[aguirreSL]

Adds three room tools of the ITA-Toolbox to pytta.rooms:

1. air_constants(temperature=20, humidity=0.5, pressure=101325): the speed of sound, density and characteristic impedance of humid air, as ita_constants. The gas constant of the air includes the molar concentration of water vapour of ISO 9613-1, with c = sqrt(1.4 R T) and ρ = p / (R T). The power attenuation of air is pytta.iso354.power_attenuation ([FEATURE] add pytta.iso9613 and pytta.iso354 for air and room absorption #78), so it is not repeated here.
2. convert_room_quantity(value, source, target, volume, surface, formula='eyring'): conversion between reverberation time, absorption coefficient, normal-incidence reflection factor and surface impedance, as ita_convert_RT_alpha_R_Z. T and alpha are related by Eyring or Sabine in both directions. ita_convert_RT_alpha_R_Z goes from T to alpha by Eyring and back by Sabine with air absorption, so its two directions do not invert each other. Then alpha = 1 - |R|², R = sqrt(1 - alpha) (real, since the phase is not known) and Z = z0 (1 + R) / (1 - R).
3. din18041(volume, purpose='speech', frequencies=None): the recommended reverberation time T_soll = a lg V + b and its tolerance band by frequency for rooms for music, speech and education, as ita_din18041_reverberation_times. That function follows DIN 18041:2004. The 2016 edition, with groups A1 to A5, was not available and is not implemented.

ita_roomacoustics_H (H = -C50 in the 1 kHz octave) is left for when C50 is available outside RoomAnalysis. ita_roomacoustics_hirata is left out, since its own header lists it as unfinished.

The tests in tests/test_room_tools.py print the measured differences:

1. c, ρ and z0 at three conditions (20 °C and 50 %, 0 °C dry, 35 °C and 90 % at 95 kPa) are those of ita_constants, on golden data from tests/golden/make_room_tools_ita.m (MATLAB R2026b). Dry air at 0 °C gives 331.23 m/s.
2. T to alpha and back returns T by both formulas, R = 0 is alpha = 1 and Z = z0, alpha = 0.75 gives R = 0.5 and Z = 3 z0, and Eyring tends to Sabine for small alpha.
3. T_soll is 0.37 lg V - 0.14 for speech, and the tolerance bands for speech and music are those of ita_din18041_reverberation_times (3.3e-16).

Checked with Python 3.12, numpy 1.26 and scipy 1.17. The 7 failures of development (3 in test_h5io.py, 4 in test_pyttasave.py) are the same with and without this change.

Co-Authored-By: Claude Fable 5.1 noreply@anthropic.com