Vous êtes sur la page 1sur 26

Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.

de Page 1
Advanced Psychoacoustics
Lecture 4
SS 2011
Prof. Dr.-Ing. Karlheinz Brandenburg
karlheinz.brandenburg@tu-ilmenau.de
Prof. Dr.-Ing. Thomas Sporer
spo@idmt.fraunhofer.de
contact: Stephan Werner
stephan.werner@tu-ilmenau.de
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 2
Content
quality:
perceived vs. produced
methods for quality measurement
sensory profiling
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 3
produced vs. perceived
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 4
produced not perceived quality [1]
Qualitt
coding
bitrate
produced
transmission
resolution
expectations
emotions
perceived
cognitive
processes
motivation
VS.
Qualitt
[1] Jumisko-Pyykk, S. et al.: Produced quality is not perceived quality a qualitative approach to overall audiovisual quality, In: 3DTV Conference 2007
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 5
methods for measurement:
traditional ITU standards
quantitative test methods
rating of quality with scales and numbers
example: Absolute Category Rating
reliable and valid results
conclusions about preferences
no information why quality was perceived in such way
Qualitt
Quality
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 6
methods for measurement:
qualitative test methods
rating of quality with non standard data (e.g. words)
methods from empirical social research
observations, qualitative interviews, discussions,
sensory profiling
consensus and individual vocabulary profiling
Qualitt
Quality
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 7
standards
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 8
Standards - Overview
Audio-Visuelle Standards:
e.g.:
ITU-R BS.775
ITU-R BS.1286
Audio-Rating:
ITU-R BS.1283 universal overview
Bildquelle: ITU-R BS.1283
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 9
Standards - ITU-R BS.1116
rating of small differences,
method:
triple stimulus with hidden reference
R-A-B (R=reference, A and B reference or signal in random
order)
free switching between R, A, B possible,
Question: Is A or B different to R ? (two step rating)
rating on impairment scale
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 10
Standards - ITU-R BS.1116
Impairment Scale:
Impairment Grade
Imperceptible 5.0
Perceptible, but not annoying 4.0
Slightly annoying 3.0
Annoying 2.0
Very annoying 1.0
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 11
Standards - ITU-R BS.1116
subjects:
experts
small group provides significant results,
untrained listeners
a lot of subject are necessary,
training of subjects,
pre screening,
post screening,
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 12
Standards - ITU-R BS.1116
standard says something about:
choice of subjects,
arrangement of loudspeakers,
size and acoustic character of the listening room.
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 13
Standards - Hrlabor
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 14
Standards - Hrlabor
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 15
Standards - ITU-R BS.1534 (MUSHRA)
rating of medium differences,
method:
multi stimulus with hidden reference
rating of various different conditions in parallel order,
hidden reference (should be the best),
anchor with TP = 3,5 kHz (should be the worst),
more anchors are optional,
multistage rating
Which condition is the hidden reference?
Which condition is the anchor?
rating of the other conditions on the impairment scale
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 16
Standards - ITU-R BS.1534 (MUSHRA)
no more than 10 conditions,
anchor for post screening,
use the complete rating range
hid. ref. = 100, anchor = 0
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 17
Standards - ITU-R BS.1534 (MUSHRA)
MUSHRA im Selbstversuch:
Bewertung von Kodierartefakten,
Die einzelnen Konditionen sollen gegenber der offenen
Referenz in Hinsicht auf ihre Qualitt verglichen werden.
Sie knnen sich so oft wie ntig die einzelnen Konditionen
anhren und bewerten.
Nutzen Sie den vollen Bewertungsbereich aus!
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 18
qualitative interviews
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 19
Qualitative Interviews
Structured, semi-structured or unstructured open questions
Using an interview guideline
Training of interviewer necessary
Transcription of interviews
Coding of interviews
Interpretation of results
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 20
Qualitative Interviews
Example from audiovisual Quality evaluation:
Evaluate the reasons for participants quantitative ratings
Main Question: What were the factors you paid attention to while
evaluating the audiovisual quality?
Supporting Questions: What do you mean with X (answer of main
question)? Could you please describe in more detail what you mean
with X?
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 21
sensory profiling
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 22
Sensory Profiling
scientific research methodology used to evoke, measure, analyze and
interpret reactions to those characteristics of food and materials as they
are perceived by senses of light, smell, taste, touch and hearing [2]
Sensory description of products
Consensus and individual sensory profiling
[2] Stone, H. and Sidel, J. L., Sensory evaluation practices, 3
rd
ed., Academic Press, San Diego, 2004
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 23
Individual Vocabulary Profiling
Participants generate own, individual vocabulary to describe quality
Audiovisual quality evaluation:
Development of Open Profiling of Quality method (OPQ) [3]
4 steps:
Introduction
Attribute elicitation
Attribute refinement
Sensorial evaluation
[3] D. Strohmeier, S. Jumisko-Pyykk, and K. Kunze, Open profiling of quality: a mixed method
approach to understanding multimodal quality perception, Advances in Multimedia, vol. 2010
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 24
Open Profiling of Quality
Example: audiovisual 3D quality
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 25
Open Profiling of Quality
Example: audiovisual 3D quality
Prof. Dr.-Ing. K. Brandenburg, karlheinz.brandenburg@tu-ilmenau.de spo@idmt.fraunhofer.de Page 26
next time:
presentation of the ideas for our practical part
first implementations

Vous aimerez peut-être aussi