
P(0) is the set HRV spectrum power for monofrequency TRRS. This index is calculated for monofrequency TRRS, separately for those ranges in which the set power is equal to the zero. The index of level of noises (Nz) is meant by a credible false power-level in relation to the set veritable power - then level in one of monofrequency TRRS ranges. General level of noises for each monofrequency TRRS is calculated as a sum of indexes of Nz in two ranges with set zero values of HRV spectrum power.

4. Exactness of frequency calculation of HRV spectrum. Testing on exactness determination of the set frequencies is executed by test sequences N4-N5 (see of tables.). Test N4 is  has on the left spectrum border power of 1000 ms2. In a test N5 the same power is set on the right frequency border.

5. The estimation of equivalence of dispersion and calculated general power was executed by a simple correlation of the calculated general HRV spectrum power and a value of undisplaced TRRS dispersion. This test was done with all other TRRS tests and was summarized as graphs. In the HRV ideal system this index must approach to 1.