I did listen to a pair of Magnepan 1.7's yesterday with some very expensive electronics (Audio Research Reference Series), in a large room. The speakers were placed 1.5 meter from the front wall, and there was some diffusion and damping on the front wall. The room was a bit more "live" than I personally prefer though.
They sound nice, no doubt! Very open sound and freedom from grain, harshness and coloration. If I wanted to buy a set of 2000$ speakers, it would be Magnepan 1.7.
That being said, they are not perfect. I found that the soundstage lacks depth and precision, and that the top end is a bit closed-in and maybe lacks a little in level. The midrange is where this speaker really shines, although there is a bit too much of it.... I think that the speaker would sound much better if the area around 1 kHz was dropped maybe 2-3 dB in level.
Its a three-way and using low-slope crossovers, and the tweeter sits on the side of the midrange. This makes the horizontal listening angle very critical. Its absolutely necessary to have the right toe-in, or else the tonal balance gets really bad.
Bass? Dont know, there was a subwoofer in action here....
Executive summary:
- Magnepan 1.7 very good, with some limitations
- RD-75 very good, fewer limitations
Thanks for sharing your impressions.
I wonder if you could be a bit more precise about the lack of precision.
I see three different areas. Was it:- uneven distribution of phantom images along the stereo base (clustering at the speakers or the center)?
- phantom images well distributed and defined, but each too large in size?
- general fuzziness of location and size of phantom sources?
I assume there was no digital correction for the Magnepans!?
Thanks for sharing your impressions.
I wonder if you could be a bit more precise about the lack of precision.
- uneven distribution of phantom images along the stereo base (clustering at the speakers or the center)?
- phantom images well distributed and defined, but each too large in size?
- general fuzziness of location and size of phantom sources?
I assume there was no digital correction for the Magnepans!?
Imaging: Your last two points, and with lack of depth in addition.
No correction. I think using that could have improved the situation a lot!
Thanks again, Stig Erik.
This is somehow what I expected. I would like to expand on it and hope to be not too much OT:
Lately our small group got an introduction to FIR filtering by the developer of acourate. Objects of demonstration were first a pair of really cheapo "NXT" like panels - just a 60x80 cm piece of styrofoam board with an exciter applied. Second a pair of two-way coax TML boxes.
FIR filtering could excellently linearise the poor frequency response of the panels, but it did not much to the fuzziness of the phantom sources. They stayed to be like the sun behind clouds: You knew where it was bright, but you didn't see the disc.
The coax had a much better behaved FR from the start with decent localisation and definition of the phantom sources. Applying the FIR filters was more like removing a halo from the sun: Better contrast against the background and less "noise" around each source. In fact I would not have expected that the definition of phantom sources can be made THAT sharp.
Coming back to topic:
If planars are still like what I heard some years ago (mouthes much wider than even Julia Roberts', pianos reaching across the complete front wall), then digital filters might not help very much to improve that IMHO.
If it is about taking that famous "last veil" from the scene, then digital filtering can do a lot - regardless of the speaker construction principle or the size of the baffle.
Rudolf
This is somehow what I expected. I would like to expand on it and hope to be not too much OT:
Lately our small group got an introduction to FIR filtering by the developer of acourate. Objects of demonstration were first a pair of really cheapo "NXT" like panels - just a 60x80 cm piece of styrofoam board with an exciter applied. Second a pair of two-way coax TML boxes.
FIR filtering could excellently linearise the poor frequency response of the panels, but it did not much to the fuzziness of the phantom sources. They stayed to be like the sun behind clouds: You knew where it was bright, but you didn't see the disc.
The coax had a much better behaved FR from the start with decent localisation and definition of the phantom sources. Applying the FIR filters was more like removing a halo from the sun: Better contrast against the background and less "noise" around each source. In fact I would not have expected that the definition of phantom sources can be made THAT sharp.
Coming back to topic:
If planars are still like what I heard some years ago (mouthes much wider than even Julia Roberts'
, pianos reaching across the complete front wall), then digital filters might not help very much to improve that IMHO.If it is about taking that famous "last veil" from the scene, then digital filtering can do a lot - regardless of the speaker construction principle or the size of the baffle.
Rudolf
My guess is that the fuzziness has more to do with non-linear distortion than with the frequency and phase response. Digital correction filters can do wonders with frequency and phase deviations (linear distortion), but very little to remove non-linear distortion (additional frequency components that were not in the original signal).
I'd like to add that the room where I listened to the Magnepans was very live, and that I would expect the speakers to sound better in a more appropriate listening room. The Magnepans did not present a very large image with instruments taking on giant proportions. Its just not anyway as sharp or vivid as my own RD-75 based system, but then I have a room with almost no early reflections, which does improve the imaging a lot.
very interesting...
My own take on advanced EQing (like done with Ulli's Accourate, Bodzio's Ultimate Equalizer - which I both have - and any other software or hardware tools the like) is that there are brick wall limitations for correcting ASAR patterns - though - in *general* these advanced EQing tools are great at any measure.
(Acoustic Source & Acoustic Room patterns)
++++
Hi Stig Erik - welcome back - hope you got enlightened by your one day trip!
Michael
Last edited:
Thanks StigEric
Thanks StigEric for sharing your evaluation of the MG 1.7's.
I have the MG 1.7's myself, they are nice speakers but as you say "with some limitations".
I will probably keep my 1.7's ... for now anyway. I have also decided to go on a long term DIY adventure looking for something more satisfying.
I have just recently seen that Magnepan has come out with their DWM woofer panel, it is supposed to be good for ~ 40 Hz --- 5 KHz. I think it would be interesting to get a pair (or more) of DWM's, and see what can be done with them mated to a pair of RD50's or maybe Raal Dipole ribbon tweeters.
Thanks again!
Thanks StigEric for sharing your evaluation of the MG 1.7's.
I have the MG 1.7's myself, they are nice speakers but as you say "with some limitations".
I will probably keep my 1.7's ... for now anyway. I have also decided to go on a long term DIY adventure looking for something more satisfying.
I have just recently seen that Magnepan has come out with their DWM woofer panel, it is supposed to be good for ~ 40 Hz --- 5 KHz. I think it would be interesting to get a pair (or more) of DWM's, and see what can be done with them mated to a pair of RD50's or maybe Raal Dipole ribbon tweeters.
Thanks again!
I finally got everything together and had a chance to take measurements of the Magnepan 1.7's.
I really want to take measurements outside, but that will take some time before I get the chance.
These measurements were taken indoors, 5' (~1.53 meter) from front wall, 5.9' (~1.8 meter) from the side walls, and 8' (2.44 meter) ceiling.
I used the Dayton's OmniMic measurement system. The microphone was 1 meter from speaker, gating set at 4 milli-seconds, no audio equalization. Five sets of measurements were taken with different smoothing levels, no smoothing, 1/24 octave smoothing, 1/12 octave smoothing, 1/6 octave smoothing, and 1/3 octave smoothing.
I will try to attach a screen capture of the 1/6 octave smoothing, and the NO (0)smoothing Impulse Response text file. The Impulse Response text file is in the .frd file format, the file suffix was changed to .txt so that it could be uploaded, it can simply be changed back to .frd if you want.
I hope the attachments work the way I think they do!
Let me know if you want any of the other files.
At the bottom of the photo is a 10 measurement average of the Frequency response everything else is just one measurement.
I really want to take measurements outside, but that will take some time before I get the chance.
These measurements were taken indoors, 5' (~1.53 meter) from front wall, 5.9' (~1.8 meter) from the side walls, and 8' (2.44 meter) ceiling.
I used the Dayton's OmniMic measurement system. The microphone was 1 meter from speaker, gating set at 4 milli-seconds, no audio equalization. Five sets of measurements were taken with different smoothing levels, no smoothing, 1/24 octave smoothing, 1/12 octave smoothing, 1/6 octave smoothing, and 1/3 octave smoothing.
I will try to attach a screen capture of the 1/6 octave smoothing, and the NO (0)smoothing Impulse Response text file. The Impulse Response text file is in the .frd file format, the file suffix was changed to .txt so that it could be uploaded, it can simply be changed back to .frd if you want.
I hope the attachments work the way I think they do!
Let me know if you want any of the other files.
At the bottom of the photo is a 10 measurement average of the Frequency response everything else is just one measurement.
Attachments
MigeO,
The situation is simpler than a corrupted file.
By default: "OmniMic always aligns "0 ms" to occur at the highest peak in the impulse response. Normally, the frequency response calculates from this point forward, and the peak of the response will be the time reference plane. In advanced mode, you can also select to have OmniMic start calculation from a different first point."
I realize the Frequency Response calculation begins at the zero reference point (the highest peak) ... as the manual says, but I just assumed the program would save the entire selected 4 ms (shown in red on graph) to file, apparently it does not. I wonder though if this is an OmniMic program bug.
Well, live and learn. The user manual really does need a lot of work to be more informative and useful.
I will be happy to do the measurements again as soon as I can, actually it is sort of fun, it's just a matter of timing and circumstances.
The situation is simpler than a corrupted file.
By default: "OmniMic always aligns "0 ms" to occur at the highest peak in the impulse response. Normally, the frequency response calculates from this point forward, and the peak of the response will be the time reference plane. In advanced mode, you can also select to have OmniMic start calculation from a different first point."
I realize the Frequency Response calculation begins at the zero reference point (the highest peak) ... as the manual says, but I just assumed the program would save the entire selected 4 ms (shown in red on graph) to file, apparently it does not. I wonder though if this is an OmniMic program bug.
Well, live and learn. The user manual really does need a lot of work to be more informative and useful.
I will be happy to do the measurements again as soon as I can, actually it is sort of fun, it's just a matter of timing and circumstances.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.