Loudspeaker compression (why a 200+ WPC amp doesn't help KHorns)

[stereohermit]

109 db at failure threshold.

[stereohermit]

OOPS, I lied. Actually its 114 db w/ 4 % distortion

[stereohermit]

Here's the left side of the graph. It's the center spread of the catalog, not able to be scanned at once.

[garyrc]

Islander:

• It looks like the failure threshold for the Belle is 114 dB at the distance they used, with their test enviornment, with their (unspecified) two tones. I assume the tones were continuous. The 120 dB (Belle) and 121 dB (Khorn) figures we often see must be 1) Peaks? 2) Much closer up? I wish there were many more specifics stated in the report. PWK used to claim that with a Belle or La Scala as a center channel, with 2 Klipschorns as front right and left, 20 watts RMS power into each would produce a combined level of 115 dB peaks "at your ears." I have produced this with 60 wts into each (from listening position) with more probably available, but I didn't dare, because I don't have much faith in my meters.
• Sorry about the big type (12). I will reform. I often slide into my computer chair without the location of my glasses known, and the 12 type helps me read what I'm writing, then I often forget to reduce the size before pasting.

[Oblio]

Sorry about the big type (12). I will reform. I often slide into my computer chair without the location of my glasses known, and the 12 type helps me read what I'm writing, then I often forget to reduce the size before pasting.

Gary, if you are using Firefox you can increase the size of the print AND images by using 'Ctrl +' or 'Ctrl mouse-wheel' without affecting the font used to post. IE might have a similar feature. I use this when I've misplaced my glasses.

[quadklipsh]

hi

as i ran across ur post i always have fancied khorns,but never heard or saw em live

do u have other klipsch speakers esp reference line ones

if yes how is ur experience and wots ur verdict

im having a RF 3 pair

and struggling to get the best sound out of em

a warm,mellowy and dreamlike

is that possible at all in this digital world?

plz comment

thanks

dr shaharyar

[stereohermit]

The max sound pressure quote for example 121 db for the Klipschorn and La Scala is at 1 meter on axis.

Dont confuse this spec with ones averaged into a reverberant space for a pair.

It would be hard to compare the two...complex variables with room averaging...how reflective, what distances, etc.

Islander:

• It looks like the failure threshold for the Belle is 114 dB at the distance they used, with their test enviornment, with their (unspecified) two tones. I assume the tones were continuous. The 120 dB (Belle) and 121 dB (Khorn) figures we often see must be 1) Peaks? 2) Much closer up? I wish there were many more specifics stated in the report. PWK used to claim that with a Belle or La Scala as a center channel, with 2 Klipschorns as front right and left, 20 watts RMS power into each would produce a combined level of 115 dB peaks "at your ears." I have produced this with 60 wts into each (from listening position) with more probably available, but I didn't dare, because I don't have much faith in my meters.
• Sorry about the big type (12). I will reform. I often slide into my computer chair without the location of my glasses known, and the 12 type helps me read what I'm writing, then I often forget to reduce the size before pasting.

[stereohermit]

Haven't heard the RF3's, but I relate to the type of sound you referred to. It took me almost 20 years to get it out of my La Scalas.

Klipsch's tend toward dry, crisp, clean 3rd-5th row sound.

For a more mid to rear hall sound it take quite a bit of modding, but I'm pleased to report it is possible.

[Islander]

Islander:

• It looks like the failure threshold for the Belle is 114 dB at the distance they used, with their test enviornment, with their (unspecified) two tones. I assume the tones were continuous. The 120 dB (Belle) and 121 dB (Khorn) figures we often see must be 1) Peaks? 2) Much closer up? I wish there were many more specifics stated in the report. PWK used to claim that with a Belle or La Scala as a center channel, with 2 Klipschorns as front right and left, 20 watts RMS power into each would produce a combined level of 115 dB peaks "at your ears." I have produced this with 60 wts into each (from listening position) with more probably available, but I didn't dare, because I don't have much faith in my meters.
• Sorry about the big type (12). I will reform. I often slide into my computer chair without the location of my glasses known, and the 12 type helps me read what I'm writing, then I often forget to reduce the size before pasting.
  
  Thanks for the info, Gary. It's likely that test tones would destroy a speaker at a volume level that might be safe with music, so maybe the higher numbers are with music, but that's just a guess. At any rate, 114dB is pretty loud already. I've had my system up to 110dB once or twice, but only for very short periods, like one song or less. More typical listening for me is in the 80-95dB range.
  
  Also, thanks for the regular-size type.

[Islander]

im having a RF 3 pair

and struggling to get the best sound out of em

a warm,mellowy and dreamlike

is that possible at all in this digital world?

Your source could be a factor. Have you read about the "loudness wars"? Lots of CDs produced in the last five years were mixed for max volume, but they're overloaded to the point of very frequent or even almost continuous clipping, making for a very harsh sound.

[garyrc]

Am I understanding the discrepancies correctly if I integrate the info in the following way? I'd be interested in what any Klipsch staffer who happens to be reading this thinks, as well as the forum people.

Sources: 1) Keele's Dope From Hope memo of January 1977, as revised in November 1980 (uncredited) in Dope From Hope Vol 16, No 1, "Amplifier Ratings to drive Klipsch Systems," 2) "Klipsch is Music" an undated brochure (probably in the ' 80s) containing the report of Klipsch's dynamic range tests, taking the Belle Klipsch and 4 other (unnamed) speakers "from leading manufacturers" through various SPLs, noting distortion, until each speaker failed. 3) Current Klipsch advertisements (c2008-2009).

Two important disclaimers:

• Sound Pressure Level Meters that use a needle will read about 13 dB low on instantaneous peaks, due to needle ballistics (this according to PWK somewhere in the Dope From Hope collection), even on "fast.". I don't know if some other meters that use other devices -- like LEDs -- are adjusted to mimic the old standard needle meters.
• As several people have pointed out, the room, number of speakers/channels operating, distance, and liveness make a big difference. I guess it would be safest (although not necessarily the most accurate) to think of the SPLs below as maximums for the wattage given, in a relatively live room, relatively close, and with one speaker operating (only because the reports used cited one speaker, always a Klipschorn or Belle Klipsch, or La Scala for the data I'm including below).

For the Klipschorn, Belle, and La Scala:

Maximum peak output with ever varying music (this would be for instantaneous peaks only, and would exclude music that involves steady tones at maximum volume, conceivable with some electronic music, or sustained electric guitar):

120 / 121 dB, requiring, in a 3,000 cu. ft. room, 200 watts for the fraction of a second of the peak, producible by some 20 watt amplifiers, without clipping, according to both Keele and his reviser. I have not seen that kind of margin between rated power and peaks anywhere else, so I assume that it holds only for the briefest possible peaks. Also, if the one set of trials that we saw reported on the forum a few years ago can be trusted, some tube amps can put out 6.25 times as much power on peaks only as some solid state amps with the same power rating RMS, so if the information Klipsch was using (10 dB headroom on peaks without clipping) came from tube amp theory, those of us who use solid state might need something like a 125 watt amp to produce that instantaneous 200 watts and 120 / 121 dB SPL without clipping.

Failure Threshold for the (steady?) tones Klipsch happened to use in the research leading up to the "Measuring Dynamic Range" section of "Klipsch is Music" in an anechoic environment:

114 dB @ 165 watts input

It should be noted that the long term maximum power cited in Keele's original article was 75 watts, and it was changed to 100 watts in the revised version.

Average (ever varying, not continuous) Music Levels in the 3,000 cu.ft. room Keele cited (excerpted, including characterizations, from the original document) for a single speaker:

90 dB requiring a good 0.20 watt amplifier -- "Loud"

105 dB requiring a good 6.3 watt amplifier -- "Very Loud"

115 dB requiring a good 63 watt amplifier -- "Too Damn Loud"

120 dB requiring a good 200 watt amplifier -- "Too Damn Loud"

These amplifier ratings would be needed to produce the average levels shown. As discussed above, if all one wants is a peak level of 120 dB, a smaller amplifier could be used, because the 200 watts would only be needed for a fraction of a second.

Note that the tests of failure level of the Belle in "Klipsch is Music" were done in a "free field anechoic environment." I find it plausible that what people describe as room gain or reflections, or the simultaneous existence of newly generated sound superimposed on decaying sound could account for ever varying music reaching 120 dB being producible with 200 watts in a typical 3,000 cu. ft. room, but that 114 dB steady tones in an anechoic chamber would require 165 continuous watts, and would blow the speaker away. Nonetheless, I am uncomfortable with the 120 dB/200 watt line being cited as one of the possible average levels in the Keele/reviser chart..

What about normal efficiency speakers, typically 90dB @ 1w @ 1M? Because the most conservative rating I have ever seen for the Klipschorn was 98 dB @ 1w @ 1M (not in a corner or against a wall) it seems to me that if we take 115 dB in 3,000 cu ft as a requirement, the typical speaker would have to make up that 8 dB difference (double the wattage for every 3 dB increase) by receiving about 420 watts, instead of the 63 watts needed by the Klipschorn, La Scala, or Belle. If a Klipschorn sealed into a corner (in Klipsch's wooden trihedral corner containing anechoic chamber) really produces 105 dB @ 1 watt @ 1 M, then it seems to me that the typical 90 dB/w/M (in anechoic) speaker would need 2,016 watts to equal the 115 dB figure.

[DrWho]

...

Music Levels in the 3,000 cu.ft. room Keele cited (excerpted, including characterizations, from the original document) for a single speaker:

90 dB requiring a good 0.20 watt amplifier -- "Loud"

105 dB requiring a good 6.3 watt amplifier -- "Very Loud"

115 dB requiring a good 63 watt amplifier -- "Too *** Loud"

120 dB requiring a good 200 watt amplifier -- "Too *** Loud"

These amplifier ratings would be needed to produce the average levels shown. As discussed above, if all one wants is a peak level of 120 dB, a smaller amplifier could be used, because the 200 watts would only be needed for a fraction of a second.

...

That is assuming that the lower power amplifier is able produce the 200W for a short period of time...that's not always the case.

As you mentioned earlier, solid state amps (especially Class D) will tend to have lower peak output ratings versus their maximum continuous ratings. However, this gets a bit more complicated because how accurately are those peaks being reproduced in the amplifier? I think it is more reliable to choose the amplifier power such that its maximum continuous output meets the peak power demands for the application. However, that is not to say that a lower continuous power amplifier can't also accurately reproduce transients...it just becomes very difficult to ensure accuracy.

Basically, the problem is that the "power rating" of an amplifier doesn't fully describe its behavior. Likewise, the power rating of a speaker doesn't fully describe its behavior either. Sadly, the transient nature of dynamic recordings is such that people are frequently running the amps and speakers into their nonlinear range. I truly believe the majority of difference in sound between various well-behaved electronics really comes down to differences in noise floor and clipping behavior. It also seems like horns are the only speakers capable of actually reproducing transients without clipping - even when listening at very moderate levels.

[Islander]

Basically, the problem is that the "power rating" of an amplifier doesn't fully describe its behavior. Likewise, the power rating of a speaker doesn't fully describe its behavior either. Sadly, the transient nature of dynamic recordings is such that people are frequently running the amps and speakers into their nonlinear range. I

This would indicate, then, to sum it up, in non-technical language, taking all factors into account, for most realistic production of transients and most convincing reproduction of music's dynamics, all else being equal: "MORE POWER = MORE BETTER"

Assuming good clean power, of course.

[garyrc]

Islander wrote:

"This would indicate, then, to sum it up, in non-technical language, taking all factors into account, for most realistic production of transients and most convincing reproduction of music's dynamics, all else being equal: "MORE POWER = MORE BETTER"

Assuming good clean power, of course."

I agree, but if you have a speaker that requires 800 or 1,000 wats to produce a very loud transient, and its power handling capacity is average, you're in the danger zone. There are some speakers around like that. So I would say "

MORE POWER

COMBINED WITH HIGH EFFICIENCY = MORE BETTER"

[DrWho]

Assuming good clean power, of course.

That's a pretty big assumption....

In order to get more good clean power, the dynamic range of the amplifier needs to increase so that the low level details don't get affected. With truly high efficiency speakers, it's actually quite difficult to get the amplifier noise floor sufficiently below the noise floor of a quiet room....and adding more gain to the amplifier is just going to make it worse.

Just enough power allows one to maximize low level detail without the drawbacks of clipping.

[DrWho]

Ok, since I'm a nerd...

The D75A is one of the better solid state amps I know of in terms of dynamic range, which is 106dB unweighted (20Hz to 20kHz):

http://www.crownaudio.com/pdf/amps/136711.pdf

It has 16 dBW of gain into an 8ohm load, which means the minimum noise
floor at 1m would be the sensitivity of your speaker + 16dBW - 106dB =
90dB below the sensitivity of your speakers.

If your room was
quiet, say 30dBspl, then the maximum sensitivity you'd want from an
8ohm speaker would be 10dBspl (20dB below the noise floor of the room
for complete masking) + 90dB = 100dBspl. A 100dB sensitivity 8ohm
speaker would put your max SPL to 116dBspl, and thus your maximum
unclipped music listening with 20dB crest would be 96dBspl average. Of
course there's some play there depending on how well you like the sound
of your amplifier's noise floor. With a 110dB sensitivity 8ohm speaker,
the noise floor would be 20dBspl (110dB - 90dB) and only 10dBspl below
the noise floor of the room. However, the max SPL with 20dB crest for a
110dB sensitivity speaker would be 106dB average (126dB peak).

Is
that too many numbers all at once? In summary, the D75 driving a 100dB
8ohm speaker can't play quieter than 10dBspl and can't play louder than
96dBspl average (116dBspl peak). The D75 driving a 110dB 8ohm speaker
can't play quieter than 20dBspl and can't play louder than 106dBspl
average (126dBspl peak).

Since
a speaker impedance isn't flat,
this all gets a bit more complicated since the maximum power for the
D75 changes with load impedance. Also, a speaker's frequency response
isn't flat either, so the noise floor of the amplifier will get
augmented by the shape of the speaker's frequency response. This can
become extremely problematic when running active crossovers on a horn
with good polar control (since this boosts the on-axis response as you
go lower in frequency). Adjusting the signal before the amplifier
doesn't change how the amplifier noise floor gets augmented by the
speaker response. If you have 10dB of gain from the horn at lower
frequencies, then your system noise floor will be 10dB higher than at
the higher frequencies. That means you need yet another 10dB of dynamic
range in order to keep
the loudest noise floor quiet enough without sacrificing max output.
And all that is assuming your music signal has equal energy at all
frequencies, which definitely isn't the case. Also, some amplifiers
designs don't have a flat broadband noise floor either so add that on
top of everything else too...

In
the end, it's probably easier to pick an amp that has just enough power
to go as loud as you want to listen. If the amp is a good design, then
its dynamic range should be relatively optimized for the power you
need.

Btw, a CD source can contain at best only 96dB of dynamic range and it's usually going to be closer to 90dB. So in this example, if you're using a CD player for the playback source, the amplifier isn't going to be the limiting factor in loudest versus softest signal. It also doesn't help that digital noise floors tend to sound relatively nasty. Vinyl is going to be even worse, but it has the advantage of a much more pleasing sound to its noisefloor. 24-bit audio (like on dvd's) is capable of 144dB, but you're usually only going to get about 130dB or so. I believe human hearing is usually quoted to be around 140dB total, which puts some perspective on just how good our ears are.

[Islander]

Actually, I was trying to be humourous and speak very much in generalities and perhaps sum up the sixty lines or so of the previous posts in four words. It's very true that a high-power amp needs to have a low noise floor, especially with high-sensitivity speakers.

[garyrc]

There are noise floors, and then there are noise floors.

A strange downhill movement has occurred in my audio world .......

In at least one room in my house I have always had high efficiency speakers of the level of the Klipshorns or just 2 dB below (the harsh JBL 030). I had no real trouble with the noise floor until my most modern component arrived. In the past, Dyna 40 wt tube amps, with a Dyna tube pre-amp*, McIntosh 60 wt amps with a McIntosh C28 preamp, a Luxman integrated amp, and, later, the pre-amp section of the Luxman feeding a Yamaha 135 wt power amp in the front, with the power amp in the Luxman integrated powering the rear channels, all produced noise I could hear, but it was not unpleasant, and not intrusive. Usually, the natural noise (including room noise from the original venue) on the tape, Lp, or CD made it even less noticeable. Enter the NAD T163 AV Tuner Preamplifier I bought in 2004. The separate NAD C272 power amps it feeds are as quiet as can be... but the T163 produces an annoying 120 Hz hum, with a slightly softer 60 Hz hum underneath. The problem was between the Volume control and the output, because turning down the volume did not turn down the hum. It went back to the NAD factory, and they did a few things to it, but the remaining hum was pronounced "normal." The 105 dB / wt / M sensitivity of the Klipschorn was blamed. I accept the idea that what is "normal" and acceptable with an 89 dB sensitivity speaker can be objectionable with a 105 dB sensitivity. Back when many speakers (EV, JBL, Altec, and, of course, Klipsch) were very efficient, manufacturers were motivated to produce quiet pre-amps. I reluctantly turned down the input controls on the power amps (reluctantly because I was told that the resolution would be a bit higher without them turned down). That reduced the hum to a barely noticeable level (my family can't hear it without training).

* After being used for about 18 years in my home, then 21 years on loan to always impoverished San Francisco State University, I recently hooked up the Dyna Preamp and discovered it has developed some unpleasant hiss.... but after 39 years of service without a tube or capacitor change, it has a right!

[DrWho]

.... but after 39 years of service without a tube or capacitor change, it has a right!

lol [Y]

So have you had any thoughts of hotrodding your NAD to bring down its hum?

[DrWho]

Actually, I was trying to be humourous and speak very much in generalities and perhaps sum up the sixty lines or so of the previous posts in four words. It's very true that a high-power amp needs to have a low noise floor, especially with high-sensitivity speakers.

lol, so in due form I respond with another hundred lines of rambling []