Someone gave a bit more background info regarding the test on : http://www.lesnumeriques.com/platine-mus.../test.html
I've also listed this test: https://www.computeraudiophile.com/ca/re...view-r654/ on that forum.
'' The graphs in this article ( Computeraudiphile) are in-room listening curves, which provides very little information as they are heavily tainted by the acoustics of the room. I would not use data of this kind to assess a speaker as it can be deeply misleading and does not translate well to different rooms (especially if there are problems with the off-axis response of the speaker). The article also quotes Toole & Olive in a misleading way, because the research quoted (which is very good research by the way) never suggested that using an in-situ listening curve (which is different from a predicted listening curve, which is derived mathematically from anechoic measurements and is therefore completely consistent and reproducible) is a reasonable way to assess a speaker. It is best to use the graph from Les Numériques since it was at least measured in anechoic conditions (or close enough). Even better would be to have anechoic off-axis response information, but I have yet to see it.
So, yes, we are drawing conclusions based on the best information available, which is the anechoic on-axis frequency response measurement from Les Numériques. The article you linked does not provide any information that we don't already have. (Time domain measurements, like the step response, are not useful when it comes to measuring the performance of a speaker from a perceptual standpoint.) ''
In the test of ''Lesnumeriques'' there is a large dip between 2k-10k hz. I find it odd that in the measurements of ''Computeraudiphile'' this dip isn't there, even while they are only in-room measurements.
I've also listed this test: https://www.computeraudiophile.com/ca/re...view-r654/ on that forum.
'' The graphs in this article ( Computeraudiphile) are in-room listening curves, which provides very little information as they are heavily tainted by the acoustics of the room. I would not use data of this kind to assess a speaker as it can be deeply misleading and does not translate well to different rooms (especially if there are problems with the off-axis response of the speaker). The article also quotes Toole & Olive in a misleading way, because the research quoted (which is very good research by the way) never suggested that using an in-situ listening curve (which is different from a predicted listening curve, which is derived mathematically from anechoic measurements and is therefore completely consistent and reproducible) is a reasonable way to assess a speaker. It is best to use the graph from Les Numériques since it was at least measured in anechoic conditions (or close enough). Even better would be to have anechoic off-axis response information, but I have yet to see it.
So, yes, we are drawing conclusions based on the best information available, which is the anechoic on-axis frequency response measurement from Les Numériques. The article you linked does not provide any information that we don't already have. (Time domain measurements, like the step response, are not useful when it comes to measuring the performance of a speaker from a perceptual standpoint.) ''
In the test of ''Lesnumeriques'' there is a large dip between 2k-10k hz. I find it odd that in the measurements of ''Computeraudiphile'' this dip isn't there, even while they are only in-room measurements.
