DIY Loudspeakers ?

[RJohan]

My current brain child is a hybrid. Starting with my KEF R100 and adding bass drivers to make them into something like poor mans Blades. Right now I'm focusing on two SEAS L26 ROY's per side, divided at something like 200 Hz. I will start by using the built in active crossover in my current HT receiver and gradually turning it into a complete, passive speaker. If it doesn't turn into anything good I can allways use the SEAS drivers as subs.

 

[h.g.]

You really can't fault the drivers, but the crossover is far, far more important - and incredibly difficult to get right (especially so when passive). It might be me but I can't see any measurements of this speaker.

Measurements are in Klang & Ton and possibly from some of the builders.

You only have to Google Stereophile measurements to see that a lot of well respected manufacturers can't achieve a flat frequency response. Yes, a flat frequency response doesn't guarantee a perfect speaker, but a wonky one is knackered from the start, and will sound how it measures (wonky).

That is not wholly true because we hear a combination of direct and indirect sound. Around a crossover on a flat baffle the larger driver will have a narrow beam width and the smaller a wide beam width. To compensate a non-flat frequency response may be perceived as more tonally neutral. B&W take this to an extreme degree when crossing a large midrange at a high frequency to a non-baffled tweeter as can be seen in the Stereophile measurements despite the 3D directivity plots being nigh on impossible to read values from.

Of course controlling directivity so there is no need to compensate for sharp changes in beam width is preferable if the objective is a neutral presentation. However the "sound effect" of a sharp increase in beam width at higher frequencies can give the impression of more detail and so may be considered a positive attribute by some speaker manufacturers. Maximising a neutral sound clearly isn't the prime objective for most home audio speaker manufacturers. A few audiophiles may seek a neutral sound but I suspect most seek something that sounds good to them. This seems perfectly reasonable given the prevalence of record players, valve amplifiers, etc... which also do not seek to provide a neutral presentation.

A flat frequency response to within a small fraction of a dB is achievable at modest cost if a manufacturer gears up for it (e.g. preaching tolerances section at bottom). It requires drivers built to tight tolerances, an active crossover, measurement and setting the active crossover parameters for each individual speaker. It also allows the response to be recovered as the drivers drift out of spec with age. Audiophile interest in this sort of "real" technical performance is low. It simply isn't what the hobby is about for most.

 

[Fatmarley]

Measurements are in Klang & Ton and possibly from some of the builders.



That is not wholly true because we hear a combination of direct and indirect sound. Around a crossover on a flat baffle the larger driver will have a narrow beam width and the smaller a wide beam width. To compensate a non-flat frequency response may be perceived as more tonally neutral. B&W take this to an extreme degree when crossing a large midrange at a high frequency to a non-baffled tweeter as can be seen in the Stereophile measurements despite the 3D directivity plots being nigh on impossible to read values from.

Of course controlling directivity so there is no need to compensate for sharp changes in beam width is preferable if the objective is a neutral presentation. However the "sound effect" of a sharp increase in beam width at higher frequencies can give the impression of more detail and so may be considered a positive attribute by some speaker manufacturers. Maximising a neutral sound clearly isn't the prime objective for most home audio speaker manufacturers. A few audiophiles may seek a neutral sound but I suspect most seek something that sounds good to them. This seems perfectly reasonable given the prevalence of record players, valve amplifiers, etc... which also do not seek to provide a neutral presentation.

A flat frequency response to within a small fraction of a dB is achievable at modest cost if a manufacturer gears up for it (e.g. preaching tolerances section at bottom). It requires drivers built to tight tolerances, an active crossover, measurement and setting the active crossover parameters for each individual speaker. It also allows the response to be recovered as the drivers drift out of spec with age. Audiophile interest in this sort of "real" technical performance is low. It simply isn't what the hobby is about for most.

Yes I agree, in some situations a small dip in the right place will help with a directivity mismatch, but some of the horror stories I've seen in the Stereophile measurements, by respected manufacturers are nothing to do with that.

 

[h.g.]

Yes I agree, in some situations a small dip in the right place will help with a directivity mismatch, but some of the horror stories I've seen in the Stereophile measurements, by respected manufacturers are nothing to do with that.

What do you think are likely reasons for strong departure from a flat response? Is it that they lack technical know-how compared to DIY hobbyists? Perhaps they would like to design speakers with a flat response but allocate almost no resources to it compared to more important activities such as production, marketing and such? Or might the deviations from a flat response be deliberate? If deliberate might the objectives for the speaker be well met and so not a horror story?

 

[Fatmarley]

What do you think are likely reasons for strong departure from a flat response? Is it that they lack technical know-how compared to DIY hobbyists? Perhaps they would like to design speakers with a flat response but allocate almost no resources to it compared to more important activities such as production, marketing and such? Or might the deviations from a flat response be deliberate? If deliberate might the objectives for the speaker be well met and so not a horror story?

Sometimes I'd imagine a ramped up bass or treble is for showroom appeal. If it's a peak in a place that's only ever going to sound nasty, then I can only think it's a f*ck up.

With old, established brands it could be that the real talent has either left, retired or died, or the company has been taken over, or just the company name bought. I'm not saying that diy hobbyists are guaranteed to produce a flat frequency response either, although the knowledge is out there to do that.

 

[h.g.]

Sometimes I'd imagine a ramped up bass or treble is for showroom appeal.

Is it only showroom appeal or might it be sustained appeal? For example, a tiny speaker like an LS3/5a has little bass extension and so when it was designed the level above roll-off was raised in order to give an increased impression of bass. Is this the correct response rather than a flat response (assuming no subwoofer)? 10-15 years ago I was waiting in an Apple store and was surprised to see tiny and relatively expensive Adams speakers on sale among the usual flashy looking computer speakers (the store was about 2 miles away from the factory which possibly explained the presence). Anyway, while waiting I asked a salesman what he thought of them and he said not much and demonstrated his preference for some plastic computer speakers at half the price. The Adams had a flat response with little bass extension and at low volumes in a large showroom the lack of a bass bump drew attention to the missing bass.

With old, established brands it could be that the real talent has either left, retired or died, or the company has been taken over, or just the company name bought.

I think you will find that modern engineers tend to be significantly better informed than older engineers as new technology has come along: detailed 3D computer simulations, cheaper more flexible measurements, new materials, active crossovers, DSP, etc... The larger speaker companies tend to be a bit better equipped and more knowledgeable than we DIYers but the same steps forward have been made in the DIY world and, indeed, much of the older DIY generation is slow to move with the times.

I'm not saying that diy hobbyists are guaranteed to produce a flat frequency response either, although the knowledge is out there to do that.

Yes with DSP it is a doddle to produce a ruler flat measured frequency response and linear phase response for speakers under anechoic conditions. Some DIYers happily do this but it is rarely going to be optimum in use due to off-axis response, in room behaviour, possible DSP issues like pre-ringing, stereo vs binaural perception issues, etc... Of course one could make a case for a flat on-axis speaker response with all the compensation in a separate DSP stage but home audio playback is rarely configured accordingly.

 

[Fatmarley]

Is it only showroom appeal or might it be sustained appeal? For example, a tiny speaker like an LS3/5a has little bass extension and so when it was designed the level above roll-off was raised in order to give an increased impression of bass. Is this the correct response rather than a flat response (assuming no subwoofer)?

That's a trick I've used myself with speakers that had limited low frequency extension. I think it's fair to say a hump to compensate for limited low frequency extension is common knowledge.

I think you will find that modern engineers tend to be significantly better informed than older engineers as new technology has come along: detailed 3D computer simulations, cheaper more flexible measurements, new materials, active crossovers, DSP, etc... The larger speaker companies tend to be a bit better equipped and more knowledgeable than we DIYers but the same steps forward have been made in the DIY world and, indeed, much of the older DIY generation is slow to move with the times.

That's not what Stereophile and other measurements are telling me (I'm talking about passive speakers though).

 

[bhazen]

Assembling a kit speaker can be a way to get a speaker that approaches something 2-3X the price. Wilmslow Audio in the UK has reverse-engineered a number of great-sounding speakers by famous names (I won't say who, but look at WA's Facebook page. The inspirations are obvious!) and offers close approximations of those designs.

I'm thinking of having a go with their LS3 speaker kit, which is an LS3/5A-inspired speaker for less than 1/3rd the price of what licensed LS3/5A's go for today.