Validating Loudness Alignment via Subjective Preference: Towards Improving ITU-R BS.1770-4

Berendes HU, Travaglini A, Uhle C. Validating Loudness Alignment via Subjective Preference: Towards Improving ITU-R BS.1770-4. In: AES Convention 153. Audio Engineering Society; 2022. Paper 27. Available from: https://aes.org/publications/elibrary-page/?id=21908

@inproceedings{Berendes2022_21908,
  author = {Berendes, Hans-Ulrich and Travaglini, Alessandro and Uhle, Christian},
  title = {{Validating Loudness Alignment via Subjective Preference: Towards Improving ITU-R BS.1770-4}},
  booktitle = {AES Convention 153},
  note = {Paper 27},
  year = {2022},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=21908}
}

TY  - CPAPER
TI  - Validating Loudness Alignment via Subjective Preference: Towards Improving ITU-R BS.1770-4
AU  - Berendes, Hans-Ulrich
AU  - Travaglini, Alessandro
AU  - Uhle, Christian
T2  - AES Convention 153
M1  - Paper 27
PY  - 2022
DA  - 2022/10/06
UR  - https://aes.org/publications/elibrary-page/?id=21908
PB  - Audio Engineering Society
LA  - en
AB  - Loudness measurement and control is important for many audio applications such as audio, television broadcasting and streaming. The recommendation ITU-R BS.1770-4 specifies means for measuring the loudness of audio signals. This paper presents experiments and results to investigate deviations between these measurements and loudness perception of human listeners. We conducted two loudness adjustment tests which quantify these deviations and one subsequent validation test to evaluate the effect of these deviations on the quality of loudness alignment. Adjustment gains obtained by ITU-R BS.1770-4 and human listeners differ on average by 2.2 dB. The listener agreement, however, is rather weak for some items, which brings up the question if the obtained data is useful for improving loudness measurements. The results of the validation test show an improvement of perceived loudness alignment quality by 16 % (of the used Mean Opinion Score scale) for gains obtained in the adjustment tests compared to ITU-R BS.1770-4. Since different listener groups participated in each test we conclude that the results of the adjustment tests are representative and can be used in future work to make loudness measurements more robust across a wider range of audio content and for object-based next-generation audio.
ER  -