I have been a music lover all my life, preferring classical and a little folk, bluegrass, and Jazz. I have a huge collection of LPs and CDs that have been accumulating since the 1950s. I bought my first "real" HiFi system in the early 1980s. It consisted of a Pioneer SX-1980 receiver (the best and most powerful receiver ever made), Magnepan MG-1 speakers (in black), A Phase Linear Turntable, and a Teac reel-to-reel tape recorder. I also invested in DBX (a linear expander/compressor) to improve the tape quality. I still have all these components 30 years later.

With the advent of computers, I listen a lot while working at the computer, and found that it was most convenient to rip my CDs and some LPs to my computer. I could then play them around the house using my (sadly discontinued) Logitech Squeezebox server and remote receivers. When I started to do this, disk storage was expensive, so I recorded all of these rips in mp3 format, albeit in the highest quality available—320 kbps. About four years ago, I started to download many FLAC (uncompressed) recordings from eclassical.com, and discovered that I preferred the uncompressed format. So since then, all of my rips have been made using iTunes to the apple m4a lossless format. Hard disks are cheap nowadays, and I keep five copies of my music collection.

A year ago, because I have a bad back, I treated myself to a Shiatsu massage chair, which is situated in the living room in the "sweet spot" for listening to my HiFi. Since I now spend about an hour a day in that chair, my HiFi listening hours have increased. So I decided to get some new speakers after all these years. I am hooked on dipolar sound (where the speakers radiate forwards and backwards also). I have listened to my MG-1s for many years, and even had the use of two pairs of KLH Model 9 full-range electrostatic speakers for a year. After reading many rave reviews about the new Magnepan MG-3.7i speakers, I bought a pair without even listening to them. My nearest Magnepan dealer is in Atlanta, and I negotiated over the phone for them to get them drop-shipped to me directly from the factory.

My HiFi system

As you can see, I have a large living room (about 19' x 45') with a sloped cathedral ceiling. The above picture is taken from my listening position. I spent several months adjusting the speaker positions, and opted for the narrower seating coverage by putting the tweeters on the insides. In the picture in my HiFi cabinet (which I built myself), you can see the SX-1980 and the Squeezebox Transporter. My turntable and tape recorder are in the end units, obviously not used a lot because their access doors are covered.

I also decided that the 3.7is were rather unsteady due to their long lever arm from the supplied feet, and looked into getting speaker stands. I decided on a set from Mye Sound. They were well worth the $630 I paid. It did take almost two months to receive the stands because they are built to order, and the owner was unsatisfied with the first paint job. These stands are heavy and sturdy and come with nice adjustable brass legs.

The stands attach to the holes used for the original legs, and the top of the strut attaches under an existing screw. Aside from the weight of the stands and the speakers, the stands were easy to install. But I would need a dolly to move them now. Sharp eyes will see my new SVS SB13-Ultra subwoofer in the left corner of the room, and my old Velodyne 12" subwoofer along the right-hand wall. Alas, the Velodyne literally melted its surround when I was testing it along with the 3.7is. I got the surround replaced for under $100, but the electronics is in some kind of oscillatory mode. 

The speakers did sound better with the Mye stands, and are totally immovable now. I was, however, missing bass below about 45 Hz, and decided to buy a new subwoofer. The SVS SB13-Ultra has gotten rave reviews (especially for the price), including in the latest issue of Stereophile, so I bought one (without listening again). I have 45 days to return it to the factory if I an dissatisfied, but I love it.

I listened to the old Mercury 1812 overture with Antal Dorati conducting, and the music sent shivers down my spine. Literally. The canon were in the room, and I realized for the first time that the bells at the end included those from a real church carillon. I do not think I need a second subwoofer, but time will tell. I have also not yet tried to calibrate my system except to set the level by ear. The SB13s have two filters I can use to fix the room resonances. I will report later on my equalization efforts. Of course, there is nothing built into my ancient SX-1980 to do this. When I listen to Magneplanar speakers, it sounds to me as if the players are in my living room, as opposed to me being in the concert hall.

With the new speakers, I can easily tell the difference between mp3s and lossless rips. The mp3s sound grainy during loud or congested passages. Sort of like someone was rubbing sandpaper over the sound. About half of my 71,000+ ripped tracks are still in mp3, and I have embarked on the tedious task of re-ripping them to lossless.

A word on the SX-1980. I am lucky to have a local guru who can rebuild/fix high-end HiFi equipment. Almost all of the electrolytic capacitors in my unit have been replaced. The power supply has been rebuilt. In its last overhaul, they found a blown diode and transistor that caused increased distortion in one channel. But the SX-1980 still sounds better than all more modern amplifiers that I have auditioned. And it is very difficult to find a great two-channel system nowadays.

Measuring the system

I came across a very nice program called REW for making speaker measurements. You will have to register with hometheatershack.com to download it—a straightforward but tedious operation. The user's manual is available here.  I installed the software in the Windows 8.1 partition of my Lenovo laptop. But the internal sound system in the laptop is inadequate for these measurements, so I purchased an external preamp, a TASCAM US-322, which I got from B&H Photo for $79.99. But then I learned that the microphone I had planned to use only went down to 70 Hz, which would not cover the subwoofer range. So, I needed a sound level meter. I settled on the T Tocas Professional Sound Level Meter from Amazon for $48.99, primarily because it has a USB interface, as does the TASCAM preamp.

But I was not done because I needed cables to connect everything together. The TASCAM has 1/4" phone jacks and XLR connectors. The meter has 1/8" output jacks in addition to the USB. I also needed a long RCA phono cable to go from the TASCAM output to my receiver's tape output jack. I was rather clueless about how I would get the signal generator in the REW software. Here is the setup I used:

The measurement setup

I had to override the driver signing requirement of Windows 8.1 in order to install the driver for the sound level meter. The first day after that, it appeared as a sound input device, but not today. so I had to run a cable from the dc output of the meter to the input of the TASCAM, which I had to set to guitar. It took a day to figure this all out. 

The REW software is quite capable, and I still have not delved into all its complexities. Nonetheless, I was able to measure the frequency response of my system

Frequency response graph

The frequency response of the meter is rated to be from 31.5 Hz to 8.5 kHz, so the high end of this plot should be ignored. But the meter seems to go a lot lower than its specs imply.

The first thing I discovered was that I wanted to move the crossover between the Maggies and the subwoofer from 50 Hz to 40 HZ. I Raised the gain of the subwoofer significantly. Then I used the two parametric equalizer settings to try to remove notches at 38 Hz and 78 Hz. The above plot is the average of about 8 frequency sweeps. I then listened to the 1812 overture again, and the canon were even more awesome, and there was no boominess. I may reduce the woofer output a db or two after more listening.

More measurements

I bought a better microphone, and Audio-Technica AT2022, which covers the full range of 20–20000 Hz. So I replaced the sound level meter with the AT2022. First I disconnected the subwoofer and measured the Magnepan 3.7i speakers alone:

MG-3.7i withouyt subwoofer

As you can see, the Maggies make it down to about 45 Hz. I have averaged multiple plots to smooth out the high frequency hash. Each run averages 4 sweeps. So a subwoofer really was needed. I do not trust the measurement above about 15 kHz, because I can hear the signal generated getting lower in frequency instead of higher. Next, I reconnected the subwoofer:

Varying the filters

The subwoofer certainly fills in the bottom part of the spectrum. I fiddled with the first equalizer, set it to 31.5 Hz and varied the Q. Finally, I removed the higher frequency filter and ended up with my "final" plot:

Fanal Tweaks

This plot has been averaged over 4 runs of 4 sweeps each to clean out the high frequency noise a bit. I also moved the microphone forward and backwards about 4 inches. Note that the scale here goes down to 15 Hz, where I can feel the bass. The volume of the subwoofer has been significantly reduced from the first settings attempt.

Measurements redux

The REW forum answered my question about the tone going down rather than up:

JohnM wrote: View Post
The sweep extends 1 octave above the highest frequency to be measured, or to half the sample rate being used, whichever is lower. If you hear the frequency go down again it is likely that your soundcard is configured to operate at a lower sample rate than has been selected in REW, probably 32 kHz, so aliasing occurs above 16 kHz. On Windows you can check that on the Advanced tab of the sound properties for the soundcard output.

Indeed, the US-322 was set at 32 kHz. I changed it to 44.1 kHz (it goes higher), and this solved the problem. So, I measured things yet again

Averaged response

The dip is at 85 Hz. The SB-13 equalizer has a setting at 80, and another at 90 Hz, and they did not help to tame this. Note that the response now is measured up to 20 kHz. You can have REW5 smooth this for you:

Smoothed response

I also measured other things via REW5:



Waterfall plot

The proof of the pudding is in the eating I guess, so I made two short videos of the speakers at work, captured with the new AT2022 microphone.

Microphone setiup

A clip from the Dvorak Requiem (Mariss Jansons and the Royal Concertgebouw Orchestra)


And one by Edgard Varese played by Pierre Boulez with Ensemble Contemporain


Of course, they sound a lot better "in person."

After more listening, I increased the toe-in of the speakers so that they cross in front of my listening position. This improved the imaging and sound stage. I also decreased the subwoofer output by 1 dB.

Last week (August 17, 2018) I drove to Atlanta to hear the new $30k Magnepan MG30s at HiFi Buys.


The room was really too small for these behemoths, and they kept playing lots of pop music with reverb and purposeful distortion, which prevent anyone from judging speakers.  But I got them to play a percussion disk from Kroumata, and the speakers sounded really good. But not an extra $20k good.

I did go home, and swapped my twp speakers, and I think they sound better now.


Submitted by Eugen on Fri, 07/31/2015 - 05:14


Aren't you bothered by that 6-8dB of presence band boost? Or do you only listen to quality classical recordings?

You are looking at the top of all the noise hash. If I averaged it, it would not be a big boost. Also, I was really measuring for the bass/subwoofer intersection. Finally, I did not subtract the microphone response from the plot, so take it with a big grain of salt. 

I am still really pleased with my Maggies, but they are not yet in the correct position because the imaging could be better. However, when I go to an actual symphony concert and close my eyes, I do not hear the pinpoint imaging that people seem to like. Things are much more regionally located. If I listen to a percussion ensemble (e.g., Kroumata), things are nicely spatially separated.

I think the Audio-Technica AT2022 XY-stereo microphone. If it is a directional microphone, it is not suitable for measuring. You need a uni-directional microphone for that.

Submitted by Qunn on Mon, 12/05/2016 - 12:14


Hi, would you mind updating more distortion results, like the SPL at 90, 95 and 100db.

Submitted by Mike on Wed, 01/30/2019 - 14:17


Nice article, you gave me some good ideas for further tuning of my Magnepan 3.6/R. Your FR looks reasonably flat, which subjectively can sound quite lean. I tuned mine to have a gradual, linear attenuation of about -1.5 dB / octave. Psychoacoustically, most people find this to be a "neutral" response. I listen mostly to classical and other acoustic music and I find with most recordings it is as close as I can get to the sound of being at the concert. Also, let me make a suggestion. I built big floor-ceiling tube traps and strategically located 4 layers thick RPG acoustic foam. This made a huge improvement in impulse and spectral decay. Makes the music even more clear, focused and dynamic.

Add new comment


  • No HTML tags allowed.
CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Image CAPTCHA

Enter the characters shown in the image.