Quality speakers for classical music with high output/volume

[excelsius]

Another one--floor reflections make themselves noticed in the frequency range of about 400..600Hz. What can a 2..3(??)cm rug do about it? As my own(!) measurements revealed, not too much. So, if people use rugs, it is because of wishful thinking, or, in more contemporary words, cargo cult science.

Does listening to 'classic' prevent you from getting used to do measurements? More into 'culture' than in 'tech'? That would be bad in case of longing for technological excellence, indeed!

Found this (1in = 2.5cm):

At the risk of stating the obvious, the purpose here is to get rid of the sound waves hitting the floor and getting to your ear. In acoustics 101, we would want to block everything above transition frequencies of a 200 to 300 Hz. That would call for an absorber that is at least 4 inches thick. Here is a simulation of one:

View attachment 2178

The horizontal scale is frequency and vertical is amount of absorption. And "alpha" if 1.0 means perfect absorption. We see that this 4 inch absorber achieves very good absorption down to 200 to 300 Hz. We call these "broadband" absorbers and it is what you need to use if you deploy absorbers in your room.

A 4 inch absorber on the floor would be problematic of course. Fortunately listening tests show that the coloration that occurs are above 500 Hz. Here is the simulation of a 1 inch absorber:

View attachment 2179

Good absorption starts around 1000 Hz or so which is fine for this application. A bit thicker would be nicer but not mandatory.

The operating principle here is simple: movement of air molecules is converted to heat. That is why you want the rug and underlayment to be porous. The larger wavelengths cause the molecules to be active well above the rug (highest velocity) so they are not affected by the relative thin rug. So if our aim is to absorb smaller wavelengths/higher frequencies of sound, we don't need as thick of an absorber.

 

[Newman]

Another one--floor reflections make themselves noticed in the frequency range of about 400..600Hz. What can a 2..3(??)cm rug do about it? As my own(!) measurements revealed, not too much. So, if people use rugs, it is because of wishful thinking, or, in more contemporary words, cargo cult science.

Does listening to 'classic' prevent you from getting used to do measurements? More into 'culture' than in 'tech'? That would be bad in case of longing for technological excellence, indeed!

Toole's research has convinced him that highly suppressing floor reflections is an overrated priority, and that psychoacoustically a good carpet on a quality underlay provides the important attenuation. So I don't support your bit in bold above.

 

[Deleted member 48726]

We “chose deliberately to ignore” because it has been covered in previous threads and it is not an issue. The real-life delays of internal subwoofer ADA processes only represent a few degrees of phase at sub frequencies, not nearly enough to cause an audible change in the FR.

Suggest you stop bringing up a non-issue like it is The Big Issue, and especially stop implying that people are ‘turning a blind eye’. Not so.

I think that's a blanket statement that doesn't hold water IRL.

Using YPAO, even though the sub is closer to the MLP than the speakers, it sets the distance to the sub to 8 m. (L+R = 3.2 m.) SVS SB2000.

In my current setup, to integrate the sub SVS PB2000, a delay of 5 ms is added to the L+R. This delay is measured.
That would correspond to phase error of 144 ° @ 80 Hz (typical Xover F) EDIT: forgot to add the sub is about 0.5 m. further away than the left speaker ( in the corner).

Also an argument against your statement is that not all subs have the same amount of DSP corrections added to it, OR have the same DSP processor speed for that matter. These differences will of course mean that not all subs with DSP are alike in terms of latency.

 

[ZolaIII]

@Holdt I used to do expert assessments of class leading high performance DSP designs from CEVA and Tensilica including programing sets and flexibility. What's put in speakers and sub's is nowhere close to such and it is a sync problem.

 

[Deleted member 48726]

@Holdt I used to do expert assessments of class leading high performance DSP designs from CEVA and Tensilica including programing sets and flexibility. What's put in speakers and sub's is nowhere close to such and it is a sync problem.

I'm sorry, but I don't know what you mean by that?

 

[ZolaIII]

I'm sorry, but I don't know what you mean by that?

Fastest and most flexible DSP designs there are up to date at and over 1 GHz. All do CEVA had better scaling (up to 8 core vs 6) design Tensilica P6 won thanks to better programmable scalability and adjustability. Those are the biggest player's on DSP stage by the way.

 

[Deleted member 48726]

Fastest and most flexible DSP designs there are up to date at and over 1 GHz. All do CEVA had better scaling (up to 8 core vs 6) design Tensilica P6 won thanks to better programmable scalability and adjustability. Those are the biggest player's on DSP stage by the way.

Still don't know how it relates to my post? Maybe I haven't had enough coffee yet.

 

[ZolaIII]

Still don't know how it relates to my post? Maybe I haven't had enough coffee yet.

May be it doesn't have anything directly with your post. I wanted to say that I know a lot better DSP architectures their bottlenecks and how hard they are to programme than most at least. I did evaluation for complex scientific workloads (like simulations and ML) with emphasise put on flexibility for a sake of ending black silicone abuse. To simplify it DSP as multipurpose scalable accelerator for cost sensitive applications where you can't either put nor afford FPGA and GPU's aren't suitable (nor will ever be). One of main point's in minimising latency whose high integration (SoC) design with fast wide RAM like HBM and HMC on multilayer async incoherent bus against VLIW. In other words how much is possible to achieve to eliminate bottlenecks and latency.
Neadles to say you don't get any of those in embedded one's in audio equipment.
Best regards and enjoy your coffee.

 

[fineMen]

Toole's research has convinced him that highly suppressing floor reflections is an overrated priority, and that psychoacoustically a good carpet on a quality underlay provides the important attenuation. So I don't support your bit in bold above.

Thing is, that I don't know of anyone who tested the carpet magic by himself. (I ignore myself in this regard.)

So it is left to the whishful consumer to make sense out of the terms "psychoac/", "carpet", "underlay" in combination, also taken the attributations of "good" and "quality" into account.

From wiki (https://en.wikipedia.org/wiki/Cargo_cult_science):

"An example of cargo cult science is an experiment that uses another researcher's results in lieu of an experimental control."

"So I call these things cargo cult science, because they follow all the apparent precepts and forms of scientific investigation, but they're missing something essential, because the planes don't land."

What people in high end land are missing is exactly that: control. The refusal to just measure a thing themselves, because of the details ... namely the difference between "predicted response2" and actual perfromance in situ. E/g that the in-room response drastically depends on the room's individual design.

There was ONE single person to recommend to buy a microphone in order to evaluate things. Of course he was utterly ignored by all. People might stop to think of themselves as "scientific" if they just and only ventilate "complicated" papers.

 

[garyrc]

Real Dynamic Range

I didn't have time to read all of this long thread, but I hope the following has been reported.

Regardless of the ambient noise level, very high instantaneous peaks (often 1/4 second or less) are important in selections like Mahler or Beethoven symphonies, Copland's Fanfare for the Common Man, The Great Gate of Kiev, and many, many others.

What are these levels from a close seat? These figures are based on my own measurements and also published levels measured by others, almost always at "fast" and C or Z weight:

1. For a full symphony orchestra, instantaneous peaks of 100 dB to 115 dB.
2. A very large symphony orchestra or opera at fff, or ffff, can briefly reach 120 dB, or more. The leading edge of a sfz moment could reach this level. Rachmaninoff invoked sffff in Prelude in C#. Ligeti used ffffffff in Le Grande Macabre.
3. At home, in a 4,000 to 5,000 cu. ft. room, these figures should be lowered by 5 dB, according to THX, due to distortion simulating and loudness imitating very early reflections that don't occur in very large venues like a concert hall or commercial movie theater. With high distortion (like that of the terrible, tiny transistor radios of the 1960s) the music can sound much louder that it really is in dB SPL terms.
4. With a speaker with sensitivity ("efficiency") of 90 dB at 1 watt at 1 meter, heard instead at 13 feet, the lowest peak level in this post (100 dB) would take a little less than 60 watts, but a peak of 117 dB, in the ballpark of the SPLs discussed above, would take 2,000 watts per channel into a 90 dB/1 w/1 M speaker. McIntosh can supply such an amplifier, but most speakers won't take it. Speakers up to 11 dB more sensitive are readily available for $$$
5. One of the most frequently occurring myths concerning hearing damage is that SPLs of around 80 to 85 dBs are dangerous, and can cause hearing loss. What these folk are failing to mention is duration. Here are the figures given by OSHA, and also a more conservative source. Notice that the "dangerous" 85 dB comes into play when you are exposed to it for 8 hours, with the most conservative reckoning (NIOSH). Let's say you heard a piece of music with 10 peaks in the finale -- great punches of sound of 100 dB each, with leading edges of 1/2 second long apiece. Well, you have used up 5 seconds of 100 dB peak time. See chart. Of course, music that is constantly loud, like some Rock, Metal, Electronic, etc. must be considered differently.

To build in a safety factor, and make a more meaningful measurement, don't use dBA -- the A weighting ignores the bass, which is clearly represented in music and movies. Use dBC or, preferably, dBZ and "Fast."

 

[Galliardist]

Thing is, that I don't know of anyone who tested the carpet magic by himself. (I ignore myself in this regard.)

So it is left to the whishful consumer to make sense out of the terms "psychoac/", "carpet", "underlay" in combination, also taken the attributations of "good" and "quality" into account.

From wiki (https://en.wikipedia.org/wiki/Cargo_cult_science):

"An example of cargo cult science is an experiment that uses another researcher's results in lieu of an experimental control."

"So I call these things cargo cult science, because they follow all the apparent precepts and forms of scientific investigation, but they're missing something essential, because the planes don't land."

What people in high end land are missing is exactly that: control. The refusal to just measure a thing themselves, because of the details ... namely the difference between "predicted response2" and actual perfromance in situ. E/g that the in-room response drastically depends on the room's individual design.

There was ONE single person to recommend to buy a microphone in order to evaluate things. Of course he was utterly ignored by all. People might stop to think of themselves as "scientific" if they just and only ventilate "complicated" papers.

There is a difference between conducting scientific experiments, and establishing best practice based on the results.

This place may be called "Audio Science Review", but in threads like this, we are not doing science (and therefore not doing "cargo cult science") but installing appropriate technology for a particular purpose.

There are threads around here where people use other researchers' results in all sorts of different ways. If I was doing an actual experiment in audio, as opposed to say, taking measurements of a device (also not science), discussing someone else's experiment (where the comparisons are going to be to other researchers' results, unless someone is in a position to repeat a valid experiment exactly) or any number of other things that are not conducting an experiment, then I may need a control depending on the nature of the experiment. I don't need a control for most chemistry experiments, for example, unless I am comparing one reaction to another.

This thread is not about an experiment but a selection of appropriate technology. We didn't test what the dogs are capable of knocking over, either.

We can't necessarily use someone else's result in lieu of a control if we are experimenting to find out what a rug on the floor does, but we can use the result in installation.

Consider the alternative. By your assertion, I can't determine, for example, ever, if I need a Category 3 or a Category 6 Ethernet cable for a particular connection until I carry out an experiment to determine which will work, and one with a control. Of course I can simply use the standard and buy a cable that meets the category I need. What has been done in this thread is the same thing.

When you find people here doing science, then judge them as scientists. It doesn't apply to anything done or said here.

 

[Sokel]

Real Dynamic Range

I didn't have time to read all of this long thread, but I hope the following has been reported.

Regardless of the ambient noise level, very high instantaneous peaks (often 1/4 second or less) are important in selections like Mahler or Beethoven symphonies, Copland's Fanfare for the Common Man, The Great Gate of Kiev, and many, many others.

What are these levels from a close seat? These figures are based on my own measurements and also published levels measured by others, almost always at "fast" and C or Z weight:

1. For a full symphony orchestra, instantaneous peaks of 100 dB to 115 dB.
2. A very large symphony orchestra or opera at fff, or ffff, can briefly reach 120 dB, or more. The leading edge of a sfz moment could reach this level. Rachmaninoff invoked sffff in Prelude in C#. Ligeti used ffffffff in Le Grande Macabre.
3. At home, in a 4,000 to 5,000 cu. ft. room, these figures should be lowered by 5 dB, according to THX, due to distortion simulating and loudness imitating very early reflections that don't occur in very large venues like a concert hall or commercial movie theater. With high distortion (like that of the terrible, tiny transistor radios of the 1960s) the music can sound much louder that it really is in dB SPL terms.
4. With a speaker with sensitivity ("efficiency") of 90 dB at 1 watt at 1 meter, heard instead at 13 feet, the lowest peak level in this post (100 dB) would take a little less than 60 watts, but a peak of 117 dB, in the ballpark of the SPLs discussed above, would take 2,000 watts per channel into a 90 dB/1 w/1 M speaker. McIntosh can supply such an amplifier, but most speakers won't take it. Speakers up to 11 dB more sensitive are readily available for $$$
5. One of the most frequently occurring myths concerning hearing damage is that SPLs of around 80 to 85 dBs are dangerous, and can cause hearing loss. What these folk are failing to mention is duration. Here are the figures given by OSHA, and also a more conservative source. Notice that the "dangerous" 85 dB comes into play when you are exposed to it for 8 hours, with the most conservative reckoning (NIOSH). Let's say you heard a piece of music with 10 peaks in the finale -- great punches of sound of 100 dB each, with leading edges of 1/2 second long apiece. Well, you have used up 5 seconds of 100 dB peak time. See chart. Of course, music that is constantly loud, like some Rock, Metal, Electronic, etc. must be considered differently.

To build in a safety factor, and make a more meaningful measurement, don't use dBA -- the A weighting ignores the bass, which is clearly represented in music and movies. Use dBC or, preferably, dBZ and "Fast."

Let's see it visually.

Brandenburg Concerto No.3 in G Major by J.S. Bach

is that what you purpose?

 

[excelsius]

Real Dynamic Range

I didn't have time to read all of this long thread, but I hope the following has been reported.

Regardless of the ambient noise level, very high instantaneous peaks (often 1/4 second or less) are important in selections like Mahler or Beethoven symphonies, Copland's Fanfare for the Common Man, The Great Gate of Kiev, and many, many others.

What are these levels from a close seat? These figures are based on my own measurements and also published levels measured by others, almost always at "fast" and C or Z weight:

1. For a full symphony orchestra, instantaneous peaks of 100 dB to 115 dB.
2. A very large symphony orchestra or opera at fff, or ffff, can briefly reach 120 dB, or more. The leading edge of a sfz moment could reach this level. Rachmaninoff invoked sffff in Prelude in C#. Ligeti used ffffffff in Le Grande Macabre.
3. At home, in a 4,000 to 5,000 cu. ft. room, these figures should be lowered by 5 dB, according to THX, due to distortion simulating and loudness imitating very early reflections that don't occur in very large venues like a concert hall or commercial movie theater. With high distortion (like that of the terrible, tiny transistor radios of the 1960s) the music can sound much louder that it really is in dB SPL terms.
4. With a speaker with sensitivity ("efficiency") of 90 dB at 1 watt at 1 meter, heard instead at 13 feet, the lowest peak level in this post (100 dB) would take a little less than 60 watts, but a peak of 117 dB, in the ballpark of the SPLs discussed above, would take 2,000 watts per channel into a 90 dB/1 w/1 M speaker. McIntosh can supply such an amplifier, but most speakers won't take it. Speakers up to 11 dB more sensitive are readily available for $$$
5. One of the most frequently occurring myths concerning hearing damage is that SPLs of around 80 to 85 dBs are dangerous, and can cause hearing loss. What these folk are failing to mention is duration. Here are the figures given by OSHA, and also a more conservative source. Notice that the "dangerous" 85 dB comes into play when you are exposed to it for 8 hours, with the most conservative reckoning (NIOSH). Let's say you heard a piece of music with 10 peaks in the finale -- great punches of sound of 100 dB each, with leading edges of 1/2 second long apiece. Well, you have used up 5 seconds of 100 dB peak time. See chart. Of course, music that is constantly loud, like some Rock, Metal, Electronic, etc. must be considered differently.

To build in a safety factor, and make a more meaningful measurement, don't use dBA -- the A weighting ignores the bass, which is clearly represented in music and movies. Use dBC or, preferably, dBZ and "Fast."

Interesting about the 5dB lower figure at home (presumably compared to anechoic space). If you have a citation for that, I'd like to add it to my notes for future reference. Even better, would be curious about how they calculated it so it can be tuned to volumes other than 4K ft^3.

Couple points, per my novice calculations about the bullet point 4:

• To achieve 117 dB at 13ft (~4m) with a sensitivity of 90 dB (i.e., at 1m) would take 8000W for 1 speaker in open space. I.e., must produce 129 dB at 1m
• But for two speakers, additive effects drops down to 4000W per speaker (3dB less, i.e., 126 dB at 1m)
• Still, these values are for open space. To compensate for lack of loss in a closed space, here using the -5dB value you mentioned from THX, this results in 1250W per speaker (121 dB at 1 m)

I think your figure of 2K W per channel is about correct (would be closer to 2500W), but maybe it doesn't take into account the extra sound added by the second speaker, which brings it down to 1250 W/channel. Of course, these are still pretty high values. I still have no idea how does one determine how much power can a certain speaker take before mechanical damage.

 

[excelsius]

Let's see it visually.

Brandenburg Concerto No.3 in G Major by J.S. Bach


View attachment 274868

is that what you purpose?

What software is that?

 

[excelsius]

Quote from Neumann about KH-150:

The maximum level that a monitor can reach is primarily determined by the drivers and the available amplifier power. For the drivers, the natural limitation is based on the one hand on the maximal excursion of the diaphragm, which is particularly true for woofers, and by the thermal load capacity of the voice coil. With that in mind, all the relevant states are monitored in the KH 150’s DSP system and, if necessary, limited through the use of limiters. Specifically, these limiters include power amplifier clip limiters for both paths, thermo limiters for the drivers, an excursion limiter for the woofer and another limiter that prevents the power supply from overloading.

I also own a Wiim Pro streamer which is very reliable and enjoyable to use. I can verify that you can control the digital output volume using the remote. The remote is an optional accessory so make sure to order one.

By the way, I went through the WiiM Pro manual and it says that when the volume is not set to fixed, the output is not bit-perfect. I'm not sure how critical this is, but it seems to be the trade-off for being able to control the volume with WiiM.

 

[Sokel]

What software is that?

Room Eq Wizard (REW)

REW room acoustics and audio device measurement and analysis software

REW is free software for room acoustic measurement, loudspeaker measurement and audio device measurement and analysis

www.roomeqwizard.com

 

[excelsius]

Room Eq Wizard (REW)

REW room acoustics and audio device measurement and analysis software

REW is free software for room acoustic measurement, loudspeaker measurement and audio device measurement and analysis

www.roomeqwizard.com

I'm assuming you're using it with UMIK-1 or something similar. Presumably I can get similar sort of graphs with the MA-1 from Neumann, which would be cool to see. But I'm not sure if it's so automated that maybe it doesn't display the graphs like that.

 

[fineMen]

There is a difference between conducting scientific experiments, and establishing best practice based on the results.

We can't necessarily use someone else's result in lieu of a control if we are experimenting to find out what a rug on the floor does, but we can use the result in installation.

When you find people here doing science, then judge them as scientists. It doesn't apply to anything done or said here.

I actually appreciate the reminder. This is about--engineering, based on science. Not every decision is necessarily made on scientific grounds. E/g, the safety margin one needs to sleep well. Things are done just "for good measure" once in a while. But this opens the field into over-egineering. One could argue that cables of exaggerated thickness, least skin effect etc are chosen for that reason.

As long as the causality is not rediculed (molecule dynamics in speaker cones) this may be in order.

But what about the causality of scientifically found 'on average' preferrence for certain speakers and the personal pleasure of an individual listener at his home? I don't want to detail out my caveats here again and again ad nausea.

Regarding (a) beryllium tweeters, (b) wide directivity or (c) that rug, all recommended here, I'm tempted to get into the specifics. But I won't.

Again, thanks for the reminder! Only, please, say engineering, not 'psychoacoustics'.

 

[Norcal]

OP, maybe I missed it somewhere, but if the KH150 accepts digital input via SPDIF coax, why would you ditch the Bluesound Node in favor of the WiiM Pro? The only thing 'wrong' with the Node is the subpar DAC, but using digital out of the Node circumvents the Node's DAC and that passed through digital signal was measured to be transparent. If you are used to the Bluesound software and are happy with the UI, you could save $150 for now and stick with it.

 

[Galliardist]

Again, thanks for the reminder! Only, please, say engineering, not 'psychoacoustics'.

I used neither term.

I also don't think that chasing absolutes is the first goal in purchasing a first or any other system. I've always advocated for subjectively neutral sounding systems, but had the OP chosen the B&W speakers, say, I'd have been just as happy.