This is a summary of the Audio Engineering Society Conference paper, http://www.aes.org/e-lib/browse.cfm?elib=12799
DVD-Audio versus SACD
Perceptual Discrimination of Digital Audio Coding Formats
Listening Comparison Test between DSD and High Resolution PCM (24-bit / 176.4 kHz)
by
Dominik Blech and Min-Chi Yang
Erich-Thienhaus-Institute (Tonmeisterinstitut), University of Music Detmold, Germany
http://www.hfm-detmold.de/hochschule/eti.html
It was published in 2004 when the format war between SACD and DVD-Audio were going on. It was an ABX test using the following set-up:
Simply put, they are talking analog audio and digitizing them with the dCS ADC/DAC into either 176 Khz PCM or 1x DSD:
"To avoid any influence of a mixer on the sound quality,
the stereo music examples were recorded with two
microphones and the surround examples with five. All
the microphones had extended frequency response to 40
or 50 kHz (Schoeps MK 2S, MK 4 and MK 41 capsules
with CMC 6-- xt amplifiers, and Sennheiser MKH 800);
one microphone was simply assigned to each playback
loudspeaker. The microphones were connected to
microphone preamplifiers (Lake People F/35 II) which
raised the signals to line level, then these signals were
sent to the control room via 50-meter low-capacitance
cables (Klotz M1 series). At that point the five analog
signals were split via “Y” adapters and converted to
digital, with one set of three two-channel dCS 904 units
used for DSD and another such set used for 176.4 kHz,
24-bit PCM. The resulting digital signals were then
stored on a “Pyramix Virtual Studio System” (Merging
Technologies) as “non-audio” files by using the “data
bitmapping” system of the converters to generate 24-bit,
44.1 kHz files (i.e. two channels of DSD were stored as
six channels on the workstation).
For playback, the audio signals were converted back to
analog again using dCS converters (a separate pair of
two-channel dCS 954 for the L, R, LS and RS of each
encoding system, plus separate two-channel dCS 955s
for each system’s center channel signal), and sent
through a high-quality stereo and surround monitor
control unit developed by the Emil Berliner Studios
(type MU 2000). The listener could switch between
DSD and PCM signals by using the ABX software (also
developed by the Emil Berliner Studios) to operate this
monitor control unit. A software-controlled delay is
introduced at the moment of switching between these
signals to prevent any accidental overlap. Loudspeakers
by Manger, distinguished by their very precise impulse
response and frequency response up to 35 kHz, were
used for playback. If the test subject opted for a stereo
listening sample, he or she could furthermore listen on a
pair of Stax headphones. All connections were carried
out exclusively with new, high-quality analog and
digital cables from Klotz."
ABX tests were then performed to see if the listeners could tell the difference.
"The listening tests took place over a period of 28 days.
During this span of time 145 tests could be carried out
with 110 test subjects. Some participants carried out the
test twice, either consecutively or on separate days,
using different music examples. ITU guidelines [3],
which state that conclusions may be drawn on the basis
of results from 20 persons or more, were clearly met."
The Tests were mostly in stereo but some were in surround.
Jumping to the punchline, four testers achieved better than 95% probability of chance that they could tell the difference:
"Each of the four test subjects had
chosen a different music example:
• Oboe: Stereo with headphones:
75% correct responses → p = 0.0207
• Speech: Stereo with headphones:
85% correct responses → p = 0.0013
• Guitar: Stereo with headphones:
90% correct responses → p = 0.0002
• Vocal: Stereo with headphones:
100% correct responses → p < 0.000001"
p = 0.05 represents 95% confidence that the listener was not guessing. Listeners all had a choice of listening with headphones or speakers. The above results were all achieved using headphones.
The rest of the tests/testers failed to provide this confidence:
"On the other hand, for 141 of the 145 test scores
(97.24%) hypothesis “H” cannot be rejected; in these
cases one could assume that a difference between
sources A and B was not perceived by the test subjects."
What this says is that even if there is a difference, vast majority of the testers cannot hear it.
Note: a flaw as found in the test in the form of a glitch. It is possible the above four people found that and voted based on that.
DVD-Audio versus SACD
Perceptual Discrimination of Digital Audio Coding Formats
Listening Comparison Test between DSD and High Resolution PCM (24-bit / 176.4 kHz)
by
Dominik Blech and Min-Chi Yang
Erich-Thienhaus-Institute (Tonmeisterinstitut), University of Music Detmold, Germany
http://www.hfm-detmold.de/hochschule/eti.html
It was published in 2004 when the format war between SACD and DVD-Audio were going on. It was an ABX test using the following set-up:
Simply put, they are talking analog audio and digitizing them with the dCS ADC/DAC into either 176 Khz PCM or 1x DSD:
"To avoid any influence of a mixer on the sound quality,
the stereo music examples were recorded with two
microphones and the surround examples with five. All
the microphones had extended frequency response to 40
or 50 kHz (Schoeps MK 2S, MK 4 and MK 41 capsules
with CMC 6-- xt amplifiers, and Sennheiser MKH 800);
one microphone was simply assigned to each playback
loudspeaker. The microphones were connected to
microphone preamplifiers (Lake People F/35 II) which
raised the signals to line level, then these signals were
sent to the control room via 50-meter low-capacitance
cables (Klotz M1 series). At that point the five analog
signals were split via “Y” adapters and converted to
digital, with one set of three two-channel dCS 904 units
used for DSD and another such set used for 176.4 kHz,
24-bit PCM. The resulting digital signals were then
stored on a “Pyramix Virtual Studio System” (Merging
Technologies) as “non-audio” files by using the “data
bitmapping” system of the converters to generate 24-bit,
44.1 kHz files (i.e. two channels of DSD were stored as
six channels on the workstation).
For playback, the audio signals were converted back to
analog again using dCS converters (a separate pair of
two-channel dCS 954 for the L, R, LS and RS of each
encoding system, plus separate two-channel dCS 955s
for each system’s center channel signal), and sent
through a high-quality stereo and surround monitor
control unit developed by the Emil Berliner Studios
(type MU 2000). The listener could switch between
DSD and PCM signals by using the ABX software (also
developed by the Emil Berliner Studios) to operate this
monitor control unit. A software-controlled delay is
introduced at the moment of switching between these
signals to prevent any accidental overlap. Loudspeakers
by Manger, distinguished by their very precise impulse
response and frequency response up to 35 kHz, were
used for playback. If the test subject opted for a stereo
listening sample, he or she could furthermore listen on a
pair of Stax headphones. All connections were carried
out exclusively with new, high-quality analog and
digital cables from Klotz."
ABX tests were then performed to see if the listeners could tell the difference.
"The listening tests took place over a period of 28 days.
During this span of time 145 tests could be carried out
with 110 test subjects. Some participants carried out the
test twice, either consecutively or on separate days,
using different music examples. ITU guidelines [3],
which state that conclusions may be drawn on the basis
of results from 20 persons or more, were clearly met."
The Tests were mostly in stereo but some were in surround.
Jumping to the punchline, four testers achieved better than 95% probability of chance that they could tell the difference:
"Each of the four test subjects had
chosen a different music example:
• Oboe: Stereo with headphones:
75% correct responses → p = 0.0207
• Speech: Stereo with headphones:
85% correct responses → p = 0.0013
• Guitar: Stereo with headphones:
90% correct responses → p = 0.0002
• Vocal: Stereo with headphones:
100% correct responses → p < 0.000001"
p = 0.05 represents 95% confidence that the listener was not guessing. Listeners all had a choice of listening with headphones or speakers. The above results were all achieved using headphones.
The rest of the tests/testers failed to provide this confidence:
"On the other hand, for 141 of the 145 test scores
(97.24%) hypothesis “H” cannot be rejected; in these
cases one could assume that a difference between
sources A and B was not perceived by the test subjects."
What this says is that even if there is a difference, vast majority of the testers cannot hear it.
Note: a flaw as found in the test in the form of a glitch. It is possible the above four people found that and voted based on that.
Last edited:
I think it's a bit outdated.
