Speaker Driver Emulation With Shaped Noise
using a DAW, white noise, parametric EQ and a spectrum analyzer plugin
last update: July 4, 2021

Copyright 2013-2021 by H. Gragger. All Rights Reserved. All information provided herein is destined for educational and D.I.Y. purposes only. Commercial re-sale, distribution or usage of artwork without explicit written permission of the author is strictly prohibited. The original units  with their associated  trade-names are subject to the copyright of the individual copyright owner. The Author is by no means affiliated with any of those companies. References to trade names are made for educational purposes only. By reading the information provided here you agree to the Terms of Use.


The Pain Of Choosing A Suitable Driver For Your Instrument Cabinet
Creating Emulator Filters Using A DAW And Shaped White Noise
Sound Samples
Update 2021

The Pain Of Choosing A Suitable Driver For Your Instrument Cabinet

If you are one of those who make their own cabinets for HIFI, you are in the wrong place. This is about making an assessment on a driver
for a musical instrument cabinet purely based on the supplied frequency charts .

If you believed into the manufacturerīs marketing slogans, all speakers would be equally excellent, but this is not the case.

I recently built a 10" detuned cabinet and developed and perfected the subsequent method which lets you  audibly compare a driverīs response with other specimen of interest.

Of course, the cabinet will modify the driverīs response to an extent. A too small cabinet will diminish bass response for example, a big cabinet will exhibit beaming. A speaker may sound differently under heavy load. A tube amp might emphasize low frequencies. But what would those speakers sound like if they were all built into a cabinet where they īfeel wellī, all other parameters equal?

Back To Index

Creating Emulator Filters Using A DAW And Shaped White Noise

Many of us musicians have a PC, and running a DAW on it is cheap and unproblematic these days.
All we basically need for our task is a short audio file containing white noise. I made one myself using Audacity, a freeware wave editor program. It can be short because it is run in a loop during replay.

Next I used a freeware parametric EQ with multiple bands (NCL parametric EQ) and a freeware spectrum analyzer type display (voxengo SPAN). You are going to need several bands to get all the dips and dents and rolloffs right.

I went and created a few EQ settings that (hopefully) emulate the responses of a driver. This is easier than you think.

I used the following setup recipe:

  • find or create an audio file (.wav) that contains white noise. Audacity can create noise (white noise contains all frequencies with equal levels.)
  • Loop playback this track in your DAW.
  • load the EQ and the spectrum analyzer into the effect chain of the noise signal.
  • Adjust the spectrum analyzer until you see about the same range of decibels on the screen as on the graph from the driver datasheet displays (Note: the absolute level is irrelevant)
  • Set a LP and a HP filter on the param eq with 12 dB/Oct rolloff (that is what speakers do...) set the frequencies, Qīs (resp. Bandwiths) and gains to complement the driverīs graph. You may need several 12dB filters cascaded to emulate the correct rolloffs.
  • Set additional peak filters for notches and peaks
  • Use shelve filters for gain steps.
  • Save this "patch" under the name of the driver you are emulating.
  • Check the decibels for peaks and notches twice, also the octave rolloffs.
  • Make a couple of such responses of the drivers of interest.
  • Make note of the signal level at one kHz of the loudest driver you have and for all others adjust the overall eq gain so that the response reflects the dB level relative to the loudest. (For example, if you have a -45 db noise signal at 1kHz for the most efficient driver, say 100dB, and some other will have an efficiency number in the datasheet of, say, 98dB, adjust the overall eq gain to -47dB and store that.)
Note: Due to the logarithmic nature of the ear, 3dB changes in loudness are barely audible anything below that will be inaudible. However, you may notice that a speaker model appears louder than another, although it has a lower efficiency figure but has a high peak in a frequency band that the ear is most sensitive at. Above that, not all maker suse the same system when creating their charts.

This procedure leaves us with some EQ settings that resemble our speakerīs frequency behavior and provides us with some hard and fast criteria for audible comparison of unaltered instrument recordings if we were careful.

DAW fx list
White noise file with effects looped over in the DAW (click to enlarge).

In the effect chain, the first effect loaded is the equalizer, followed by the spectrum analyzer.

The white noise file is looped over and over.

The DAW you see is Mixcraft 6.

param EQ
10x param EQ (NCL parametric EQ) (click to enlarge).

A look at the equalizer settings for the deltapro-12a.

Note the overall gain setting (right) is set so that the display shows -45db@1kHz.

Voxengo SPAN (SPectrum ANalyzer) (click to enlarge).

Note the overall gain is set so that the display shows -45db@1kHz for the driver with the highest efficiency (deltapro-12a). The visible dynamic range has been chosen similar to the datasheet values (about 35dB overall)

                                                          analyzer with
Voxengo SPAN with driver graph overlayed (click to enlarge).

The graph of the deltapro-12a has been overlayed (source: Eminence datasheet).

Note that the scale for SPAN is uneven. Both graphs meet at -45dB (SPAN) with one section of the overlayed graph meaning 5dB steps, so they both meet again at -75dB.

Not a bad emulation.

Faital Pro 10PR300 emulator (click to enlarge).

The same visible range of dBīs as in the graph above has been used.

Any alleged relative difference in efficiency (value at 1 kHz) has been taken into account. Although the graph does not seem very different to the 12" driver above the sound inside the (half size) enclosure is different. This is what cannot be taken into account.

Back To Index
Sound Samples

Note as per 2021: with the discontinuation of Adobe flash player those samples may not be available.
I later recorded some piece of:

    * clean guitar (chord comping with western guitar)
    * distorted guitar (electric guitar with overdrive)
    * steel guitar (clean)
with a full frequency range (not using any speaker emulators or real speakers) in a direct-to-console recording method and then looped over those tracks one at the time while cycling through the driver simulation filters. The shaped noise signals give you a good guess at how the filter sounds.


none (unprocessed signal)

Eminence Delta-Pro12A

Reference dB level (efficiency); 12" guitar speaker w/ ferrite magnet. EVM-12L clone
Eminence EPS-15C

15" guitar speaker w/ ferrite magnet and aluminium dustcap, used for steel guitar.
Sica PL15B25S

15" bass speaker, used for steel guitar. Neodym magnet.
Eminence Kappalite 3010MB

10" power mid/bass PA pro speaker w/ neodym magnet
Faital Pro 10PR300

10" power mid/bass PA pro speaker w/ neodym magnet
Eminence  Delta-10a

10" bass/mid driver w/ ferrite magnet

Back To Index

I finally chose the Faital Pro 10PR300 for my 110 detuned cabinet with the help of the above method. My favorite was the Kappalite, but this seemed hard to get. Both sound really similar to the Delta-Pro12A, which is sitting in my 112 enclosure, with is my reference sound for all purposes. Inside the box, for the earlier mentioned reasons, the Faital Pro sounds more different than the comparison of soundfiles would lead to believe, but good. So this is a method for selection, where no method was existing before. This is a clear advance.

Back To Index

Update 2021

Some years have passed, and some VST plugins have changed. I no longer use the NCL parametric, and it may potentially create some problems with current DAWs. Although I use Mixcraft, I prefer Reaperīs free Rea plugins, here the ReaEQ for filtering. This is a fair standard even Proīs use. If you use Voxengo SPAN, you should remove the SLOPE they use for all their presets for reasons I do not understand. If you run white noise into it, it will show a line that starts bottom left and goes to top right. Go to the EDIT menu and set SLOPE to zero. You may want to wind up AVG TIME to 6-7 seconds for a steadier display.
While you may continue using any source of white noise, I use MELDA MNoisegenerator, which is a free, uncomplicated plugin that does what it says - make noise.

I found another good use for such a speaker emulation filter. My amp has a (balanced) DI out, which is just an unprocessed copy of the preamp signal. I made up a filter whose response mimics the graph from the datasheet, using ReaEQ. This lets you save presets, so that I can recall the steel amp filter any time after I have made a recording. Fact is, that although you hear a change in tone listening to white noise while playing with the filter parameters, there is little you hear on an actual steel guitar recording. You hear the top end changing slightly, itīs becoming more nasal due the 3.5kHz high roll-off of the 15" driver, but not as much as you would expect. You also donīt hear the little dips and peaks in the dB range that are there.
However, this is a way to hear on your recording what you heared through the speaker.

This approach to a speaker simulator does not take any complex phasing into account, that is inevitably introduced by using complex filters. If this is of concern, you may resort to a simple high pass + low pass  filter, particularly if you feel the high end on the recording needs more taming.

Back To Index

Update History
  • July 4, 2021: first release
  • Aug. 4, 2013: first release
Back To Index


(c) 2013-2021 AQUATAUR Musik & Elektronik