Noponyheretoholdme
New Member
This is a collection of measurements of Diptyque DP 107 planar-magnetic/quasi-ribbon floorstanding loudspeakers. I was curious how they measure and was not able to find any reasonable measurements of the loudspeakers from this brand, so I tried to did them myself. All measurements are done in a room with dimensions of 5m x 3m and height of 2.45m using calibrated umik-1, not gated. From what is available on the Internet I'm 99% sure that exact woofer and tweeter are also being used on newer Diptyque DP 115 (and DP 140 mk1 but only woofer).
First off, individual woofer and tweeter measurements - taken at less than 1cm from the speaker, 1/6 smoothing, rms avg of L and R combined
Crossover frequency at around 1200Hz (1600Hz stated by specification) and it looks like first order crossover being used (6dB per octave), although 12dB per octave being specified by producer (that's very weird). Huge peak at around 45Hz is not that bad in my opinion, being open baffle speaker sub bass is not that strong anyway as you listen at some distance from the speaker. Tweeter have gradual 5db peak at 9kHz and honestly I like it, see my preferred eq at the end.
Woofer was not simple to measure. Single measurement produce big fluctuations (maybe due to push-pull design, as magnets covers around 50% of membrane from front and back), so I did 9 of them at different places of the woofer and then combine them.
Right speaker woofer measurements - taken at less than 1cm from the speaker, var smoothing, rms avg of all 9 measurements as white line without smoothing
Left speaker woofer measurements - taken the same way as on right speaker
Left and right woofer comparison - var smoothing, green - right, red - left
Left and right woofers are tuned differently, as right woofer resonance goes down to around 44Hz, while left fall off below 49Hz. There is an idea to tune resonance frequency of left and right speakers differently, so they will complement each other while human hearing should not be aware from which direction low frequency sound comes from. This technique is widely used in Magnepan loudspeakers. Hopefully here it was intended
About 1.5dB difference between left and right from 110Hz up.
Left and right tweeter comparison - no smoothing, single measurement per speaker, green - right, red - left
Uneven response of right tweeter, no matter how I measured it, peak at 9kHz was always present.
THD measurement at 86dB, 1 meter away measured from tweeter center, on tweeter axis, only left speaker being measured from now on
THD measurement at 96dB, 1 meter away measured from tweeter center, on tweeter axis
3rd harmonic stronger than 2nd - characteristic for push-pull design. I didn't wanted to push them further, as at 96dB they obviously started to suffer. As surface area of woofer is roughly equal to 21 inch conventional cone driver, I would consider those thd results as poor, although still good enough to enjoy music. Also want to point out, to achieve 86dB of SPL it required 4.2V rms output from amplifier, which given that those are 6 Ohm loudspeakers, would mean roughly 81dB of sensitivity @1W, 1 meter.
On axis response at different distances, 1/2 smoothing, fixed volume, green measurement taken 30 degrees offside, still on tweeter axis
Not sure if it translates to other room dimensions, but in my case, the further away from the speaker, the weaker the lower frequencies become. After long time trying different listening and loudspeaker positions, I end up sitting 75cm away from the tweeter. Further away requires too much eq bass boost for them to handle and even subwoofers won't help, as they would need to go as high as 300Hz, too high for subs to blend in nicely with main speakers.
I'm using those loudspeakers for almost 2 years now and from my listening experience they are unlistenable without eq. Even smallest portable bt speaker like JBL GO gives me more pleasure then them, however, with my eq, I love how they sound very much and don't plan to change them to anything.
Here is frequency response that I end up with - L&R loudspeaker avg with subs, psychoacoustic smoothing, at listening position with eq
Not sure why my target curve differs so much from standard ones like harman curve, but I'm confident that 4kHz dip is mandatory and 9kHz "peak" is not perceived as a peak (at least in this case with those loudspeakers in this room). It sounds correct to me and everyone I've let listen to it.
First off, individual woofer and tweeter measurements - taken at less than 1cm from the speaker, 1/6 smoothing, rms avg of L and R combined
Crossover frequency at around 1200Hz (1600Hz stated by specification) and it looks like first order crossover being used (6dB per octave), although 12dB per octave being specified by producer (that's very weird). Huge peak at around 45Hz is not that bad in my opinion, being open baffle speaker sub bass is not that strong anyway as you listen at some distance from the speaker. Tweeter have gradual 5db peak at 9kHz and honestly I like it, see my preferred eq at the end.
Woofer was not simple to measure. Single measurement produce big fluctuations (maybe due to push-pull design, as magnets covers around 50% of membrane from front and back), so I did 9 of them at different places of the woofer and then combine them.
Right speaker woofer measurements - taken at less than 1cm from the speaker, var smoothing, rms avg of all 9 measurements as white line without smoothing
Left speaker woofer measurements - taken the same way as on right speaker
Left and right woofer comparison - var smoothing, green - right, red - left
Left and right woofers are tuned differently, as right woofer resonance goes down to around 44Hz, while left fall off below 49Hz. There is an idea to tune resonance frequency of left and right speakers differently, so they will complement each other while human hearing should not be aware from which direction low frequency sound comes from. This technique is widely used in Magnepan loudspeakers. Hopefully here it was intended
About 1.5dB difference between left and right from 110Hz up.Left and right tweeter comparison - no smoothing, single measurement per speaker, green - right, red - left
Uneven response of right tweeter, no matter how I measured it, peak at 9kHz was always present.
THD measurement at 86dB, 1 meter away measured from tweeter center, on tweeter axis, only left speaker being measured from now on
THD measurement at 96dB, 1 meter away measured from tweeter center, on tweeter axis
3rd harmonic stronger than 2nd - characteristic for push-pull design. I didn't wanted to push them further, as at 96dB they obviously started to suffer. As surface area of woofer is roughly equal to 21 inch conventional cone driver, I would consider those thd results as poor, although still good enough to enjoy music. Also want to point out, to achieve 86dB of SPL it required 4.2V rms output from amplifier, which given that those are 6 Ohm loudspeakers, would mean roughly 81dB of sensitivity @1W, 1 meter.
On axis response at different distances, 1/2 smoothing, fixed volume, green measurement taken 30 degrees offside, still on tweeter axis
Not sure if it translates to other room dimensions, but in my case, the further away from the speaker, the weaker the lower frequencies become. After long time trying different listening and loudspeaker positions, I end up sitting 75cm away from the tweeter. Further away requires too much eq bass boost for them to handle and even subwoofers won't help, as they would need to go as high as 300Hz, too high for subs to blend in nicely with main speakers.
I'm using those loudspeakers for almost 2 years now and from my listening experience they are unlistenable without eq. Even smallest portable bt speaker like JBL GO gives me more pleasure then them, however, with my eq, I love how they sound very much and don't plan to change them to anything.
Here is frequency response that I end up with - L&R loudspeaker avg with subs, psychoacoustic smoothing, at listening position with eq
Not sure why my target curve differs so much from standard ones like harman curve, but I'm confident that 4kHz dip is mandatory and 9kHz "peak" is not perceived as a peak (at least in this case with those loudspeakers in this room). It sounds correct to me and everyone I've let listen to it.
Last edited:
