Zitat: [FONT=&amp]Crossover Group Delay Audibility Testing[/FONT]
INFORMATION AS OF AUG 23, 2024

[FONT=&amp]About:[/FONT]
Group delay differences that are generated by loudspeaker crossover filtering cause some parts of the signal to emerge from the crossover at different times than others. Even when the frequency response is unchanged, non-uniform group delay causes the signal waveform in the time domain to be distorted. It has been established that when the group delay exceeds certain thresholds the changes to the time domain waveform become audible. The audibility threshold is lowest when listening via headphones to synthetic signals containing sharp transients. When listening to music-like signals via loudspeaker in a room, the audibility threshold is typically much higher.

[FONT=&amp]How to Take the Tests:[/FONT]
The following tests are organized by signal type. Each test consists of a reference signal and N adulterated signals with various levels of group delay added, presented in a random order. The frequency response and level of each adulterated signal is the same as the reference signal. The reference signal is always the original, unprocessed audio. All signals are stereo WAV audio files in various sample rates and bit depths, each lasting 5-10 seconds.
The subject listens to each adulterated signal and compares them to the reference signal. The adulterations include likely inaudible changes as well as likely audible ones. Only when a difference is audible should the corresponding checkbox be checked.
The changes between reference and adulterated tracks will be subtle. These are sometimes characterized as changes in tone or timbre. You will need to listen carefully. It is OK when you cannot hear a difference – that is the expected outcome for some adulterations and some test signals.
You may play the reference and adulterated tracks as many times as desired. Please be as honest as possible when selecting your answers. Once you are satisfied with your choices on the test form, click the button at the bottom of the form to submit your results. Submissions are anonymous. When a sufficient number of responses have been received the data will be analyzed and the results made available along with the details of each adulteration.
You can take a dummy/practice test in order to familiarize yourself with the form and format here: http://audio.claub.net/GD_testing/test_example/Example_Test.html
The results are only statistically significant when there are a large number of tests returned. You may take the test more than once, but please only do so after waiting a few hours or overnight.
Please invite others from the audio community to participate!
If you wish to contact me regarding these tests, please use virtually_me AT claub DOT net

[FONT=&amp]ONGOING TESTS:[/FONT]
SYNTHETIC SIGNALS:
TEST 1 : Signal = 100Hz square wave
TEST 2 : Signal = Rectangular Impulse
TEST 3 : Signal = Pink Impulse (from Liski and Välimäki)
MUSIC SIGNALS:
TEST 4 : Signal = Castanets
TEST 5 : Signal = Jazzy Piano Riff
TEST 6 : Signal = Latin Samba
TEST 7 : Signal = Xylophone

Gaga schrieb:[...] Ich konnte bei den synthetischen Signalen gut Unterschiede hören, [...]