speakers for classical music

When you've all the strings going, the brass playing at it's most brilliant and the percussion pounding away, many systems lose track of the character of each instrument and instead present a homogenized, loud, stressful mess.

That would be typical midrange compression (heat and non-linearities at high excursions).

Soundstage is the first reviewer to start measuring compression. Since 2006, they include a chart comparing output at 70 db SPL with output at 90 db SPL and higher levels. The difference is thermal compression - so you can see Watt Puppy 8's tweeter starting to compress at 95 db SPL on Chart 4. Whilst Revel F12 seems pretty good at 95 db SPL but starts to look dodgy at an albeit very impressive (for a consumer speaker) 100 db SPL (although a glance at Chart 3 further up shows that distortion is starting to look scary at 95 db SPL)!

Soundstage state "Very few speakers can be tested at 100dB without damage"....so if you were wondering where is that all important 10 db SPL headroom...it ain't there in most designs playing at 90 db SPL (at listening position), as you are already stressing the speaker.

This paper explains venting and the importance of voice coil diameter in a driver and the choice of magnet. The JBL 2225H and EVX 150 with 4 inch voice coil takes 20 seconds to compress whereas a small 3 inch voice coil compresses within seconds. (Note that some subwoofers and many consumer woofers use a mere 2 INCH Voice coil or less - and these get hot really quickly and have trouble to get rid of heat)

The highly respected TAD driver with 3 inch Voice coil uses Alnico which loses magnet strength with heat and also compresses within seconds.

There is so much more to speakers than veneer/cabinetry. Almost nobody asks the salesman about the woofer voice coil size but, as it turns out, it is a highly relevant question that will govern how the speaker performs.

BTW - if you get 3 or more db of compression then you can bet that the whole crossover design is completely inappropriate at this playback level as the voice coil resistance has probably more than doubled in the drivers.

What does it sound like - "a homogenized, loud, stressful mess." as Dave so eloquently put it!

I hope the last post I made above was at least useful to one or two people or perhaps it was another case of my misguided efforts to explain something that is largely falling on "deaf ears".

They say "ignorance is bliss" and it may be so for the many people with non pro drivers.

A friend of mine once told me "I'd really rather not know why my speaker sounds awful. Besides, I prefer to think that I can compensate for things through spending more on a warmer/incredibly powerful amp, source and better cables..."

I have not heard Duke's speakers but if his design philosophy and choice of drivers is any indication then they are definitely worth investigating, as Atmasphere suggested.

Sadly however, I suspect this particular area of system performance will remain fundamentally one of "ignorance is bliss". The majority of speaker manufacturers are forced to use cheap mass produced drivers to compete in a tough market with hard to beat price points - besides it is much more important what the drive looks like on the outside than the size of the motor - so I don't think you'll hear much about this topic except on DIY forums and in pro audio.

Harbeth C7ESIII or the Super HL5

Shadorne, thanks for you links and comments. I read it and just didn't think a further comment was needed. Now you've prompted me to ask:

How long is the typical musical peak? If I'm listening at 87dB on average and I have peaks of 105dB, which I have observed and note seem appropriate for the music, if it takes a few seconds for compression to begin, will I experience it often, if ever, in music listening???

BTW, I play electric guitar, so I've actually purposely compressed guitar alnico drivers to get a certain tone quality only available from a compressed speaker (or a good speaker emulator). So, I understand the concept pretty clearly. I'm only wondering how much of a real world factor it is with high quality, high fidelity speakers.

Also, the compression shown on the graphs was actually very small in comparison to the in-room EQ variations of the speakers shown. I'm wondering how much we'll hear those fractional dB losses when taken with all the other interactions.

Don't take this as anti-horn. I plan to investigate Audiokenesis at RMAF, as well as other high efficiency speaker designs. I'm just trying to understand some of the real world implications.

Dave

Hi Atmasphere, you said, "This recording puts most speaker/amp combinations right flat on their face nearly right away, if you are unwilling to change the volume as it plays.'

Could you explaine to me what you mean by this. Are you saying you need to turn up/down the volume for this recording due to its dynamics within the recording?

One volume for all!
Bob