An externally hosted image should be here but it was not working when we last tested it.
Finally got 2 drivers EQ's and level-matched. Works good so far, except for the plateamps that allows only 1db increment... Still, i was able to level-match 0.3db difference.
Also got a first guinea pig 
And i already can see few problems, starting by the 4-drivers-Cube.
It might be to much to test 4 drivers at the same time. Might have to move to good old ABX method.
Short term audio (auditory?) memory won't do very well with too much drivers to process at the same time, unless you practice for hours, which is not possible.
Maybe we were badlucky with 2 similar sounding drivers, but i'm afraid 4 drivers (as a starting level) will just be too much for most people.
ABX could work, but that means we need to do a lot to avoid ''lucky'' picks.
And i already can see few problems, starting by the 4-drivers-Cube.
It might be to much to test 4 drivers at the same time. Might have to move to good old ABX method.
Short term audio (auditory?) memory won't do very well with too much drivers to process at the same time, unless you practice for hours, which is not possible.
Maybe we were badlucky with 2 similar sounding drivers, but i'm afraid 4 drivers (as a starting level) will just be too much for most people.
ABX could work, but that means we need to do a lot to avoid ''lucky'' picks.
i'm totally brainstorming here, but maybe the good way to proceed would be to make a ''practice session'' with only 2 drivers (with good contrast between them, even if level-matched/EQd) in order to filter the ''good ears'' from the bad ?
I'm starting to see it coming... A lot of time and effort could be wasted for maybe only 10-20% of participants that will actually be able to identify the drivers...
Practice session of, lets say, 10min. If you don't start to show improvments after 10min, then the test stops for you, next participant, thank you have a nice day.
Make sense ?
I'm starting to see it coming... A lot of time and effort could be wasted for maybe only 10-20% of participants that will actually be able to identify the drivers...
Practice session of, lets say, 10min. If you don't start to show improvments after 10min, then the test stops for you, next participant, thank you have a nice day.
Make sense ?
FYI: i was able to cross as low as 330hz the ATC dome with very little need of EQing... and all the way up to 6.8khz.
Took me all 6 bands for that one, though.
Lucky me, i didnt have 1/200 oct measurement gears of Tomahawk! 😛
Took me all 6 bands for that one, though.
Lucky me, i didnt have 1/200 oct measurement gears of Tomahawk! 😛
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
FYI: i'm using Earthworks M50 mic for the EQ but the SPL calibrated Audiocontrol mic (images) for the level-matching.
Comfort listening level seems to be between 77db and 86db (peaks) from a distance of 1.7m.
Doesnt seem much but a single driver in mono making +/- 350hz-7khz can be quite irritating when it's too loud. 😱
Choice in Music
I believe the participants, trained and untrained, need to have a vested interest in this comparison to be willing to participate.
Have You thought about letting the participants bring their own music? I would say at least five songs, and allow them enough time to listen and compare.
It would be interesting to hear what the drivers sound like with different kinds of music, even the poorly recorded ones. 🙂
I believe the participants, trained and untrained, need to have a vested interest in this comparison to be willing to participate.
Have You thought about letting the participants bring their own music? I would say at least five songs, and allow them enough time to listen and compare.
It would be interesting to hear what the drivers sound like with different kinds of music, even the poorly recorded ones. 🙂
Good point.
In fact i thought about that already. I need to make a preselection because of the 400hz-6khz thing. Also excerpts needs to be short, and probably even shorter of what i did so far. Short-term memory is said to be 18 seconds...
Bottomline, i strongly believe it would be counter-productive to let people bring their own music. But that is sure possible ''outside the business hours'' just for the sake of trying it. In fact, i'll give it a try. 🙂
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Got our first participant who did a positive identification. No statistical validity yet, but at least some hope... 🙄
Real test runs are not started yet, need to make some tweak in the equipment and methodology, but its going very well. Very promising and exciting so far, i'm confident we'll have some answers! 🙂
I'm thinking about a way to ''filter'' participants... One idea i'm having is to use two identical drivers but with one of the following twist:
1. different xover points (say 1/3 oct wide on both hi/low)
2. difference of +/- 1db
3. difference in the EQ
The goal would be to discard the ''bad'' ears as soon as possible in the process..
Any comments or idea about that ?
1. different xover points (say 1/3 oct wide on both hi/low)
2. difference of +/- 1db
3. difference in the EQ
The goal would be to discard the ''bad'' ears as soon as possible in the process..
Any comments or idea about that ?
It seems clear we'll have to proceed with ABX test format. At least at first...
Since it won't be possible to test ALL drivers against all others, what about individual sport event format such as tennis ?
Either randomly chosed matchs, or we select it.
by example:
Match 1
ATC SM75-150 v.s. Alpair 7.3eNc
Match 2
Radian PB950 v.s. Scanspeak 10F
Then we could later match the ''winners'' between them and the ''losers'' between them.
Or, we could try a minimum of 2 or 3 matchups for each, but with a limited number of participants.
There will be some trade offs:
either many participants on a limited selection
or
few participants(test rounds) for more driver's matchups
Since it won't be possible to test ALL drivers against all others, what about individual sport event format such as tennis ?
Either randomly chosed matchs, or we select it.
by example:
Match 1
ATC SM75-150 v.s. Alpair 7.3eNc
Match 2
Radian PB950 v.s. Scanspeak 10F
Then we could later match the ''winners'' between them and the ''losers'' between them.
Or, we could try a minimum of 2 or 3 matchups for each, but with a limited number of participants.
There will be some trade offs:
either many participants on a limited selection
or
few participants(test rounds) for more driver's matchups
In the upcoming days, i'll test different musical excerpts lenghts as well.
I had the comments that our actual excerpts are too long (+/- 25-40sec)
according to Wikipedia, short-term memory is +/- 18 seconds, so i'll try excerpts between 10-15sec
I had the comments that our actual excerpts are too long (+/- 25-40sec)
according to Wikipedia, short-term memory is +/- 18 seconds, so i'll try excerpts between 10-15sec
So adding an Eton isn't going to happen? How about a 5" PRV, XRK had good things to say about them.
I think doing a tree is a good idea. But I'd welcome throwing one that lost back in to compare to say a polar opposite, maybe not for the win, but to help people understand the differences.
I think doing a tree is a good idea. But I'd welcome throwing one that lost back in to compare to say a polar opposite, maybe not for the win, but to help people understand the differences.
Last edited:
More drivers ain't the issue here, as i can see so far...
Second matchup tried today: same story.
Most people (if not all) ARE NOT ABLE to identify in a ABX test.
In the next few days i'll have to find some contrasty drivers OR some golden ears! 🙁😕
Second matchup tried today: same story.
Most people (if not all) ARE NOT ABLE to identify in a ABX test.
In the next few days i'll have to find some contrasty drivers OR some golden ears! 🙁😕
