News

Common language to describe sound

A new ITU-report describes how to develop a vocabulary for sound characteristics. This enables audio professionals to communicate efficiently when talking about sound perception in the context of audio reproduction.

By Christer P. Volk & Torben H. Pedersen

Speaking about audio…

The International Telecommunication Union (ITU) develops standards and recommendations for measuring and assessing equipment – both within telecommunication (ITU-T) and radiocommunication (ITU-R). Within the broadcasting (BS) division of ITU-R a new report, ITU-R BS.2399-0 [1], have recently been released, which aim to help audio professionals talk about sound reproduction with a common language use with objective and unambiguous terms. To communicate efficiently it is important to have a concise and well-defined language, such that everyone understands precisely what is meant. Until now, most audio professionals have used a colorful and often personal language to describe perception of an audio reproduction system. In magazines, the language resembles poetry and must describe the tested products in both subjective and objective terms, while manufactures searches for adjectives (often superlatives), which will impress consumers. Even within the scientific community, a common language is lacking, e.g. in perceptual modelling, which is rapidly expanding in these years. One example is an article where “Punch” is modelled without a definition of the term [2] and another that a (very similar) percept is described by “Apparent source width (ASW) in one place and “Sound image width” in another.

In short, for communications among audio professionals a common language for describing auditory perception makes a world of sense.

Audio vocabulary in the making

Back in 2005, Torben Holm Pedersen from DELTA SenseLab (now part of FORCE Technology) published a vocabulary of sound describing words [3]. Danish and English words with translations in from one to the other, this was the first step in finding out how to assemble a list of words. This work has been widely recognized and was for instance the catalyst for a similar vocabulary for evaluations of sound in an automotive context [4]. Within perceptual audio evaluation, the set of characterizing words are referred to as attributes. In 2006 Nick Zacharov (SenseLab) and Søren Bech (Bang & Olufsen) included a list of attribute requirements in their popular book Perceptual Audio Evaluation [5]. It is the continuation of their work, which led to a Sound wheel of well-defined attributes for evaluation of reproduced audio described in a more recent paper from 2015 [6]. The paper includes a detailed description of how attributes in the sound wheel where established and validated and it is an updated version of this process which has now made it into ITU-R BS.2399-0, along with the sound wheel (now “Audio wheel”) as an example.

Figure 1. Sectional view of the Audio wheel developed by DELTA SenseLab. Attributes are in the outer ring, categories in the middle ring and main groups in the inner ring. Basic Audio Quality describes the objective overall impression.

Impact and perspective

The International Telecommunication Union dates back to 1865, celebrating its 150th anniversary in 2015. With its current 193 member countries, it has a global reach and is an ideal platform for spreading the knowledge from the few to the masses. With the publication of this report, the vocabulary used by audio professionals can be gradually streamlined, not only by using a common set of words, but just as importantly, by having a common understand of the meaning of those words, as each one comes with a definition.

Sensory product information for audio products

One very practical utilization, which we in SenseLab hope this will bring, is an increase in the use of sensory product information on audio products as known from the food industry, e.g. the beer label shown in Figure 2. Also, an example of perceptual characteristics for a loudspeaker is shown in figure 2.

Figure 2. examples of sensory product scales. Left, the label of a Carlsberg beer: Sweetness, bitterness and colour. Right, perceptual characteristics of a loudspeaker.

The published audio wheel is not static, and the development continues. Zacharov et al. has published an updated version with more spatial attributes [7]. Lindau et al. has published an alternative set of spatial attributes (SAQI) [8] and others will follow in the future. Thus, the ITU report may serve as a support document for ensuring that the continued development leads to high quality contributions and further allows a wider segment of audio professionals to contribute. The Audio wheel is ready for a worldwide debate, but is even now an amazing tool for easing communication of the characteristics of reproduced speech and music.

More info

For more information about perceptual audio evaluation contact Søren V. Legarth, Team Manager, SenseLab, FORCE Technology, svl@force.dk, +45 72 19 46 10 or Torben Holm Pedersen, Senior Specialist, SenseLab, FORCE Technology, thp@force.dk, +45 72 19 46 17.

References

[1] ITU-R. ‘Methods for Selecting and Describing Attributes and Terms, in the Preparation of Subjective Tests’. Report. Geneva, Switzerland: International Telecommunication Union Radiocommunication Assembly (ITU-R), March 2017.

[2] Fenton, Steven, and Hyunkook Lee. ‘Towards a Perceptual Model of “Punch” in Musical Signals’. In Proc. of the Audio Engineering Society Convention 139, 1–10. New York, NY, USA: Audio Engineering Society, 2015.

[3] Pedersen, Torben H. ‘The Semantic Space of Sounds – Lexicon of Sound-Describing Words’. Lexicon. Hørsholm, Denmark: DELTA, May 2008.

[4] Altinsoy, M. Ercan; Jekosch, Ute. ‘The Semantic Space of Vehicle Sounds: Developing a Semantic Differential with Regard to Customer Perception’. J. Audio Eng. Soc 60, no. 1/2 (2012): 13–20.

[5] Bech, Søren, and Nick Zacharov. Perceptual Audio Evaluation: Theory, Method and Application. Chichester, England; Hoboken, NJ: John Wiley & Sons, 2006.

[6] Pedersen, Torben H., and Nick Zacharov. ‘The Development of a Sound Wheel for Reproduced Sound’. In Audio Engineering Society Convention 138, 1–13. Warsaw, Poland: Audio Engineering Society, 2015.

[7] Zacharov, Nick, Torben Pedersen, and Chris Pike. ‘A Common Lexicon for Spatial Sound Quality Assessment – Latest Developments’. In 2016 Eighth International Conference on Quality of Multimedia Experience (QoMEX), 1–6. Lisbon, Portugal: IEEE, 2016. doi:10.1109/QoMEX.2016.7498967.

[8] Lindau, Alexander, Vera Erbes, Steffen Lepa, Hans-Joachim Maempel, Fabian Brinkman, and Stefan Weinzierl. ‘A Spatial Audio Quality Inventory (SAQI)’. Acta Acustica United with Acustica 100, no. 5 (2014): 984–94. doi:10.3813/AAA.918778.

Read more


DANSK VERSION


Et fælles sprog til at beskrive lyd

En ny rapport fra ITU giver forslag til, hvordan man kan udvikle fælles betegnelser for lyde og deres egenskaber. Lydeksperter får dermed et begrebsapparat til mere effektivt at drøfte lydopfattelse i forbindelse med lydgengivelse.

Af Christer P. Volk og Torben H. Pedersen

Når man taler om lyd…

International Telecommunication Union (ITU) udvikler standarder og anbefalinger til måling og vurdering af udstyr – både inden for telekommunikation (ITU-T) og radiokommunikation (ITU-R). Radiodivisionen (BS) under ITU-R har for nylig udsendt en ny rapport, ITU-R BS.2399-0 [1], der har til formål at skabe et fælles sprog med objektive og entydige termer, som audioeksperter kan benytte, når de tale om lydgengivelse. Effektiv kommunikation kræver et præcist og veldefineret fælles sprog, så alle deltagere i kommunikationen har det samme udgangspunkt for at forstå hinanden præcist.

Hidtil har audioeksperter benyttet et farverigt og ofte meget personligt sprog til at beskrive deres opfattelse af et lydgengivelsessystem. I magasiner anvendes et nærmest poetisk sprog, der beskriver de testede produkter i både subjektive og objektive termer, og producenterne har tendens til at benytte tillægsord (ofte superlativer), der kan gøre indtryk på forbrugerne. Selv i videnskabelige kredse mangler man et fælles sprog, fx inden for modellering af lydopfattelse, der vinder hastigt frem i øjeblikket. Som eksempel kan vi nævne en artikel, hvor ordet ’punch’ benyttes i en model, uden at termen defineres nærmere [2]. I en anden artikel, hvor en to meget ens faktorer i lydopfattelsen beskrives med ’apparent source width’ (ASW) ét sted og ’sound image width’ (dansk: ’bredde af lydbilledet’) et andet sted.

Kort sagt giver det rigtig god mening at fastlægge et fælles sprog til beskrivelse af lydopfattelse til brug for kommunikation mellem audioeksperter.

Begrebsapparat under udarbejdelse til audioeksperterne

Allerede i 2005 udgav Torben Holm Pedersen fra DELTA SenseLab (nu er en del af FORCE Technology) en ordliste for lydgengivelse [3]. En liste over danske og engelske ord med oversættelse mellem de to sprog var første skridt til at udvikle et nyt og omfattende begrebsapparat. Hans arbejde blev bredt anerkendt og dannede for eksempel udgangspunkt for udarbejdelse af en lignende ordliste for vurdering af lyde i forbindelse med arbejde med biler [4].

Inden for perceptuel vurdering af lyd kaldes de kendetegnende ord ’attributter’. I 2006 indarbejdede Nick Zacharov (SenseLab) og Søren Bech (Bang & Olufsen) en liste over krav til attributter i deres populære bog Perceptual Audio Evaluation [5].

Som en videreførelse af deres arbejde er lydhjulet blevet udviklet med et sæt veldefinerede attributter til vurdering af lydgengivelse, der beskrives i en nyere artikel fra 2015 [6]. Artiklen indeholder en detaljeret beskrivelse af, hvordan lydhjulets attributter er blevet fastlagt og valideret, og det er netop en opdateret udgave af denne proces, der nu indgår i ITU-R BS.2399-0 med lydhjulet (nu kaldet “Audio wheel”) som eksempel.

Figur 1. Udsnit af lydhjulet udviklet af DELTA SenseLab. Attributterne står i den yderste ring, kategorierne står i den midterste ring, og hovedgrupperne står i den inderst ring. Basal lydkvalitet beskriver det generelle, objektive indtryk.

Perspektiver og konsekvenser

International Telecommunication Union (ITU) blev stiftet i 1865 og kunne altså fejre 150 års jubilæum i 2015. ITU har i øjeblikket 193 medlemslande, og dermed favner sammenslutningen globalt og er den ideelle platform til at udbrede ny viden effektivt. Med udgivelsen af rapporten kan lydeksperternes begrebsapparat gradvist ensrettet, ikke kun ved at benytte en række fælles termer, men lige så vigtigt ved at have en fælles forståelse af termernes betydning, da de enkelte termer nu er præcist defineret.

Sensorisk prodiktinformation for audioprodukter

Hos SenseLab håber vi, at en af de praktiske anvendelser kan blive en øget brug af sensorisk produktinformation om audioprodukter, som vi kender det fra fødevareindustrien, fx øletiketten vist til venstre i figur 2. Til højre i figur 2 ses et tilsvarende eksempel på sensoriske parametre for højttalere.

Figur 2. sensoriske produktattributter. Til venstre, etiketten til en Carlsberg øl: sødhed, bitterhed og farve. Til højre, sensorisk produktinformation for højttalere.

Det publicerede lydhjul skal ikke ses som en statisk størrelse, og det skal løbende videreudvikles. Zacharov et al. har udgivet en opdateret udgave med flere rumlige attributter [7]. Lindau et al. har udgivet et alternativt sæt rumlige attributter (SAQI) [8], og yderligere vil komme til fremover. På den måde skal ITU-rapporten ses som en vejledning, der skal sikre, at der fremover kommer bidrag af høj kvalitet, og at en bredere kreds af audioeksperter får mulighed for at bidrage. Lydhjulet kan indgå i en global debat, og samtidig kan det allerede nu bruges som et effektivt værktøj til at støtte kommunikationen om egenskaber ved gengivelse af tale og musik.

Mere info

For yderligere oplysninger om perceptuel vurdering af lyd kontakt Søren Vase Legarth, Team Manager, SenseLab, FORCE Technology, svl@force.dk, +45 72 19 46 10 eller Torben Holm Pedersen, Senior Specialist, SenseLab, FORCE Technology, thp@force.dk, +45 72 19 46 17.

Referencer

[1] ITU-R. ‘Methods for Selecting and Describing Attributes and Terms, in the Preparation of Subjective Tests’. Report. Geneva, Switzerland: International Telecommunication Union Radiocommunication Assembly (ITU-R), March 2017.

[2] Fenton, Steven, and Hyunkook Lee. ‘Towards a Perceptual Model of “Punch” in Musical Signals’. In Proc. of the Audio Engineering Society Convention 139, 1–10. New York, NY, USA: Audio Engineering Society, 2015.

[3] Pedersen, Torben H. ‘The Semantic Space of Sounds – Lexicon of Sound-Describing Words’. Lexicon. Hørsholm, Denmark: DELTA, May 2008.

[4] Altinsoy, M. Ercan; Jekosch, Ute. ‘The Semantic Space of Vehicle Sounds: Developing a Semantic Differential with Regard to Customer Perception’. J. Audio Eng. Soc 60, no. 1/2 (2012): 13–20.

[5] Bech, Søren, and Nick Zacharov. Perceptual Audio Evaluation: Theory, Method and Application. Chichester, England; Hoboken, NJ: John Wiley & Sons, 2006.

[6] Pedersen, Torben H., and Nick Zacharov. ‘The Development of a Sound Wheel for Reproduced Sound’. In Audio Engineering Society Convention 138, 1–13. Warsaw, Poland: Audio Engineering Society, 2015.

[7] Zacharov, Nick, Torben Pedersen, and Chris Pike. ‘A Common Lexicon for Spatial Sound Quality Assessment – Latest Developments’. In 2016 Eighth International Conference on Quality of Multimedia Experience (QoMEX), 1–6. Lisbon, Portugal: IEEE, 2016. doi:10.1109/QoMEX.2016.7498967.

[8] Lindau, Alexander, Vera Erbes, Steffen Lepa, Hans-Joachim Maempel, Fabian Brinkman, and Stefan Weinzierl. ‘A Spatial Audio Quality Inventory (SAQI)’. Acta Acustica United with Acustica 100, no. 5 (2014): 984–94. doi:10.3813/AAA.918778.

Læs mere

Tags:

This is a unique website which will require a more modern browser to work! Please upgrade today!