Quality speakers for classical music with high output/volume

[ZolaIII]

@fineMen white noise reference calibration point is 83 dB. It's not +20 dB a top with EBU R128 (eventually 10 dB). I have good neighbours, live in a hause and have good isolation and even so I hardly ever go over 83 dB calibration point. I did it even over 100 dB on short critical listening seasons but with really great headphones. I worked on stages and club's and their levels of 100 dB or more and with bad both materials and equipment I never ever considered as anything else then horrible amplification of noise.
Whose referring to subjective like feal, desire and love (against which you really can't go) emotional reactions and objective reasons like state of hearing when referring to different "preference" in main and low bass regions. Let's face it we are not all the same (and that's a good thing).

 

[bodhi]

I lost the breadcrumb trail. Are we talking about an individual bass lift, maybe even record by record, or the necessary output capabilities at home?

Yes. The amount of loudness compensation to reach the same balance that the record was produced varies record by record. Even worse, you have no idea if the balance that mixer preferred would be the same you yourself like the best, even if you knew the exact volume level and equipment the mixer used and could compensate accordingly.

As it is almost impossible to accurately use loudness compensation for any record, song or even a piece of a song we can give up on the idea. Just let the song play however it does with your personal parameters.

Then again, we can use some generalizations and assume that records are usually produced with approximately standard parameters and equipment and use that information with our preferred listening levels and compensate. We can also use dynamic compensation which scales with the volume.

Now, it seems that as the latter approach is preferred by majority of people. It's certainly not more "wrong" than not doing any compensation at all, so it seems logical to give it a try.

 

[Vladimir Filevski]

This was explained before, but I feel it need additional explanation again , with examples:

1. An active "satellite" speaker (here: Neumann KH 150) with ideally flat frequency response is capable of, for example, maximum SPL = 110 dB (with less than 3% THD) above 200 Hz, and SPLmax = 100 dB (with < 3% THD) below 100 Hz. Because active loudspeakers have internal amplifiers, maximum SPL is determined either by maximum available power from the amplifier before clipping or by maximum THD from the loudspeaker. Of course, asking (internal) amplifier to output more power lead to amp clipping and excessive THD.
If we use 100 Hz high-pass filter for that "satellite" and add sub below 100 Hz, available maximum satellite SPL above 200 Hz still is 110 dB (with less than 3%THD)!
Why then the "trouble" with adding the sub? Because sub can deliver more than 110 dB clean (<3% THD) SPL below 100 Hz, instead of 100 dB without it (sub). Now, we have clean SPL = 110 dB for all frequency spectrum (including 100 Hz), instead of only 100 dB at 100Hz without the sub.

2. Let say Neumann have made passive “satellite” model "KH 150 PASSIVE" with the same drivers and enclosure from KH 150, but with passive crossover. We can connect this passive speaker to 1 kW amplifier, but no avail - again, maximum SPL above 200 Hz still is 110 dB irrelevant of the input wattage (from the 1 kW amplifier), because of the 3% THD limit of the loudspeaker drivers! We do want clean and undistorted sound, don’t we?
Of course, 1 kW can deliver higher dynamic SPL peaks (than KH 150 internal amplifiers) above 200 Hz – but with higher than 3% THD (coming from the speaker, not the amp)!

 

[Steve Dallas]

To reiterate, it's a simple question: what specs does one use to make an informed decision about which sub to integrate into a given system? Hopefully you agree that choosing a sub by cone diameter does not make sense.

It is not a simple question, especially since your use case is well outside the mainstream. To reiterate, no one attempts to listen as loud as you claim to want to with your size constraints, space constraints, and budget constraints. What you are attempting is not normal, therefore you will encounter little actual practical experience with that application.

The answer for most people in domestic situations is "good enough is good enough." This assumes passive speakers. Allow me to elaborate on the process using myself as an example.

My application is a combined home theater and listening room. Passive speakers are Revel F206, which are sized appropriately for the space and have an in-room F3 of 32Hz. My goals are to:

1. Extend in-room bass extension down to below 20Hz
2. Increase low bass SPL at least to THX reference level
3. Smooth bass response in the room (help with room modes)
4. Reduce distortion in the stereo pair.

Other criteria include at least two subs to fill in room nulls, and they must be small-ish.

I read objective reviews and settle on a pair of SVS SB-2000 12" sealed subs. They extend below 20Hz (meaningful bass will extend below 10Hz with boundary reinforcement) and are each capable of ~115dB at 30Hz according to CEA 2010 measurements, which is more than enough.

I take a day to carefully set them up via measurements: position, phase, crossover frequencies and slopes, levels, verify my results, and in the end, "good enough is good enough." The selection approach is hardly scientific, and the setup process is informed experimentation, AKA trial and error.

The result for critical listening is I can now push my stereo pair as loud as I want to without audible distortion--far louder than I can stand, and midbass clarity is improved because cone excursion of the 6.5" woofers is greatly reduced, as they are no longer trying to produce frequencies below ~70Hz.

^ The problem with the above result is that my speakers were not distortion-limited in the first place. Adding the sub made them cleaner, but I do not listen any louder, because I do not like to, which limits that benefit. My preference for loudness is high 70s to low 80s nominal, and that does not change because I technically can listen louder without distortion. I do not benefit from goal number 4 in real life. Bookshelf speakers benefit far more from this than towers.

Your situation is more complex using active speakers:

1. Active speakers are power-limited
2. Active speakers have protection circuits
3. No one does what you are trying to do with small actives.

Taking these one at a time...

1. No matter what you do with subs, actives always run up against power limitations. You will gain some SPL by high pass filtering the stereo pair, as the amps are less stressed by not having to attempt to drive low frequencies, but there are still limits. And, you cannot throw more power at them like you can with passives.

2. Similarly to the above, actives often have several protection circuits to protect the circuits and drivers. You cannot get around this.

3. The closest area of expertise I can think of to your application is a sound designer who sets up studio control rooms, mixing studios, mastering studios. But, they do not try to reach anywhere near your expected SPLs. They are designing for the high 70, low 80 nominal dB used by those engineers. They do not have criteria for damaging people's hearing. Further, those who mix and master classical music often use passive speakers with high power amps and/or huge active monitors. There is little to no precedent for your use case.

A few words on using subs to reduce distortion...

As has been stated in this thread, you can reduce woofer distortion in the stereo pair by high passing them and filling in the missing frequencies with subs. We are not terribly interested in low bass distortion, because we are not terribly sensitive to it. In a two-way speaker, this reduces midrange distortion, because the high pass filter reduces cone excursion of the driver responsible for bass and midrange up to the crossover frequency, and this is where the real benefit is. Distortion reduction is not limited to the sub / woofer crossover frequency; the benefit extends much higher than that. Things get even better with a three-way for obvious reasons.

TLDR:

1. The process is to find objective data and choose a sub that has the FR extension you desire, will play at the desired SPL, and has the distortion and compression characteristics you desire. Then carefully integrate it into your system, then measure the results.
2. Adding a sub DOES allow you to play music at higher SPL, but the benefit is limited in active speakers. If you want louder, you really need bigger amps and drivers.

 

[Vladimir Filevski]

Here are some 105 dB results taken from magazine reviews for both the possibly more meaningful IM distortion (which I think might be the same as "modulation distortion?") and THD
Everything else being equal, three way speakers often have less modulation (Doppler) distortion than two way...

Yes, three or four-way loudspeakers have less intermodulation or Doppler distortion, but OP asked for a good bookshelf loudspeaker. Bookshelf loudspeakers mostly come in a two-way format.

Now, going back to the misconception that subs are not needed, this is a larger version of the ISO equal loudness contours.
View attachment 275312
Assume that I am monitoring music so the midrange is roughly 80 dB. What SPL do I need to sense equal loudness at 50 Hz? Is it not about 102 dB?
If I am monitoring music through a KH150 that measures flat via a calibrated measurement mic from ~40 Hz to 20 kHz, what SPL will it provide in that lower register versus the midrange?

There is a widespread misconception that listening to music at particular SPL have to follow Flecher-Munson (or ISO 226:2003) curves. That is wrong!
Suppose we are listening to a solo piano recording, at the same moderate SPL=80 dB (average at mid frequencies) as in live concert at the fifth row. Should we boost low frequencies, say at 50 Hz, by +21 dB, to 101 dB - as the graph below indicates? Of course not - mic (recording) already had captured correct balance between low and mid frequencies (actually, piano key G1 has harmonics that are louder than 49 Hz fundamental!). That 21 dB boost (!!!) at 50 Hz will result in a catastrophic change in the timbre! The same applies to all low frequencies.

But what Fletcher-Munson curves are good for? For indication how much bass boost we should apply, when listening at much lower SPL than "normal". For example, listening the same piano recording at 60 dB, we should boost +12 dB at 50 Hz, to enjoy the same correct timbre. How we end up with this +12 dB boost? Simple - by comparing 80 phon and 60 phon equal loudness curves at 50 Hz: for 80 phon it is 101 dB at 50 Hz, and for 60 phon it is 89 dB at 50 Hz. The difference is 12 dB (101-89). Almost all "loudness" switches in amplifiers are boosting bass at around 15 dB.

 

[fineMen]

Yes, three or four-way loudspeakers have less intermodulation or Doppler distortion, but OP asked for a good bookshelf loudspeaker. Bookshelf loudspeakers mostly come in a two-way format.

=> AdamAudio A77H (dig/act, no DSP 1,2k$ each)

 

[Triliza]

Yes, three or four-way loudspeakers have less intermodulation or Doppler distortion, but OP asked for a good bookshelf loudspeaker. Bookshelf loudspeakers mostly come in a two-way format.


There is a widespread misconception that listening to music at particular SPL have to follow Flecher-Munson (or ISO 226:2003) curves. That is wrong!
Suppose we are listening to a solo piano recording, at the same moderate SPL=80 dB (average at mid frequencies) as in live concert at the fifth row. Should we boost low frequencies, say at 50 Hz, by +21 dB, to 101 dB - as the graph below indicates? Of course not - mic (recording) already had captured correct balance between low and mid frequencies (actually, piano key G1 has harmonics that are louder than 49 Hz fundamental!). That 21 dB boost (!!!) at 50 Hz will result in a catastrophic change in the timbre! The same applies to all low frequencies.

View attachment 275547

But what Fletcher-Munson curves are good for? For indication how much bass boost we should apply, when listening at much lower SPL than "normal". For example, listening the same piano recording at 60 dB, we should boost +12 dB at 50 Hz, to enjoy the same correct timbre. How we end up with this +12 dB boost? Simple - by comparing 80 phon and 60 phon equal loudness curves at 50 Hz: for 80 phon it is 101 dB at 50 Hz, and for 60 phon it is 89 dB at 50 Hz. The difference is 12 dB (101-89). Almost all "loudness" switches in amplifiers are boosting bass at around 15 dB.

I'm still confused about how to correctly implement loudness. Currently I'm using a Loudness filter via moOde/camillaDSP, which it is the same as the one RME ADI-2 DAC FS is using:

filters:
  loudnessvol:
    type: Loudness
    parameters:
      ramp_time: 200.0
      reference_level: -25.0
      high_boost: 7.0
      low_boost: 7.0

I read in another thread that Dr. Toole says that the highs don't need any boosting in low volumes, only bass. What would be an alternative for the above filter?

 

[fineMen]

I'm still confused about how to correctly implement loudness.

I might be an exception to the rule. My stereo is, virtually nerver listened to in real "stereo" sitting in the middle and so. I hate it!!

But, the speakes are pretty capable way above what 95% of the audience here may experience more often in quality and max spl. My average listening level (average, not peak!) lies around some humble 60dB, maybe 70db in my weekly wild 5 minutes. The frequency response is flat, just flat. I don't miss the "bass". It is there without pressure.

Do you actually miss the drama of distortion? What is it, that missing piece?

Toccata and Fuge, Bach, big organ, 80+dB as a show-piece. My friend looking out for the huge boom, Bach castrated, what did you do?

Answer was in the spectral analysis of the actual sound in my flat: full of 30Hz content and more. Reason why not heard? No distortion! Proof: got bad speaker in and replayed, bass was all there ... not in the spectrum, but audibly.

That is why I still recommend to have a recording microphone at hand, and some proficiency in using it.

 

[Vladimir Filevski]

=> AdamAudio A77H (dig/act, no DSP 1,2k$ each)

I presume it is an excellent very good studio monitor, but for a domestic living room it is somewhat on the big side and also is in the "wrong" landscape format.
I am curious how Adam A8H will perform, pity it is in the "landscape" format, too.

 

[fineMen]

[AdamAudio A77H] ... is somewhat on the big side and also is in the "wrong" landscape format.

C'mon

 

[Spocko]

I am building my first system where I am going to stream music from my computer to a streamer (likely Bluesound Node) and connect it to the speakers both in my office and in another room.

One of my first dilemmas is getting high quality speakers that can reproduce 24/192 sound (or at least 24/96) with high fidelity, mostly classical music, including vocals, but also be capable of high volume output (current room size is 13.5 x 20.5 ft). I have been considering KEF LS50 II, but according to a review linked on this forum, that speaker might not have a high output. Review says it's 90dB max at 13ft (4m).

My question is what speakers can I get that produce both accurate sound and also have high output? I'd prefer active speakers, but can consider passive with an amp (don't have an amp currently). The budget is under $3000, but can go a bit higher if it will make a big difference. Also would prefer bookshelf rather than floor standing speakers.

The irony of your request is that you have two objectives that may be mutually exclusive (1) highest fidelity, low distortion and (2) high SPL. For you to be able to distinguish the subtle distortions and differences at the bit level (assuming you can of course) requires you to listen at a volume below 80dB because at higher SPLs the loudness masks distortion you'd otherwise notice at 73dB (your hearing is protecting itself, attenuating for the volume) during critical listening. If you are planning to listen above 83dB, then focus on balanced/accurate bass retrieval by budgeting for a subwoofer or two (with necessary room EQ to flatten the peaks and valleys at your MLP) to match the required 90dB at 1Khz as others have noted above.

To your inquiry, there are no active monitors under $4K that will deliver the bass required for classical music at high volume output from 13ft away - HOWEVER if you increase your budget to $13K then I highly recommend the Dutch & Dutch 8C active studio monitors which will do everything you asked because it uses wall boundary and DSP to reinforce bass response and designed to do what you are asking.

 

[Vladimir Filevski]

I'm still confused about how to correctly implement loudness. Currently I'm using a Loudness filter via moOde/camillaDSP, which it is the same as the one RME ADI-2 DAC FS is using:

filters:
  loudnessvol:
    type: Loudness
    parameters:
      ramp_time: 200.0
      reference_level: -25.0
      high_boost: 7.0
      low_boost: 7.0

I read in another thread that Dr. Toole says that the highs don't need any boosting in low volumes, only bass. What would be an alternative for the above filter?

I am not using Camilla DSP, but I suppose this is OK:

reference_level: -10.0 (or your choice)
high_boost: 1.0
low _boost: 15.0

 

[sejarzo]

There is a widespread misconception that listening to music at particular SPL have to follow Flecher-Munson (or ISO 226:2003) curves. That is wrong!

.....

But what Fletcher-Munson curves are good for? For indication how much bass boost we should apply, when listening at much lower SPL than "normal". For example, listening the same piano recording at 60 dB, we should boost +12 dB at 50 Hz, to enjoy the same correct timbre. How we end up with this +12 dB boost? Simple - by comparing 80 phon and 60 phon equal loudness curves at 50 Hz: for 80 phon it is 101 dB at 50 Hz, and for 60 phon it is 89 dB at 50 Hz. The difference is 12 dB (101-89). Almost all "loudness" switches in amplifiers are boosting bass at around 15 dB.

Of course I agree with your analysis. I was posing a question to the OP about the relative power required to produce a fundamental at 50 Hz that is perceived as equally loud to music in the midrange.

I never said that such a difference was not already, as some have put it, "baked into" the recorded signal. If an ensemble were to be playing such that the midrange signal was recorded at -40 dBFS, then the 50Hz tone that was perceived as equally loud would have been recorded at about -18 dBFS.

If an 8 ohm speaker required 1 watt/2.83 Vrms to reproduce that -40 dBFS/80 dB SPL midrange, the 50 Hz signal would require 35.6V rms, or 158 watts.

My reason for doing so was trying to lead the OP to realize that one cannot consider only math ratios and flat frequency response curves when considering capabilities of real world transducers with legitimate physical limits. He seems enjoy listening at higher than normal SPLs, and I sense a real possibility that he will run his new speakers into the limiters and be unhappy with the result.

 

[ZolaIII]

@Triliza to reference calibration point you make other parameters.
I do calibration to 83 dB white noise and you don't need correction above it or a bit under.
Let's say you do calibration the same for the sake of explanation.
Let's say you won't listen below 60 dB so you set reference_level: - 23
high_boost: 12.0
low_boost: 6.0
So when you dial volume down for - 23 dB it will apply full high/low values and in between depending how much you dial it down to the max set values you entered which can be in up to 20 limits for high/low boost. I don't really like that implementation especially because bass boost starts very low. You can play with values of course.
Edit: actually scrap that!
you set reference_level: is to where you set reference_level: don't want any correction to happen. So for calibration point of 83 dB set it to - 5
It engage full values of:
high_boost:
low_boost:
parameters you set at - 25 dB from that and granular less if you set - dB volume less. So try putting tham:
high_boost: 15.0
low_boost: 3.0
And play with it.
A correction level table towards frequency and listening levels in dB.

 

[Vladimir Filevski]

C'mon

No, really: 21" wide "bookshelf" speaker is not acceptable in most living rooms.
Also, Adam A77H has lower SPLmax (<3% THD) than Neumann KH 1150 for all frequencies, except in the 100 Hz - 450 Hz band.
On the other hand, Adam A77H can reach higher SPLmax (at 10% THD) below 90 Hz.

 

[sejarzo]

I read in another thread that Dr. Toole says that the highs don't need any boosting in low volumes, only bass.

The equal loudness curves above the midrange peak at 1.5k are very consistent while they are much steeper in the bass region as you go down in perceived loudness. The curve for 80 phon perceived loudness peaks at about 82-83 dB at 1.5k, then at about 92 dB SPL around 10k for a difference of about 10 dB. That difference is pretty close to the same on the 60 phon curve, and only about 1 dB more on the 40 phon curve, which is well below typical listening levels. If you look at the difference down at 100 Hz, the required dB boost for equal loudness in phons goes from 9 dB at 80 phon, to about 16 dB at 60 phon, to about 19 dB at 40 phon.

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.

 

[ZolaIII]

This is JRiver Loudness normalisation controls (ISO 226 2003) as measured (that I use):

https://www.audiosciencereview.com/forum/index.php?attachments/1574203824523-jpeg.39415/

It works fine tho I use lower calibration point.

 

[Triliza]

Thank you all for your feedback, I got some ideas to experiment a little bit more with the topic

 

[fineMen]

No, really: 21" wide "bookshelf" speaker is not acceptable in most living rooms.
Also, Adam A77H has lower SPLmax (<3% THD) than Neumann KH 1150 for all frequencies, except in the 100 Hz - 450 Hz band.
On the other hand, Adam A77H can reach higher SPLmax (at 10% THD) below 90 Hz.

O/k, a new form factor. In a shared household the significant other may be pleased with an appropriate renewal of the whole interior.

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

 

[excelsius]

@excelsius sub's sheat:

Subwoofer Comparison

I've created a *slightly large* subwoofer comparison spreadsheet. :D Info: -==Link to my Subwoofer Comparison spreadsheet ==- https://bit.ly/SubwooferComparison

www.audiosciencereview.com

Neumann KH150 driver response:

https://www.audiosciencereview.com/forum/index.php?attachments/neumann-kh-150-two-way-studio-monitor-speaker-active-near-field-driver-frequency-response-meas-png.248932/

It shows it all including why to cross @ 100 Hz.
Your question whose wrong, it fals off a cliff little under 50 Hz like it should for it's driver diameter no magic there.

Edit: this rises a bit more concern with me. It turns out woffer can go that high because it's early crossed about 1.6 KHz compared to most designs about 2 KHz. This puts additional pressure on tweeter as in classical orchestral music there will be more overtones (by far compared to simpler vocal one) and most second order harmonics will fall on tweeters it's not a good thing for it.
View attachment 275391

Thanks for the link and the nice visual. I had found the subwoofer comparison sheet elsewhere, but was not aware of the subwoofer thread, which is helpful. I think there lies the answer: CEA-2010. This is where SPL is limits are tested, but using audible distortion metrics, which is key. This is all I was asking. The posts simply stating cone size are very misleading because it matters little. One can clearly see from that list of subs a 12" sub outperforming an 18" one at 25 Hz. To me, the way to select a sub would be to first know the frequency vs SPL response of your speaker, select where you need to crossover, and then select a sub for that frequency and below based on CEA-2010 SPL. At least this is how I would explain this to someone trying to buy a sub.

The SPL response does drop off more sharply below 50Hz for KH-150, which is why I mentioned the crossover could be at 50Hz and below, instead of 100Hz or higher.