Quality speakers for classical music with high output/volume

[garyrc]

I always thought that the loudness controls on the early 1970s receivers that my friends and I owned in college were way off, and it seems as all the later research has proven why. (Bolding mine, Garyrc)

In the 1970s 1960s, and in high school (late 1950s), my friends and I had very fine tunable "2 knobs, and switches" loudness controls. CentraLab made and sold them. Wish I had a photo, but they seem to have vanished. We put them on everything. I installed one on our 1960 RCA Lane TV, and it sounded like a high fidelity component for the first time (the compensation worked pretrty well partly because it fed a JBL D130 "extended range" woofer -- 15" diameter --

in a Karlson enclosure

, with a JBL 075 ring radiator

tweeter). The procedure was to would turn one of the knobs up to "as loud as you would ever want it," said the instruction sheet, then turn down the other knob to the desired -- softer than max -- volume. It provided a continuously variable loudness compensation. For atypical recordings and broadcasts we could adjust one or both knobs by ear. There were also two switches for adjusting treble, 8dB and 12 dB boost at the top end, IIRC.

 

[mixsit]

It would mean that even if multiple people were speaking, it would drown speech. That is, instead of a small group of people overpowering the speakers, they should have a hard time hearing each other. Much like during the peaks of an orchestra performing--you wouldn't be able to hold a conversation, much less worry that your conversation will overpower the sound.

I have put together my notes from this thread and will post a summary response likely today.

I would like to with some caution/reservation to say that level of volume -masking of speech not typical of live orchestral. A big band at forte and very close range perhaps? :>)

 

[mixsit]

In the 1970s 1960s, and in high school (late 1950s), my friends and I had very fine tunable "2 knobs, and switches" loudness controls. CentraLab made and sold them. Wish I had a photo, but they seem to have vanished. We put them on everything. I installed one on our 1960 RCA Lane TV, and it sounded like a high fidelity component for the first time (the compensation worked pretrty well partly because it fed a JBL D130 "extended range" woofer -- 15" diameter --

in a Karlson enclosure

, with a JBL 075 ring radiator

tweeter). The procedure was to would turn one of the knobs up to "as loud as you would ever want it," said the instruction sheet, then turn down the other knob to the desired -- softer than max -- volume. It provided a continuously variable loudness compensation. For atypical recordings and broadcasts we could adjust one or both knobs by ear. There were also two switches for adjusting treble, 8dB and 12 dB boost at the top end, IIRC.

Oh my. The '030 system (...don't ask :>)

 

[riker1384]

In the 1970s 1960s, and in high school (late 1950s), my friends and I had very fine tunable "2 knobs, and switches" loudness controls. ..........

The procedure was to would turn one of the knobs up to "as loud as you would ever want it," said the instruction sheet, then turn down the other knob to the desired -- softer than max -- volume. It provided a continuously variable loudness compensation. For atypical recordings and broadcasts we could adjust one or both knobs by ear. There were also two switches for adjusting treble, 8dB and 12 dB boost at the top end, IIRC.

That sounds like how Yamaha's variable loudness on their stereo receivers works.

 

[Vladimir Filevski]

THD is irrelevant. IMD, and not the least Doppler is happily ignored. That was it for me now. I'm seasick from walking circles. Never mind

Well, no. THD is relevant, also IMD and Doppler. I was comparing THD from various speakers (apples to apples comparison) only because there are no IMD or Doppler measurements for them, as far as I know. If you have IMD and Doppler measurements for the speakers which were discussed here, please do share them.

 

[garyrc]

That sounds like how Yamaha's variable loudness on their stereo receivers works.

Yes, I think so.

 

[Vladimir Filevski]

If you don't mind, could you also address this premise that crossing over lower frequencies will somehow decrease the distortion of the speaker because it has to travel less? Because to me it sounds like loudness and distortion sort of go hand in hand. If crossing over lower frequencies could actually decrease the distortion of the speaker, then this would mean the speaker could now play louder to reach the same distortion level.

Keeping the same premise for the crossover - satellite high-passed and subwoofer low-passed at 100 Hz, that will help only in lowering, to some extent, the IMD and Doppler distortion (not THD!) for the two-way satellite loudspeaker. Three-way satellite will have no benefit. Of course, subwoofer will enable very loud SPL, deep frequencies and low distortion below 100 Hz - for which the satellite speaker is not capable for.
Do not become obsessed with low distortion (of any kind) numbers - I am sure you will enjoy your Neumann KH 150 even without the sub.

 

[Vladimir Filevski]

Nice explanation. To confirm how well I understand (or don't understand) these curves:

• Let's say a composer has written a piece for a bassoon (50Hz+) and violin (200Hz+) duo--unrealistic, but simplistic for the concept illustration
• The composer would control how loud each of the instruments would have to play to match his intentions. This means if both instruments are to be heard equally, then the bassoon will already play let's say 20dB above the violin (or the violin will play quieter)

Not quite - bassoon will play as intended, with no requirements to be 20 dB louder than the violin.

• A microphone recording this work captures it just like the human ear would hear it it, i.e., it's flat and you don't need EQ to hear the bassoon

Yes.
For the bolded part (short story): Yes-ish, sort of.

• Now when this recording is played back on a 2 channel speaker, there will be no issues at all as long as the volume is kept below the highest bass SPL that the speaker can produce.

Big yes.

So for example, KH-150 can produce 50Hz at about 100 dB (3% THD, black line in the graph below). This means that music would have to play at ~80 phons or lower to capture the full bass (see annotated graph below). In fact, no subwoofer is needed at this volume.

• But now if we raise the volume to lets say 100 phons, to truly hear the bass, that's when we would need a subwoofer to go up to 110 dB at 50Hz, otherwise the bass will start fading into the background (or become distorted, assuming you can hear that)

No!
Use FM curves only for determining how much bass boost you need at lower SPL than "original, normal" SPL. Recording already have captured the correct balance between bass and mid/highs.
If "normal" SPL captured in recording was 80 phons, when you reproduce it at 80 phons you don't need any bass boost!
If you want to reproduce this recording at 100 dB, you likely will want to reduce bass slightly (nobody is doing that! ).

 

[garyrc]

If you have IMD and Doppler measurements for the speakers which were discussed here, please do share them.

Not for the speakers discussed here. If the late Richard C. Heyser were still alive and reviewed any of these, you would have the IM data -- and much more.
-------------------------------------------------------------------------------------------------------------------------

@excelsius you asked about where they got the 5 dB lower SPL recommended for home theaters or music listening rooms. Here is the chart. Some explanation is necessary. Their initial level of 85 dB is their nominal somewhat arbitrary level for huge rooms (e.g. concert halls). They figure 20 dB head room is needed above the 85 dB for peaks, bringing the total to 105 dB (10 dB more for bass, especially below 80 Hz). I was off by 2 dB; The recommendation was for 7 dB lower for 4,999 cu. ft. As you can see, these figures are approximate.

As far as cone excursion causing IM distortion (Doppler or "modulation diatortion"), at one time Klipsch made two expensive speakers with exactly the same woofer, midrange, and tweeter in them. One was fully horn loaded (including the bass; the Klipschorn) with very low woofer cone excursion, and the other one was a bass reflex design (the Cornwall). Even with the same drivers, the Cornwall had 3 times the Total Modulation Distortion as the Klipschorn, even though the Cornwall was at 10 dB lower in SPL.

​

 

[rdenney]

This discussion is exactly about speakers reproducing the music, as recorded. Assuming FM curves are correct, that automatically means that all that information is already encoded in the music for you to be able to hear it. It just becomes a matter of at which SPL can the speakers reproduce it. It's not about doing anything special to hear that music. I don't think we're in disagreement there.

@Newman, I don't quite follow when you say the frequency spectra doesn't follow the FM curve. Which spectra would that be? I don't think you can simply look at the average spectra output of a music piece and expect it to look like an FM curve. Music has an additional dimension of time. And even if you break it down and analyze any one instantaneous segment of time and chart freq vs SPL, it still does not mean that the shape will look like an FM curve. It would depend on what the artist was doing it that moment since music can have varying levels of bass to midtones depending on intention. That's why I simplified it to two musical instruments here. If you're in a room listening to a live performance, that should mean that if at any one moment the lowest note on the bassoon and the violin sound at equal volume to you at 80dB, then in that moment, the SPL of the bassoon must be 20dB higher than the violin. If that was not to be the case, then the FM curve would be inaccurate since that's the definition of that chart, as I understand it.

Nothing is “encoded”. Our hearing is what it is, whether we are sitting at home or in the concert hall. The ”encoding” (and decoding) is done in our brains.

The thing with loudness curves is not the fact that Fletcher-Munson is curved, but that it is curved A LITTLE differently at different loudness levels. But we don’t fix that with speakers, which we don’t want to be customized for one loudness level, but rather with EQ or even appropriately designed tone controls or loudness filters. That’s assuming we don’t think our brains are up to the task of compensating (because we know we’ve turned it down).

But at reference volume, we specifically can (and should) ignore even that subtlety, because that’s where our hearing spectral sensitivity is the same whether it’s live or Memorex.

Rick “getting wrapped up in a non-issue” Denney

 

[krabapple]

This thread is Exhibit A for why ASR should have a very prominent, flashing sign on its front page: READ THIS BEFORE YOU POST ANYTHING, with a link to a plain-language debunking of the most common audiophile nonsense beliefs, of the sort the OP came into this thread with.

 

[sejarzo]

This thread is Exhibit A for why ASR should have a very prominent, flashing sign on its front page: READ THIS BEFORE YOU POST ANYTHING, with a link to a plain-language debunking of the most common audiophile nonsense beliefs, of the sort the OP came into this thread with.

Mmmhmm. Maybe you are right, as way too many don't take the effort to scroll down far enough to see these, which would provide either answers to their basic questions, or at least allow them to get a better idea on how to ask a relevant question.

 

[sejarzo]

Nothing is “encoded”. Our hearing is what it is, whether we are sitting at home or in the concert hall. The ”encoding” (and decoding) is done in our brains.

The thing with loudness curves is not the fact that Fletcher-Munson is curved, but that it is curved A LITTLE differently at different loudness levels. But we don’t fix that with speakers, which we don’t want to be customized for one loudness level, but rather with EQ or even appropriately designed tone controls or loudness filters. That’s assuming we don’t think our brains are up to the task of compensating (because we know we’ve turned it down).

But at reference volume, we specifically can (and should) ignore even that subtlety, because that’s where our hearing spectral sensitivity is the same whether it’s live or Memorex.

Rick “getting wrapped up in a non-issue” Denney

I think the OP meant that when a bass region tone and midrange tone perceived as equally loud are recorded, the dBFS level of the bass tone is higher than that of the midrange tone. Some FAQs on digital recording often use a phrase like "the purpose of the ADC is to encode analog voltage signals into digital data."

 

[Steve Dallas]

I have seen more than a few comments here (meaning, not only on AV enthusiast forums where one might think it's about effects, not music) regarding how high some users typically set their sub gains that make me wonder if they were born with severely compromised hearing down in that range.

I suppose it is all about the feelz? I have Audyssey's Dynamic EQ turned off, because I do not like much bass when listening quietly, which makes me a bit of a unicorn in AV circles. Who likes booming bass when taking a nap?

 

[krabapple]

Who wants music while taking a nap?

 

[Toltek]

Nice explanation. To confirm how well I understand (or don't understand) these curves:

• Let's say a composer has written a piece for a bassoon (50Hz+) and violin (200Hz+) duo--unrealistic, but simplistic for the concept illustration

Let me proceed with this example as an aforementioned composer. So I ask the violin player to play a single tone at 80 db and he happily obliged. Then I asked a bassoon player for 100 db and he refused. Said something about how his lungs have to be at least 100 bigger or something, since when he blows as hard as he can, he can only output 80 db and going 20 db up in SPL would require him to increase his blowing power by a hundred. Also mentioned something about final SPL being not 80 db, as written by me, a famous composer, but 100db+80db=100.04 db.

I said OK, I got your point, now you blow as hard as you can, so we still aim for 80 db total SPL, then I checked the curve and advised the violinist to play at 40db. This time the violinist disagreed, since no one in the audience would hear him at all! In the end I sat right next to bassoon player, asked the violinist to step back 100 meters and I enjoyed my fabulous piece as intended.

 

[sejarzo]

It's all I got.

 

[Ciobi69]

Not for the speakers discussed here. If the late Richard C. Heyser were still alive and reviewed any of these, you would have the IM data -- and much more.
-------------------------------------------------------------------------------------------------------------------------
@excelsius you asked about where they got the 5 dB lower SPL recommended for home theaters or music listening rooms. Here is the chart. Some explanation is necessary. Their initial level of 85 dB is their nominal somewhat arbitrary level for huge rooms (e.g. concert halls). They figure 20 dB head room is needed above the 85 dB for peaks, bringing the total to 105 dB (10 dB more for bass, especially below 80 Hz). I was off by 2 dB; The recommendation was for 7 dB lower for 4,999 cu. ft. As you can see, these figures are approximate.
View attachment 275681
As far as cone excursion causing IM distortion (Doppler or "modulation diatortion"), at one time Klipsch made two expensive speakers with exactly the same woofer, midrange, and tweeter in them. One was fully horn loaded (including the bass; the Klipschorn) with very low woofer cone excursion, and the other one was a bass reflex design (the Cornwall). Even with the same drivers, the Cornwall had 3 times the Total Modulation Distortion as the Klipschorn, even though the Cornwall was at 10 dB lower in SPL.

​

I have the Klipschorn myself I didn't know about this interesting fact! Thanks!!!

 

[excelsius]

Aren't those CEA-2010 ratings meant to be compensated to free space? In which case, add 6 dB to the sub's capability for floor placement in the centre of the room, another 6 dB for moving it to the middle of a wall on the floor, and another 6 dB for moving it into a corner. Then downrate that +18 dB to +12 dB for real room gain being less than theory. So a sub might only need to score 108 dB on the CEA-based chart to deliver 120 dB in a room.

But 120 dB was assumed by you off a chart based on an, um, uniquely unreal set of assumptions, see above. If we return to the 105 dB in my post above, and even allow an extra +10 dB for LFE effects (which is no longer a discussion of music), that's 115 dB. Less 12 dB for real room gain, so 103 dB.

And that's with one sub.

And that's for peaks. I don't think the CEA rating is for momentary peaks.

Thanks for bringing this up. I looked it up and it seems you're referring to room gain. I will have to get a better understanding of this, but based on this article from SVS, it seems like it might be more complex than simply adding 6dB per surface. The article says the frequencies affected depend on the dimension of the room, the sub might need to be sealed, and the location of the sub might not have much of an effect. Have to read more to understand the nuances, but this is great to know since will certainly come in handy in picking a sub.

 

[Newman]

Perhaps re-read my post that you are quoting. I already used the term ‘room gain’ and I already down-rated the 6 dB per half-space reduction for real rooms. The answers to the issues you raise are already in my post.

cheers