Re: Question regarding ISO 532 B and DIN 45 631

[Torben Poulsen <tp@xxxxxxxxxx>]

Dear Aron and list

As far as I know, the DIN standard has been revised in order to correct for some low frequency 'errors' (< 100Hz). Ask Hugo Fastl, Munich, about this. He was/is chairman of the DIN committee. The ISO standard has not been revised accordingly.

According to the definition, 1 sone is equal to 40 Phone (or dB SPL when we talk about 1000 Hz) and a 10 phone increase will double the sone value. This means that 70 dB SPL at 1000 Hz shall give 8 sone. The 7.4 sone result may reflect the loudness difference between a pure tone and a narrow band signal(?). Another explanation: The total loudness is proportional to the area under the curve and thus the extra steps will contribute marginally making 7.4 -> 8.0.

The three stair steps - at the low frequency side - must be based on measurements with actual filters and the stairs just reflects what's coming through the two neighbour filters (800Hz and 630Hz) - where 1000 Hz is way down at the skirts of the filters. In a calculation program (with ideal filters?) there should be no input to the lower filters.

If I specify 70 dB SPL at 1 kHz as input to (old) DOS programs I get the following (all free field):
DIN 45631 program: 6.94 sone, 68.0 phone
Calculation program for B&K 2231 Sound level meter: 6.95 sone, 68.0 phone
Loudness model by Brian Moore (not DIN/ISO): 8.95 sone, 71.4 phone

regards
Torben
--------------------------
Torben Poulsen
Department of Acoustic Technology
Technical University of Denmark
DK-2800 Lyngby, Denmark
Phone +45 4525 3940 (office), +45 4525 3930 (department)
Fax +45 4588 0577    mobile +45 2326 0420
e-mail: tp@dat.dtu.dk     Web: www.dat.dtu.dk/~tp/