Using a graphic equaliser analyser to improve room and speaker anomalies

[stevehertz]

I have long been interested in the subject of using a graphic equaliser to correct for room and speaker anomalies. The main problem faced with the process is knowing how to adjust the sliders to obtain best results. Fortunately some of the higher profile Japanese hifi manufacturers made units that incorporated analysers that allowed this to be done much more accurately than simply relying on one’s ‘ear’. I recently bought a Technics SH-8055 graphic equaliser that is such a beast. It has 12 frequency sliders per channel and for analysis purposes Technics made the matching RP-3800E microphone that has an ultra flat frequency response to use with it. As you would suspect, the RP-3800E is as rare as rocking horse droppings or hen’s teeth and commands a price to match when you do come across one, there’s one for sale now at over £300. Fortunately I have a Behringer ECM8000 condenser microphone that also has an ultra flat frequency response and is essentially the same spec and the same thing.

The SH-8055 has a pink noise source and it’s a relatively simple process to initiate the pink noise and adjust the sliders on each channel in turn to achieve a near flat response for each channel. Obviously the room has a considerable effect on the settings but otherwise the process is quite simple. I experimented with and analysed the positioning of the mic and decided to position it in an ‘average’ position that aided in nulling speaker and room anomalies but accounted for the fact that not everyone would be sitting in the perfect stereo listening position. It’s quite enlightening to see the position of the sliders required to obtain a flattish response. Pink noise itself is not a ‘perfect’ source for this exercise, and the graphic display jumps around a tad at most frequencies; you cannot arrive at a perfectly flat display. But nonetheless it’s possible to obtain a pretty flat response from 40Hz to 16Khz. I purposely kept 25HZ low down as not much material is present there and it’s basically a portal for other nasties such as rumble to enter the signal chain and potentially damage the speakers as well as muddying up the amp.

Having done that, I began listening tests with the graphic on and off. My speakers are the highly rated, multi award winning ATC SCM11 units. The difference with the graphic on and off is quite pronounced. With the graphic in circuit the sound is much more natural and balanced – as you would expect. I’m not going to venture into audiophool territory with lucid, airy fairy descriptions but as an experienced musician, small studio owner/builder and erstwhile builder of high quality speakers I’m very pleased with the results. Also, the graphic has distortion and noise level figures that make its addition to the signal chain inaudible apart from its effect on frequency, but I wouldn’t expect an audiophool to accept that.

Photos L to R: The SH-8055, graphic display when playing music, graphic display when using pink noise with mic and analyser, and lastly the slider settings to achieve an improved speaker/room response.

[Electronpusher0]

Thanks for the report, an interesting read.
A surround sound amp we used to use for watching films had a microphone supplied and had an auto balance function to balance the levels of the 5 channels for best levels.
I am not sure if it also adjusted the tone levels but it made setting up very easy.

I had wondered about plotting the response of my setup & room but the price of ultra flat mics put me off. However I have checked the price of the Behringer one you use and surprisingly its only £30. I may have to get one.

Peter

[stevehertz]

Originally Posted by Electronpusher0

Thanks for the report, an interesting read.
A surround sound amp we used to use for watching films had a microphone supplied and had an auto balance function to balance the levels of the 5 channels for best levels.
I am not sure if it also adjusted the tone levels but it made setting up very easy.

I had wondered about plotting the response of my setup & room but the price of ultra flat mics put me off. However I have checked the price of the Behringer one you use and surprisingly its only £30. I may have to get one.

Peter

Yes, the Behringer mic is very nicely made and comes in a presentation case with stand mount and pop cover. How will you plot the room response?

[PJL]

Standing waves formed in rooms means you get peaks and troughs that move with frequency. Converting your room to an oversized padded cell is the only solution.

[stevehertz]

Originally Posted by PJL

Standing waves formed in rooms means you get peaks and troughs that move with frequency. Converting your room to an oversized padded cell is the only solution.

I believe I have a solution that suits my needs, I'm not looking to create an anechoic chamber, just an improvement, and I know I've achieved that. Particularly, I don't generally listen to music sitting in the perfect stereo position (as one would have to to hear a perfectly flat response), and that's why I experimented with and analysed the mic position to give me an improved frequency response in a broader range of listening positions. The phrase "only solution" suggests a perfect solution, I've been in electronics and audio/music long enough to know not to expect that, and I never set out to achieve it.

[Buzby123]

I came across this bit of software in a post in this forum, somewhere.

It is a sort of spectrum analyser which works with a modified pink noise source, and is especially designed for room acoustics.


https://www.libinst.com/SynRTA.htm

[Radio Wrangler]

Resonances in rooms can have quite high Q factors, while limits on the numbers of bands in an equaliser mean that the ability to shape the driving function may not be able to match the shape. Given a very large number of bands, multiple bands can be proportioned to synthesise the shape, but that's impractical in an analogue equalliser.

Parametric equalisers can handle only a limited number of resonances, suck-outs or slopes, but the frequency and Q can be freely adjusted and can give a better fit.

Transversal filter type equalisers implemented in DSP can throw a lot of taps at the problem and equalise quite complex responses, but sometimes the issue of overall time delay can rule out some uses.

Nothing's perfect, but if you've got a net improvement, then you're on a winner.

David

[paulsherwin]

Originally Posted by Radio Wrangler

Transversal filter type equalisers implemented in DSP can throw a lot of taps at the problem and equalise quite complex responses

It's becoming increasingly common for professional sound reinforcement systems to use these techniques, even at the comparatively low end. Bose make a PA targeted at the live music / karaoke / mobile disco market that does an amazing job of coping with typical dreadful pub acoustics. It's relatively expensive like all Bose kit, but not ridiculously so. I heard an example a couple of years ago and it basically sounded like a decent HiFi system in a domestic lounge, rather than the usual reverberant squawky mess.

in a post in this forum

My post about pre amp graphics, the software is quite good.

[ToneArm]

Gosh, I drooled over equalisers like this as a teenager in a 1980s hifi show. It hadn't dawned on me that they might now be affordable 2nd hand bargains..

Interesting to see the sliders for left and right are not in the same positions, does the adjustment improve the stereo imaging?

Cheers

[stevehertz]

Originally Posted by ToneArm

Interesting to see the sliders for left and right are not in the same positions, does the adjustment improve the stereo imaging?

Cheers

Can't say that I've 'looked' to see if it has. Saying that, due to the improved frequency response, individual instruments are tonally more 'as they should be', so yes, I guess you could say that the stereo imaging is improved.

[Bazz4CQJ]

I'm sure they are frowned up in some quarters, but I too revel in the delights of my Technics equaliser - this time a SH-GE70, shown sitting below its sibling SL-PG200A CD player. I know the CD player was bought new, but I have no recollection of buying the equaliser.

B

[stevehertz]

Originally Posted by Bazz4CQJ

I'm sure they are frowned up in some quarters, but I too revel in the delights of my Technics equaliser - this time a SH-GE70, shown sitting below its sibling SL-PG200A CD player. I know the CD player was bought new, but I have no recollection of buying the equaliser.

B

Oh for sure they are very frowned upon by the greater hifi community, but these are the same people who don't understand that what was heard in the studio is normally nothing like the sound they are hearing in their homes, and they refuse to use even basic tone controls to correct for bass and/or treble deficiencies as a result thereof. Basically the minimalist straight wire with gain belief is much important to them than a dull, bright, or boomy listening experience. Each to their own!

[paulsherwin]

The main concern seems to be that the complex analogue frequency shaping introduces phase shifts. I've no idea if there's any basis to this.

Sound perception is extremely subjective, and what matters is that it sounds better to you. I find it interesting that the pink noise flat response setup is not dissimilar to the traditional 'loudness' compensation available on many amps.

[Radio Wrangler]

Things like room resonances also introduce phase shifts, so if the correction is done by means of an overall pole-zero cancellation, then it comes out in the wash. Delayed reflections can introduce frequency response shaping with different phase shift to that of a classical pole-zero constellation model and this is where things go wrong.

However, the transversal equaliser approach has independent control of magnitude and phase so a transversal equaliser can fill the bill. The independence of phase and amplitude response allows the coefficients of an FIR transversal filter to be fiddled to produce a controllable phase shift do a pair with 90 degrees between them and you have a major part of phasing method SSB. Also the coefficient fiddle turns out to be the hilbert transform when you pick 90 degrees.

So there's a lot of prior art about playing with phase and amplitude independently.

David

[GrimJosef]

There are modern DSP units which can do a similar thing - this one https://www.minidsp.com/products/min...minidsp-2x4-hd hooks up to your PC and the same company will sell you a calibrated mic. The accompanying software uses the PC's sound card to deliver a frequency sweep through your hifi and it 'listens' to this with the mic. It works out a correction profile which it loads into the little DSP box and you can then disconnect the mic and PC and see if you like the new sound better than the one you had before.

The box's other function is as a low-level crossover, allowing you to do away with the lossy, imperfect crossovers in your speakers and connect your amplifiers straight to the drivers (of course this means you need more amps).

Cheers,

GJ

[Radio Wrangler]

One small Achtung! about sweeping hifi systems.

When you sweep a flat tone through the audio band at constant power, beware of the power rating of tweeters. Music and speech have a significant reduction in power with increasing frequency. Your bass drivers may be good for several tens of watts input, but your tweeters will only handle a few watts. It's awfully easy to burn out tweeters when testing speakers/rooms.

People run into this problem when using hifi speakers for synthesisers and guitars

David

[stevehertz]

Originally Posted by Radio Wrangler

One small Achtung! about sweeping hifi systems.

When you sweep a flat tone through the audio band at constant power, beware of the power rating of tweeters. Music and speech have a significant reduction in power with increasing frequency. Your bass drivers may be good for several tens of watts input, but your tweeters will only handle a few watts. It's awfully easy to burn out tweeters when testing speakers/rooms.

People run into this problem when using hifi speakers for synthesisers and guitars

David

And also when rewinding tape machines that don't (or choosing not to) lift the tape from the playback heads. Sizzled tweeters.

[m0cemdave]

Whilst Andy has achieved a good result (i.e. something that works and he is happy with), I would offer some words of caution to others who may be contemplating something similar.

I think Andy's 40Hz boost is making up for the small size of the SCM11's and their consequent natural bass roll-off. How about making a couple of subs using the ELF technique described by Rod Elliott on his website?

In general, trying to correct the effects of room acoustics with equalisation is frowned upon. It's really just a quick and dirty fix for situations where there isn't the time, or money, or facilities, to do it by acoustic means. One night stands in dodgy sounding concert halls are a prime example.

Equalising the stereo channels differently will result in some differential phase shifts, which will affect the stereo imaging at some frequencies. This may or not be noticeable amongst the probably greater effects of room reflections.

There are some simple techniques for improving the sound of a room that can help. A few strategically placed panels of absorption, hung like small tapestries on the side walls, can make a surprisingly big difference without wrecking the appearance of a living room. They work by reducing the early reflections that cause noticeable coloration.

If it's possible to hide a couple of rolls of wadding in the room's rear corners behind suitable cloth material (say 2ft wide and full room height, cutting off the corners in a triangular plan view), it will help reduce bass boominess.

Bookshelves can help a lot too. Lining the books up to the front of the shelves to give random shaped cavities behind them is useful.

Anything that helps to scatter sound behind the listening area is also a good idea, shelves of bric-a-brac or books are much easier to deal with than purpose built diffractors / diffusors...

Hoping this is useful, and please don't misconstrue it as pouring cold water on Andy's efforts.

By the way, that 8055 is a nice unit. Essentially an octave equaliser, but with finer control at low frequencies, where it's needed.

[Trevor]

About 50 years ago I was experimenting with linear phase speakers i.e. stepped units I set up a good quality condenser microphone with a scope
standing 2 meters away i noted a distinct change in a 1 khz square wave when i moved my head left or right
So is phase so important ?
Trev