The “problem” with impulse response for amp cabinet sounds

This is an updated version of a post that lacked good technical support. Thanks to Normen and Selig for their expert knowledge!

In modern guitar amp modelling software often convolution techniques are used to capture the sound of the speaker cabinet. Via impulse response files the sound of the cabinet is sampled and can easily be applied to any input. Although this is a very popular technique it is in my opinion not perfect for getting a great speaker cabinet sound.

Linear sound

An impulse response is like a sample that can be superimposed to a signal. It is a linear process which can’t add non-linear compression or saturation to a signal since it’s a linear technique. Using an impulse response the input signal won’t be able to drive the output in non-linear ways. This means that many of the irregularities of a real speaker are not present when using an impulse response. A speaker can breakup in a non-linear way and we’re not able to capture this in an impulse response.

An impulse response of a speaker cabinet captures the frequency spectrum of the speaker cabinet with a fixed amount of breakup plus the ambiance of the room in which the cabinet is placed.

A perfect cab filterpalmer-pdi-03

In 1993 I bought a Palmer PDI-03 Speaker Simulator. The speaker output of my Koch, Vox and Marshall tube amps can be attenuated into a line-level signals when using the PDI-03 plus its build-in filter that will simulate the sound of a speaker cabinet. The Palmer in fact is a loadbox + speaker simulating filter. Its sound is very direct, lacking any ambiance. The sound is also extremely dynamic since it lacks the irregularities and breakup a real speaker will add to a signal. But this can be a nice thing I found out! It’s a much cleaner signal, but one without the irregularities of the speaker and without colorisation of a cabinet.

A guitar amp speaker cabinet is a midrange system. It’s not like a HIFI speaker system but often uses only 10 or 12 inch speakers. It’s frequency curve can be simulated using an EQ to filter the sound. The Palmer showed me that this can create a fantastic sound which supreme dynamics.

But I also believe that the non-linearities are an important factor for getting a great guitar sound that reacts nicely on playing dynamics. An impulse response can capture the sound of a cabinet but it is not able to capture the breakup sound of the speaker. And this is a bit of an issue in my opinion.

I must say I am not a technical genius and probably never will be one (read: I am not a scientist and don’t wanna become one!). But I have trained my ears over the years so I am now more capable than ever before to listen in 3D and explain what I am hearing (frequency peaks, saturation, compression). And I am now able to design the sounds I am hearing in my head for many years. Sounds which are an improved version of the equipment I have been using for many years, like some of the tube amps I mentioned above. Because I always felt something was not right. Some amps have a great overdriven tone using the bridge pickup but start to sound muddy when using the neck pickup. That’s why I came up with the Rockmen ReFill, using nothing but the stock devices of Propellerhead Reason. I was able to create my own custom AMP which worked great on all my guitars using all kinds of settings. No impulse responses or Rack Extensions are needed.


Also published on Medium.

, ,

2 Responses to The “problem” with impulse response for amp cabinet sounds

  1. ThatOneGuy December 20, 2016 at 6:06 pm #

    The logic in this article is completely flawed.

    You basically said:
    Speaker impulse response doesn’t accurately reproduce the irregularities and break-up a real speaker produces and that means it’s bad

    My palmer doesn’t accurately reproduce the irregularities and break-up a real speaker produces…but it’s good!

    Also you can’t “listen in 3D”, you can listen in stereo because humans have two ears, but 3D is just the length, height, and width of a physical object. Sound is just fluctuation in air pressure, not a physical object.

    Yes, both impulse response and EQ simulation aren’t 100% accurate. That is a problem, you’re right, but you could get the same dynamic and clean response you find desirable in your Palmer by sampling an impulse response from a speaker that’s not being pushed into break up or compression.

    Besides most guitar speakers aren’t often pushed into break up or compression. Unless, we’re talking about a low wattage speaker being pushed by an amp that puts out a higher wattage than the speaker’s rating. Most speakers that get paired with an amp match the wattage from the amp, or can handle higher watts than the amp provides. Also, the sound of a guitar loudspeaker distorting is actually way less desirable than most people think.

    Most of the desirable break up and compression people look for actually comes boost, drive, and/or distortion pedals, or the break up of the tube preamp section.

    The best and most desirable break up and compression that your looking for is when the pedals, preamp, output tubes, and the speaker are properly balanced with one another. Which means the speaker breakup is a small part of a much bigger equation.

    If you hear someone with awesome tone, 90% of the time the speaker is adding very little break up or compression, if it is even adding any break up at all.

    • Marco Raaphorst December 20, 2016 at 6:41 pm #

      What I meant to say is that IR can only produce the sound of break-up as a sample (a one shot sample). The Palmer doesn’t produce any break-up BUT that can sound very nice as well. It is super dynamics. You are right that the break-up is probably not so important as the other parts of the signal chain. And I must say I do get great results using Impulse Reponses as well.

      You’re right about “3D”. Not sure what when through my mind when writing it. It’s hard to explain. In the end it all comes down to what you personally like or not. Playing and listening.

Leave a Reply