Tuesday, 1 April 2014

Room Equalizer Wizard with JRiver Media Center

I wanted to document a very neat feature where you can use Room Equalizer Wizard (REW) to automatically generate a list (say 20) parametric EQ's that are generated by the program automatically based on the measured frequency response of your speakers.  How cool is that?  Home theater receivers have been doing this for a long time but I've always found them to be compromised.  Using the Daytom UMM-6 Reference Microphone is a major step ahead along with the excelent REW software.  This post assumes your using a music server which has replaced your CD player.  If you haven't made the move to a music server then you should.  The sound quality is far superior to any CD player.   Just look at the recent finding by UHF Magazine.  They found that the Stello U3 USB interface > Moon 300D DAC beat out their reference Linn Unidisk 1.1 CD player.

So let's get started!

First Download the two pieces of software.

JRiver Media Center

Requires a login for Home Theater Shack

Installation of JRiver is very straight forward.  You just need to specify what output you'd like to use from your PC or laptop.   In my case I'm using my the USB output on my laptop so I selected the Kernel Streaming option.
You will notice the parametric equalizer under Options >  DSP & Output Format

Shown are the parametric EQ settings I've already created.  To create one you simply click Add  > Adjust a Frequency (Equalizer Filter).

Run REW software and load your calibration file if required for your measurement mic.

Next you'll want to set the appropriate output and input.

You can then begin measuring.

After you run your measurement you should get something like this.

You may want to apply some smoothing to your results.

Next we want the REW software to automatically generate an PEQ setting to flatten your curve.

Next  you'll want to work your way down the right side changing the setting as follows:

Continue down the right side and once everything is set click on "Match Response to Target".
Next you will see your computer generate all the parametric EQ settings required to match to the target response within a tolerance that you can specify.  In my case I selected 3 dB.

When that is complete your screen will look like this.  You'll want to click on EQ Filters to see the individual settings required.

You'll want to sort the list based on frequency.

Next you'll take these values and manually input them into JRiver.  It takes some time and you should use split screen to do this.

You are done!  Close the window and apply the settings.  If you don't like the results I suggest limiting the maximum boost permitted when you match the response to the target. Say only 3 dB instead of 6 dB.  If you really want to nail it down you can measure each speaker individually.  

Monday, 10 March 2014

Heathkit AS-10 Speakers with Jensen Alnico Drivers

Recently I acquired a pair of vintage Heathkit AS-10 speakers that use the venerable Jensen P10PF 10" woofer and two 3" Jensen alnico paper cone tweeters mounted to a sheet metal plate that angles the two tweeters slightly off axis horizontally to provide a little broader coverage.   Also the sheet metal is required because the tweeter can only be mounted on its front face.  If it was mounted directly to the front baffle it would result in the tweeter firing through a hole the thickness of the baffle.

The cabinets a fairly heavy with real wood veneer and removable back panel.  
Inside the cabinet there is no stuffng material used and all the wiring is soldered connections.  
The Jensen P10PF driver was also used in Fender guitar amps and is known for its beautiful tone.  

The closed back jensen tweeters are familiar to me  and part of the reason why I bought the speakers. These tweeters are able to achieve something special compared to any other tweeter I've encountered.  They're able to reproduce vocals and electric guitar with a realism and believability that lets you forget everything and just enjoy the music.   Sure they have some limitations.  They don't resolve complex music very well, they have limited power handling, and somewhat ragged frequency response.   But they are wide bandwidth playing from 2 KHz up to 17 kHz.   The natural Fs of the tweeter is 1.7 kHz, so a crossover point of 3 kHz is recommended if using a 2nd order slope. 


These speakers sounded pretty unflattering when I first hooked them up.  The bass was deep (down to 47 Hz) but there was a serious upper bass resonance coming from the cabinet. 

The first order of the day is to stuff the box with some polyfill to help resolve this issue.  

I went a little over board at first try.  Stuffing the box right full resulted in losing most of the bass.  Although the cabinet resonance was gone so was the bass!  

I decided on a much smaller amount that seemed to quel the resonance just enough however not to steal the bass. 

Ok so my conclusion here is that these could stand to have some much better enclosures.  

Sound Continued...

Listening to the midrange it was obvious something was amiss.  There was this nasty harshness that demanded some further investigation.  

First step, measures impedance curve and frequency response. 

Below is the frequency reponse of the speakers.

Ok so things don't look too bad, nothing seriously out of whack.  I was surprised to see the drop off in the high treble but I would deal with this later. 

In order to figure out where he harsh sound is coming from I needed to measure the woofer and tweeter separately.  Being bi-wirable speakers this was easy to do. 

Below is the frequency response of the low frequency and high frequency separately. 

Yikes, ok. That midrange bump is coming from the tweeters.  The next step would be to determine the natural resonance of the tweeters. This determines how low a crossover point can be achieved since you don't want a tweeter playing in this spectrum simply because it sounds very harsh; kinda like what I'm hearing with these speakers.

Below is the impedance curve of the Jensen tweeter. 

Ok so yes there is a natural resonance at 1.7 kHz that likely is contributing to the harsh sound that I'm hearing.  
This is indicated by the small bump in the impedance plot at our suspect frequency. 

Ok, so at this stage what does one do? Replace the cap?  The cap says 3.5 uF which is easy enough to replace but there is another easier thing to try first that could identify the solution.

Being bi-wire I can insert an additional capacitor outside of the cabinet to drive up the crossover point of the tweeters to quickly tell if there is a subjective and objective improvement.  

Shown below is a 1.5 uF cap wired to the terminal lugs.  

The resulting frequency response is shown below as the red line.  

My subjective listening does confirm that the this bump at 1.7 kHz was the source of the harshness.  The sound is much improved and the dip that is now present is slightly noticeable but not nearly as objectionable as what was before.  
Also notice the green curve that rises from 7 kHz up to 17 kHz.  This was a result of me bypassing the 16 ohm resistor found on the crossover!  Why this resistor is ther is beyond me, maybe it was there to reduce turntable hiss and pops, I don't know.   This alone was another giant step forward in improving the sound.  

Sound continued... again 

Doing these quick changes to both speakers and sitting down and listening to some good music revealed that these are some really great speakers.  The bass is the weak point (I blame the cabinet) but from there on up everything is dynamic and detailed able to reproduce details such as brushes and natural decay in reverberation.  Percussion in particular is very lively and enjoyable.  

I bought these speakers because I wanted to evaluate mainly the Jensen P10PF woofer for use in another project where I was to build a large Transmission Line enclosure utilizing my DATS speaker test system to create a design.  You see, older drivers although good, didn't have the Thiel & small parameters available so that you could design a proper box. A lot of the design was trial and error.  However with technology available today it's possible to take advantage of vintage drivers and design them into projects. Hornresp is able to optimize a design that would otherwise take years. 

Are the vintage drivers worth the trouble?  So far my most successful projects sonically have been when using modern drivers such as the B&C DE120 compression driver. There is just no comparison to the sonic beauty that this driver can produce.  Also, more recently the Celestion AN3510 driver proved out to be a clear winner.  And it was designed using the latest in Finite Element Analysis software according to Celestion.  

I would say that there is a charm to the vintage sound and I am going to continue to do some projects using vintage drivers simply because it is interesting to blend old with new and maximize the old in ways never possible before.  


In the meantime I will replace the old capacitor as I suspect that it's value has drifted.  (Maybe)

I'm not sure if the L-pad is required either.  It was obvious that the tweeters didn't need attenuating at all.  The tweeter pair struggle to match the efficiency of the woofer.  

Apon writing this I just remembered why bypassing the resistor brought back the highs.  Ceramic resistors contain a coil of wire to create the neccisary resistance.  The problem is that this coil has inductance which causes the highs to fall off.  I didn't realize this right away because I always use metal oxide resistors which have zero inductance.    There ya go.  Lesson learned.  Never use ceramic resistors.  

Sunday, 9 March 2014

Celestion AN3510 TL Spectral Decay Plots

Below are my measurements of the Celestion AN3510 
I've measured another speaker of similar size (Nuance SAT1) for comparison.  
Also I've shown the factory published frequency response a similar size driver from Fostex, the FF85WK.   I think the results certainly favour the Celestion. 

I used a protractor to ensure accurate off axis angles. 

Wednesday, 5 March 2014

Celestion AN3510 TL Project Measurements

This is a continuation of my previous post.  As promised some measurements of my Transmission Line 1/3 wave project using the Celestion AN3510;  a 3.5" aluminum cone neo wide band driver typically used in pro audio line arrays.  

The frequency response looks pretty good.  There is quite a rise starting at 7 KHz but this doesn't translate into anything overly bright because the driver is very directional at this frequency.  It also provides great extension out to 18 KHz!  Wow. 

It's only 5 dB down at 100Hz which when placed in 1 pi space brings this up to normal.  

I decided to measure the vent output as well.  It's producing solid output right up to 1khz!  I guess it's a good thing I placed the vent close to the driver!  I bet that's where the waviness is coming from between 200 Hz - 1Khz, a comb filtering effect between driver/vent output.  I'm not sure if I should want output up this high.  I will make some adjustments later to reduce this to see if it makes an improvement in sound quality. 

Next I decided to measure the impedance curve and phase using my DATS system.  Aligator clips on. 

Not your typical impedance curve.  I'm not sure what the large peak is at 130 Hz.  
The impedance curve shows that my tuning frequency is exactly 78 Hz.  The MH-Audio calculator predicted 74 Hz for this design so I am pretty close.  I will continue to experiment to see if I can lower the tuning a bit.  Also notice hat there is no cone modes until 11 KHz.  Pretty good! 

Monday, 3 March 2014

Celestion AN3510 Full Range Driver

As of late I've been looking at building what I call a 'Late Night' Speaker for a bedroom or small apartment.
I wanted it to have the following:

1) Affordable
2) Single Driver
3) Easy to build
4) Get the maximum bass possible from a small driver in the 60-120 Hz range

I wanted also to try something new, something that breaks the mold from the usual Fostex bass reflex desktop speaker.

I also wanted a test mule to verify my simulation in Hornresp so that I could rely on the software for bigger and more expensive projects.

Enter: The Celestion AN3510 full range driver.  I found this driver myself and thought it was interesting for the following reasons.

-Aluminum cone and dustcap
-Neodymium magnet
-Non-resonant plastic basket
-Flat frequency response (except for narrow peak at 10 kHz)
-High EBP value of 158.  Suitable for Horn Loading or Transmission Line.
-Affordable at $43.00 from Q Components

The other thing that interested me was how small the magnet was which would minimize the back reflection.  I also noticed after receiving the drivers that the air space under the spider is ventilated as well.

I had a hard time deciding what type of enclosure to build for this driver.  I initially was going to build a wide baffle design similar to the Jordan Acoustics VTL enclosure seen here.

I decided against this because I didn't like that the vent was so far from the driver.  I was planning on seriously relying on the vent to aid all the way up to 200 hz, so I figured the vent should be somewhat coincident with the driver.


I decided instead to go with a very simple design.  I went with a 1/3 Wave Transmission Line based on the online calculator from MH-Audio.

I also simulated this design in hornresp and was able to perform the 'combine response' command to see the resulting combined frequency response of both the vent and the driver.  Since the vent is only 5.9 cm away from the driver there was very little comb filtering effect.

Next I drew the basic box in Solidworks so that I didn't screw anything up out in my cold garage.

This is shown with the speaker complete with the side panel removable to adjust the poly fill amount.
The MH-Audio Website suggests 50% fill and this is what I did.   I used double sided sticky tape to hold the poly fill in place.

Shown in my bedroom hooked up to my Miniwatt N3 tube amp.

I am very pleased with the result.  These celestion drivers are a nice break from the forward shouty sound from some fostex drivers.  They have a nice smooth flat frequency response and produce very nice music.  I listened to them all weekend and they never got irritating or aggressive.  

This was my first Transmission Line Enclosure build and I was amazed at how well it works.  There is tons of bass coming out of the vent at the top.  So much so that when you cover it you lose everything below 200 hz.  I ran a tone generator app from my iPhone and was able to get very solid bass down to 65 Hz where it falls off quickly.  The bass is very tight and tuneful and not boomy at all.  

I was concerned about the 10 kHz peak shown in the specs but I never actually heard any peakiness in the driver.  In fact it sounds a little rolled off in this area and could use a super tweeter.

But overall I really like the pinpoint imaging and great soundstaging that these speakers produce.  

It was a fun project and highly recommended.  

It was nice to see the hornresp simulations proved entirely accurate, however I haven't measured the frequency response but I will do so in the near future.

The other nice feature of this type of design is that you can cover or adjust the top vent to control the level of bass.   Just place another piece of wood on top or use a book or something.  Especially useful if you don't want to wake the kids!