I just tried an interesting tone comparison, one I’ve never seen attempted. It concerns the search for loud amp sounds at low volumes.
Have any of you ever experimented with speaker attenuators — the passive load boxes that reside between your amp output and speaker input, which let you crank the amp while maintaining a low level from the speaker?
I’ve worked with one model before, a borrowed THD Hot Plate, and thought it performed well. I decided to purchase my own attenuator after several Premier Guitar reviews of large amps. As a small amp fan (not to mention an aging player with fragile ears), I wanted to minimize the aural assault of evaluating loud-ass amps.
But first, I wanted to determine whether it’s legit to evaluate amps at attenuated levels. Does attenuation inevitably alter the tone? And if so, can you compensate for via recording software?
Online opinions about attenuators range from “works like a charm!” to “totally killed my tone!” So I picked up a Swart Night Light and started recording and measuring. (I didn’t compare rival products. I just went with the Swart for its reasonable price, solid online reviews, and dual outputs for driving two cabs. I didn’t A/B it with a Hot Plate, though the results seem roughly similar.)
I direct-recorded a brief guitar phrase using my black Les Paul with Bigsby and PAFs, and then ran it through a ReAmp to my early ’60s Tremoverb, a 35-watt Fender with two 6L6 power tubes. I dimed the volume and left the EQ flat. Tt was insanely loud in my small studio. After recording that, I tracked the same clip again using the attenuator at each of its three settings. The lowest attenuation setting reduced the sound from insanely loud to very loud. Medium attenuation reduced to somewhat loud. Strong attenuation produced a sound quiet enough to speak over. I recorded the results through Royer R-121 ribbon mic. I added a touch of plate reverb, but no compression or EQ. (Though I did normalize the files so they played back at similar levels.) In other words, you hear the same clip four times through a head whose settings never vary.
So did the tone change? Have a listen:
Do you hear what I hear?
IMHO, none of the clips sound particularly great. (Most amps, including this one, don’t sound their best at 10.) But the unattenuated loud sound has some qualities the others examples lack. The attenuated clips have a little less low-mid impact, and the higher-register single notes that sound a bit thin and prickly even on the original sound even thinner and pricklier post-attenuation.
Why, since the amp settings don’t change, and the performance are identical? Mics can respond differently at different sound pressure levels, and the relatively restrained speaker movement alters the result as well. Conclusion: the timbres of the attenuated signals are fairly faithful to the original, but there are slight spectral differences and a bit less body/fatness, especially on single notes.
Then I introduced some additional wrinkles:
Using the same files recorded above, I tried getting closer to the unattenuated sound using EQ. I turned to Logic’s Match EQ plug-in, which can analyze two recording and plot the differences in their spectra. I measure each attenuated clip against the old original, and then added the compensatory EQ adjustment. I analyzed each clip separately, but the EQ curves all looked a lot like this:
Match EQ says: “Boost low-mids and roll off highs.”
Here are the same clips again, but with Match EQ adjustments applied. I think these sound closer — do you agree?
To my ears, the original loud sound still has a touch more impact, but we’re getting within spitting distance. I could totally work with these attenuated sounds in a mix.
Okay, one last experiment: In addition to its attenuated speaker outs, the Swart box has a direct out, so you can record the head’s output directly at line level. (This part gets really interesting ….)
The first three clips in this final set were recorded direct with no mic. But before tracking them, I made an impulse response of my cabinet’s Celestion G12H speaker. Using Logic’s Impulse Response Utility, I ran a test tone through the speaker, deconvolved the resulting file, and loaded it as a preset in Logic’s Space Designer reverb plug-in. (I discuss the miracle of impulse responses here, including the fact that they can mimic speakers as well as reverberant spaces.)
The first clip is unprocessed direct out, and it sounds as fizzy an shrill as you’d expect (though I wouldn’t be above using it in some contexts). The second clip is the same track, but with the impulse-response speaker simulation engaged. (If you’ve ever wondered which aspects of electric guitar tone are due to the amp and which are imposed by the speaker, just compare these two files!) The third clip adds Match EQ for an even closer approximation. And just for reference, the final clip is the pure, unprocessed loud sound that got this party started.
To my ear, the last two clips are pretty darn close. I perceive slight differences, but not enough to make me care.
This was an interesting experiment! Based on this, I wouldn’t hesitate to use speaker attenuators while reviewing gear, though I wouldn’t want to do so without spending at least some time with the signal wide-open. Also, I definitely plan to record direct from the head more often. I like the idea of capturing true amp overdrive, but choosing between different speaker models according to the context. (You can do this with most amp simulation software: just bypass the amp section, playing through only a speaker model.)
Most of us have been guilty of saying, “But I can’t turn down — I need to be loud for the tone.” I often make fun of guitarists who say shit like that, even though I’m as guilty as anyone! But I probably won’t say it with quite so much conviction now.
So what techniques have you tried for getting loud sounds at quiet levels. I mean, other than — god forbid! — turning down?
If the attenuator is purely resistive, then the load that the amp sees is going to change as the attenuation changes. When the attenuator is switched in it’s going to be more resistive and when you’re going straight to the speaker, it will be more reactive (inductive). I’m willing to believe this has some effect on the tone.
I just looked up some DIY attenuator schematics and one of them had an optional inductor in series with the load resistor for “a more natural speaker response”.
Any idea if the unit has an inductor inside?
Might be interesting to compare your findings to a solid state amp, if you have such a thing – I’d expect it to be less sensitive to load changes of this sort.
Far as tracking goes, I have no cred – I have always used a DSP modeling preamp or direct to DAW with amp plugins.
ah, rigorous tests with good data make me so happy. Yeah, it sounds alright and I’ve heard a couple setups that made good use of different attenuators, but I’ve never had much interest in owning an amp that I’d HAVE to soak in most settings. I get the physics (well, generally) of why that much power sounds good, but it just feels somehow wasteful to use that much power and then burn it off.
Since part of the subtext here is finding good low-power and/or output studio options, might it make more sense to attenuate on the front end of a smallish amp rather than the back end of a biggish one? I’m thinking out of my rear here because I’m no expert on amp designs, but my impression is that the main advantage of big amps is the headroom that they allow, followed by the more complex distortion that occurs when you dime it (tubes, transformer, etc). So though I’ve never heard of anyone doing it, it seems obvious to my poor brain that if you start with a small amp, then you should be able to attenuate the input down to give you a ratio that provides more headroom and space for the signal to ‘bloom’ when you drive the different stages. It won’t sound just like an existing big amp, but it might still sound great on its own. No? Yes? Brilliant? Crazy? I haven’t had my coffee yet….
It’s all math right? so instead of fixing our calculations around the source, could we try starting with a power setting and then calculating backwards to find the right source attenuation that would be optimum for it? The ‘one knob’ become an input attenuator rather than a power soak.
well, freak my brain out. i see what u r saying. i did well in bonehead physics (concepts no math). but i still miss the speaker rolling that boulder up that hill for eternity, dig? i identify w the actual strain. can you really fool me/ my ears? i do own a hiwatt and my friend that used to play ‘beatles’ w pool cues, have been playing ‘the who’ for 20 years (he drums). my poor ears. i wonder if it would be that easy to replicate loud? on paper is one thing.
One issue I have is that I never found the best sound at 10 on any amp. Always find the sweetest spot in volume and attenuate from there instead. I miss my old Original Rockman power soak, it was peeled off the back of my cab and stolen long ago.
Fascinating! I’ve always been a fan of the software “sans amp” plugins because of the flexibility you get later on in the mix, but I love seeing the impulse response and the match EQ in action here. Could there be any more flexibility from here out? I can’t imagine it.
I could certainly use any of these tones if I had to for some reason. However I bought a house with a nice dry basement so I would never have to be concerned by volume limitations in my studio/practice space. I’ve also lived in the country for a lot of my life so I can get away with being loud.
I have tried many methods for limiting volume on guitar amps over many years. I tried limiting screen grid current to the output tubes as well as limiting B+ voltages which makes the tubes clip at very low power output down to 1/2 watt or less. Unfortunately this eliminated compression from the power supply among other problems. The sound was hard edged and useful for only some uses. I also have played with speaker attenuators including using light bulbs and inductors. The compression from the power supply was restored that was lost with screen grid attenuator and voltage reduction. Again it was ok but not the same due to a lack of compression from the speaker (speakers under high power have some interesting properties which increase due to temperature of the voice coil and magnet structure. This is even more true with alnico magnet which actually change magnetic flux under high power and excursion. It is my feeling that speaker compression is very key to the sound and more importantly feel. I found that the feedback loop in a power amp is also a big factor. The feedback loop taps from the output xfrmr secondary winding back into the phase inverter and it tries to make the output signal match the signal entering the phase inverter. The resonance of the speaker and what happens when the speaker is worked hard makes a very complex interaction electronically and it would take a lot to duplicate otherwise. Finally there is a element of the interaction of the guitar with sound waves hitting the strings and body. The best I have experience with is the Eminence FDM speakers that have a variable magnet position that varies the field strength reducing volume. The FDMs keep the speaker compression (even a slight increase in compression) and the power supply compression but this would be more apparent on stage than in a studio.
But if I was recording tracks in a studio I could throw all this out the window. The best guitar tone is the on that works for the song whether it’s a raging 100 watt JCM 800 and 2-412 stack or direct into the board. There is no right or wrong only what works. I have seen guitarists make some of the worst combos of equipment sound sweet as can be.
For the stage I have never been the loudest guitarist, in fact my bandmate ask me to turn up my volume. I have never relied on working an amp hard to get my tones. I believe that most of my tone is in my hands and regardless of FX used I sound like myself playing (it really made me think he first time I realized it). I favor a clean almost hifi like tube guitar amps that are properly biased and with good clean power supplies. But I do like tube amps both for guitar and my hifi the just are more friendly to my ears especially when the clip (which happens on transients even at medium volume..
I have played with a lot of other guitarists over 35+ years and with out exception they were always much louder than me. One guy in particular that I finally convinced to get a Fender Blues Jr to use instead of his Fender Evil Twin and even reducing the power and spkrs that much did nothing to reduce his volume. I had to resort to ear plugs to play with him even with small amp. The thing is that he an astounding guitarist that can mimic just about any classic rock guitar part with a couple of pedals he’s scary like that. He was worth the effort. IMHO most of these loud guys either have hearing problems or lack critical listening skills. It seems that many players can’t hear themselves regardless of amplifier ability or placement, I believe this may actually have a psychological element to it, a form of stage fright if you will. I was a hifi freak before I started playing guitar seriously and I believe I trained my ears to listen while working at high end stereo shops as both a sales and repair person. We spent a lot of time critically listening to various combos of amps, spkrs and turntables. It really honed my hearing skils.
I think Joe has one done a lot of very good work in creating these audio comparisons. Using a recorded guitar track made for a nice comparison and really showed what the Night Light was all about. I could see this box being useful to many players they seem to have thought about the design and produced a good product. The use of the light bulb is a good technique that has been used in “optical compressors” forever in studios. We used a modified compressor in rental PAs back in the 1980s. When very the systems were turned up loud the harmonic distortion it produced mimicked the sound of overdriven amps. People loved that sound (especially partying drunks) and it protected the spkrs and amps saving repair costs. I also have a system of three switchable light bulbs I put in series with AC line when repairing amps, It limits the AC current if the amp should still have a bad component during testing. The lamps light up instead of the new parts I installed. Light bulb are cool because they change resistance as they warm up small transients can come thru but sustained power causes the lamp to heat up which makes the resistance rise. The initial in rush of current is diminished as the filament warms up and it’s resistance increases. A similar effect occurs in a spkr voice coil which is why an amp sounds best at the end of the set rather than the beginning. One studio tone trick is to have sound playing thru a guitar amp for a while before recording to warm up the voice coil or simply playing multiple takes to warm the voice coil up.
The use of the IR was interesting to hear even tho’ that IR is done with a single frequency and I assume a clean amp setting. If you could do the IR with a real guitar track it would be even better than a single frequency but the computing of that would take a lot of processing power and much more powerful software. It is that reason the modeling amps and FX leave a lot to be desired IMHO. Tube amps will be used for a long time in the future because their sheer complexity of signal interaction between the circuits, spkr and physical interaction with guitar and the sound waves.
But it all comes down to whatever floats your boat. I used to want to design a replacement for loud tube guitar amps but over many years it became clear that it was a tremendously complicated system. It seems sort of ridiculous that the mere handful of parts in a tube guitar amp could be so difficult to duplicate with modern technology but it is just that complexity that continues to keep them at the top of the musical heap.
Sorry I didn’t realize how long that post was, I got a bluetooth keyboard for my iPad for Christmas. You should blame my nephew ;-).
Been using hot plate attenuators for a long time although they are pain because they are fixed Ohms. There may be better choices but these have been reliable and amps have held up so I stay with them. I feel I play better when the volume isn’t a shock to my system. Overly loud fatigues my ears fast and I don’t hear as accurately.
I like the sound of the power tubes cooking when I dial up the guitar volume pot. For that the attenuator has to be after the power tubes. I’m using mostly small amps but there’s little volume difference between a 35 watt and 8 watt if both have an attenuator on it.
I can hear the difference on mine in the higher attenuation settings but I have hard time hearing it once the attenuation is backed off even a little. Its nice too to be able to shift the volume without effecting the sound and dynamics much – its very useful.
This demo was very interesting but the sound was of an amp that doesn’t like being dimed, as you said. It would be very cool to hear the same test with something that is more commonly maxed as well as what effect it has on some cleaner passages – right at the edge where you can apply dirt with the fingers. That’s more where I play most and it sounds pretty good to me with the attenuator.
Joe, what is the Logic Impulse Response test tone? It can’t be just a single frequency (Aidon think).
It’s a 50-second sine wave sweep that goes from subsonic to supersonic. It sounds like the world’s slowest siren.
That 50 second sweep would be much more useful. I imagined it was a shot of noise or a series of stepped tones for the impulse test. The old Sound Forge Impulse software used a single pulse that sounded like pink noise (equal energy per octave) it was better for ambiance than actual frequency domain stuff.
Very interesting comparison. I like your site's approach in general. Like mwseniff, I am a convert from HiFi, and pretty new to guitar and guitar amps. I'm always a bit nonplussed by the lack of precise vocabulary when guitarists talk about amps and tone. The kinds of distortions and where they come from, are rarely addressed with technical vocabulary, as they would by HiFi freaks (HD, and of which kind – spike on the 7th or 9th harmonic? IMD? Onset of speaker nonlinearities? and of which kind? etc).
In your test I'd suspect that the major factor comes from the speaker nonlinearities that start to amplify once the voice coil starts leaving the gap at the true high volume settings. In a guitar speaker that should produce lots of lower harmonics (2nd especially) which makes the sound fatter. If the speaker driver sees the exact same track at lower volumes, the lower voice coil excursion means that the speaker isn't adding much harmonic of its own, just the cone breakup and high end rolloff. So it has got to sound different.
At high volume the speakers seem to compress the signal as well (at least it seems that way to me). There is also the interaction with the strings of the instrument. At some point that turns into feedback but before feedback appears it makes the guitar feel more responsive.
Guitar loudspeakers are known for their non-linearity at high volumes so, as has been mentioned they do compress a little (if the coil moves beyond the magnet gap the magnet field strength falls off) also known for cone break up. I’m not sure that an impulse repsonse would fully capture that. Maybe what is also needed is a touch of soft clipping. And as NotSoFAst points out with a guitar played near an amp at mid to high volumes the overall resonance of the entire system increases until it gets high enough and self sustaining oscillatory feedback occurs. I guess is would be possible to introduce a touch of feedback. I know there is a plug-in – Softube Acoustic Feedback which emulates amplifier feedback. Fender have even built that into a pedal – the Runaway.
I’m guessing that your applied EQ may be taking care of the Feltcher-Munson effect where sound at low volume needs bass and treble boost.
FWIW HotPlates have bass and treble boost switches to compensate for the Feltcher-Munson effect – I don’t care for the sound of them myself. They also have the lamp thing but besides looking kind of cool I haven’t really cared for it, either.
The scientific way Joe approaches these tests is great – I’ll have to try to do the same and really compare the attenuation side by side with the full out. If I can stand the full out! Short clip!
I’ve been accused of a lot of things in my life (or even posthumously), but please! It’s “Fletcher”!
https://en.wikipedia.org/wiki/Harvey_Fletcher
By the way, the Fletcher-Munson effect is a psycho-acoustical effect and only relates to average humans (raise your hand, please) listening to the same sound source at different levels. Since the playback was apparently done at the same level, F-M should not be a factor.
Thanks and good night!
Harvey
Welcome back to this mortal coil, Harv. If I returned after 30+ years, my first activity pro
Okay, since I was in the room when the examples were recorded at very different volumes, I experienced a Fletcher-Munson effect. But since I subsequently normalized them, you, dear readers, did not. Neener neener neener.
I use a Dynacomp pedal as the first pedal in my effects chain with a BBE sonic maximizer in the effects loop on the mixer. I turn the knobs on the Dynacomp to 3 o clock position. At lower volumes it seems to maintain the boost and I can fine tune the tones with the BBESM. Works well for small venues and through PA systems when the amp is not mic'd. I try to give the sound man as much control as possible by keeping the noise from stage to a minimum. I make adjustments in the tone with the guitar, expression and volume pedals on stage. Keeps me from chasing the crowd away with harsh tones that get out of hand. Still no substitute for a good sound man. I've taught my kids to run sound for me. They work cheaper.
I had a Weber Mini Mass which uses a speaker motor for the dummy load.
I believe that as attenuators go the Weber is good but I decided that I didn’t need it as I’m satisfied with the tone at any volume from my amps and pedals…Although my loudest amp is just 22 watts.
That said, I only know of one way to get the cranked tone of my Deluxe Reverb thru an alnico Blue Dog… I’ve never really heard that sound conveyed through any recording but I have stood in front of it for sure.
I've had some experience with Attenuators,back in the day,I had a Power Soak from Tom Scholz,worked quite for my needs. I also remember another one I used,but I can't remember the name,but the box was kind of tan colored and it had red printing on it. The nicest one that I've found today,is made by Palmer,you can vary the impedance ,and it will handle 100watts. The price was under 200 bucks,and the tone was near original as possible. I would think just about anyone would be pleased with it. As Robin Stonestreet mentioned the Sonic Maximizer would help to replace any tone that you think you might have lost