Loudness etc: Tempest in a Teapot

Chat room for MikeC's Audio Craft, music & sound reproduction, streaming, audio PC, etc.
Post Reply
glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Loudness etc: Tempest in a Teapot

Post by glyphin » Tue Jan 05, 2016 4:51 pm

Sweet build Mike. About how your system sounds at low volumes - this is the tradeoff that happens with Class AB or D amps - distortion for these amp topologies is mostly constant across power output, which allows low distortion % at high power, at the cost of higher distortion % at low power. To do better at low volumes you have to go to Class A, unfortunately, which means much higher power consumption, and isn't practical for multi-amped designs.

Also, Linkwitz-Riley filters do crazy things to the phase of the signal (360 degrees rotation per octave even in the pass band), which some people find audible as a loss of intelligibility (including myself). LR filters, particularly active line level LR filters, allow the drivers to work comfortably within their limits by sharply attenuating power outside of the pass band, so again this design improves high volume performance at the cost of low level intelligibility.

There's no perfect speaker system, so I say enjoy your system for rock music and home theater. If you want better performance for jazz music and small ensembles at lower volumes, try building a second system with single driver speakers or first order crossover multi-driver speakers with class A amps, even tubes, and listen near-field.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Linkwitz Orion speaker project

Post by MikeC » Tue Jan 05, 2016 6:00 pm

glyphin wrote:Sweet build Mike. About how your system sounds at low volumes - this is the tradeoff that happens with Class AB or D amps - distortion for these amp topologies is mostly constant across power output, which allows low distortion % at high power, at the cost of higher distortion % at low power. To do better at low volumes you have to go to Class A, unfortunately, which means much higher power consumption, and isn't practical for multi-amped designs.
I doubt this very much, at least for me. It's not the amp. I don't like the way any hi-fi system sounds at low volume. And I've listened to more systems than I care to count or remember. They all sound wrong & unnatural when played softly. I almost always prefer silence over quiet music w/ 65 dB peaks when the original peaks were 30 dB higher. Just turn it off, I say... unless my focus is entirely off the music -- like at a social event. (Besides, even if I was hearing higher distortion at low power, if my average volume dropped by 15 dB, even a 10x increase in distortion would be negligible. ie, 0.1% at 50W vs 1% distortion at 2W.)
glyphin wrote:Also, Linkwitz-Riley filters do crazy things to the phase of the signal (360 degrees rotation per octave even in the pass band), which some people find audible as a loss of intelligibility (including myself). LR filters, particularly active line level LR filters, allow the drivers to work comfortably within their limits by sharply attenuating power outside of the pass band, so again this design improves high volume performance at the cost of low level intelligibility.
I'm not highly sensitive to phase changes, I know from having done a bunch of controlled testing at various times over the years. Change the pitch of a string by >1% on a good guitar when all the rest are in perfect tune and I can hear that instantly, but both amplitude & phase changes in small bands of the audio spectrum are much less audible to me (and to lots of others I've shared these "tests" with). In fact, I can honestly say that most people (including me) cannot hear a 1 dB change centered at say 1 kHz (with a Q of perhaps .7) when music is being played. A 2 dB change I can probably hear.... but I may not care one way or the other, again depending on the music. I can easily hear a piano that's a bit out of tune, and this can bug me. Ditto the mid-high freq ringing (of the mic, the cutter head, preamp circuits or whatever the cause) in many Ella Fitzgerald recordings, including unfortunately, some her best Cole Porter performances.
glyphin wrote:There's no perfect speaker system, so I say enjoy your system for rock music and home theater.
Agreed re- perfect speakers, but I hardly ever play rock, and these speakers are in an audio-only system. The TV on the wall is only used as a monitor (OK, once in a while, I might watch a youtube or other online video... but often not music).
glyphin wrote:If you want better performance for jazz music and small ensembles at lower volumes, try building a second system with single driver speakers or first order crossover multi-driver speakers with class A amps, even tubes, and listen near-field.
Again, I have to disagree. :lol:
1. I have listened to single driver systems, and haven't liked them; they sounded bandwidth limited and/or too non-linear (yeah, I could hear their much less than linear freq response). But I haven't heard that many, and not the best, so I withhold any blanket judgements. I suspect I won't like their limitations.
2. I do not like nearfield listening at all, even with famous BBC studio monitors. The experience seems completely unnatural to me. Latest experience was in a recording studio a month ago where I was helping out a young band. The control room had some expensive nearfield active powered monitors atop the console, maybe 1m away, that were supposed to be the cat's meow (can't recall brand)... and they sounded artificial, not very true to the sound I was hearing in the studio where the band was playing. (PS -- In contrast, I have no problem with headphones, at least with most music, tho I mostly prefer open-back as opposed to closed.)
3. My Orion system mostly plays jazz (both traditional & modern/electronic) lounge music, vocals backed by relatively small numbers of instruments. A bit of folk, even borderline country at times. Not much that's heavily orchestrated. And the system sounds fantastic with all of these.

To be perfectly clear, live jazz music, even a very small unamplified ensemble (which usually means a small venue) will hit 100 dB peaks. A single violin easily hits 90 dB in my living room, almost effortlessly. My daughter, who is a pop singer, can reach well above 90 dB with her unamplified voice in my living room. All this from ~2m away, often farther. I know, mostly because I've measured the SPL levels (& spectrum) of the music myself. Most of my hifi listening is done with peaks reaching 85~95 dB at about 2m distance, depending on what/who is playing. This is rarely as loud as the original performance, but it's not that far off, perhaps down by 10 dB (or more). I often listen quite far away from the speakers -- more than 4m.

glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Re: Linkwitz Orion speaker project

Post by glyphin » Wed Jan 06, 2016 11:46 am

If you like headphone sound then that supports my contention. Most headphones are essentially near-field single driver speakers. It's also much more feasible to get a high quality class A headphone amp than a speaker amp. Indeed, headphone frequency response is terrible - but the merits of a time/phase-coherent sound outweigh their FR anomalies. Headphones also tend to sound better than single driver speakers because the headphone driver doesn't have to work as hard to generate the same SPL, being positioned right next to your ears.

Near-field monitors for recording studios are designed for sound engineers to hear into the mix, not for listening enjoyment. They tend to use high-order crossovers to enable full dynamics and are balanced with a flat frequency response for anechoic/heavily sound-treated control rooms so the engineer can pick up on all the details. This ends up being a fatiguing sound. Home speakers tend to be softer in the treble since not many recordings have high frequencies worth listening to.

If you have the opportunity, try to audition these speaker brands that use first-order crossovers: Vandersteen, Reference 3A, Thiel, Quad ESL, Rethm, Voxative, Dunlavy (out of production), Meadowlark (also dead). Class A amps include Pass Labs, Plinius, Atmasphere, Accuphase, Berning, and a bunch of more boutiquey single ended triode tube amps. With a combination of these speakers and amps you may find lower-volume listening quite enjoyable. Just don't expect them to play as loudly as higher-order crossover speakers with class AB amps, for a given size or cost of speaker and amplifier.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Linkwitz Orion speaker project

Post by MikeC » Wed Jan 06, 2016 5:11 pm

glyphin --

Funny how you picked up on the minor comment re headphones. Here's another: I much prefer listening to my speakers. Headphones used only when necessary -- ie, late at night, traveling -- and often with lesser sources.

You miss my point -- I have NO interest whatsoever in listening to music at low levels. Sure, I can have things on in the background at low level, but then I hardly care what it sounds like, because I'm not paying attention. My system sounds about the same as any high quality hifi system at such volume. Perfectly listenable and ultimately unrealistic & unnatural.

I totally disagree with you re-audibility of distortion in good amps of any class at low levels -- the signal is already muted, and the distortion is at a level far below the signal, which makes it completely inaudible. It's more or less the same argument I have with people worrying about 1% efficiency differences in computer PSUs at 20~30W idle power -- the difference in real watts is insignificant (ie, 1% of 20W = 0.2W). I also take issue with the notion that Class AB and D amps have so much higher distortion at low power levels (which is insignificant due to inaudibility anyway) -- it may be higher than at near maximum output, but is rare to see it climb even one order of magnitude. The best Hypex amps, for example (NC400), have a THD+Noise of something like 0.001% at 200W, and it rises to perhaps 0.01% at 10mW. Anyone who thinks they can hear the latter when this amp is playing music through a loudspeaker is misguided. In fact, in the bass, THD of several %, upwards of 10%!, is often not that easy to hear, it depends a lot on the music.

I have heard, listened to and even taken apart things like Vandersteens, Quads, etc. I've had several stints between 1975 to 1990 in audio sales, and been an active enthusiasts since 1968. No interest in hearing them all again, certainly not enough to go out of my way for that. My reference is live music, both unamplified and amplified. (PS -- Quad ESLs are very nice, the first real high end spkrs I heard, all the way back around 1970, in Bangkok. They can't play very loudly, have limited bass, and a fairly small sweet spot. If you can accept the limitations, great, but they are not for me.)

Finally, despite my long interest in music, audio gear, etc, I am most definitely NOT an "audiophile".

glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Re: Linkwitz Orion speaker project

Post by glyphin » Wed Jan 06, 2016 6:07 pm

Such strong convictions, Mike! I understand you like music at live levels and prefer to listen to it at live levels. All I'm suggesting is your preferences might be affected to some extent by the equipment.

I play multiple classical musical instruments. My daily amps are in fact NC400 monoblocks. And I listen to class A tube and solid state amps regularly, with first-order or crossover-less speakers. I can hear the difference. NC400 is good, but class A is better, particularly at low level details, often described as 'warmth' and 'ambience'. And to me, the less phase distortion, the more natural I perceive the sound to be, all else being equal. I have performed ABX tests on myself on the audibility of phase distortion (specifically, 4th order Linkwitz-Riley) and performed better than p<.01 - as part of a college course on psychoacoustics, which I earned an 'A' in, I might add. A simple test I've applied at home is to use a digital parametric equalizer operating in the source digital domain (in the PC at 64-bit) to flatten my in-room frequency response. Some people may prefer the sound with the flatter frequency response, but I can hear the phase distortions introduced, and prefer the unequalized version.

Nelson Pass on why single ended class A:
http://www.firstwatt.com/pdf/art_se_classa.pdf

and on audio distortion and feedback:
https://passlabs.com/articles/audio-dis ... d-feedback

Calling someone 'misguided' because they perceive things differently is a bit rich for someone who operates a website extolling the virtues of PCs that output 20 dB of noise as opposed to 30 dB of noise where the typical residential noise floor is 40 dB!

Now, when I want to rock out, the NC400 is just fine. I shudder to think at the power bill for 200 watts of class A (2000 watts of constant power draw compared to the 5 watts idle for the NC400).

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Linkwitz Orion speaker project

Post by MikeC » Wed Jan 06, 2016 10:10 pm

I bow to your superior hearing perception. You must be one in a million.

What can I say? To me, no matter what the system, turn it down too low and there's just no point. And it's not just the equipment I have today; I've had a huge array of stuff over the past 40 years -- including class A (Classe, etc). About the only kinds of music I like quiet is when it was recorded quietly. Like a sad quiet singer unaccompanied.

To put the amp distortion issue in context, speaker drivers have a level of distortion 2 or 3 orders of magnitude higher -- possibly much more -- than good electronics.

This is not to say there aren't audible difference between amps, I hear them & appreciate them at times. But there's a lot of other things that impact the sound way more. Like whether the drapes are drawn, the windows open, or the main mic was often overloaded during the original recording.

What I'm saying is when the music is at 80 dB, and the distortion components are 70 or 80 dB down, I KNOW I can't hear it. I've done the testing.

I'm not so clear on phase distortion, despite having done some experimentation with crossovers, both passive & active. Some people are less sensitive to this than others... tho it's audible & better when drivers are in phase in the xover transition band. Will be experimenting more with my miniDSP device & another pair of open-baffle speakers in the pre-assembly stage now. I have all the parts, including Heil open-back tweeters. It's a blend of Linkwitz LX521 & Nao Note II RS designs + plywood/aerolam/plywood sandwich panels for the main baffles. Just need time and warmer weather to execute.

As for 20 vs 30 vs 40 dBA (@1m) of noise, the differences are totally & unequivocally audible in my home. Daytime ambient is probably around 25~30 dBA, but the noise of computers, HDDs, fans and other machine noises are still audible even when below the ambient. It's always more tonal, and fundamentally different from natural sounds like wind. We (SPCR) generally think of 30 dBA as being too loud, 20 as being very quiet, below 15 as being effectively inaudible.

TBH, the biggest "distortion" in my system right now is the audio PC. Its PSU fan has gone south, and rattling/buzzing constantly. Probably 25 dBA@1m. Need to deal with it soon... but meanwhile, I turn up the volume another notch or 2. :lol: :lol:

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Linkwitz Orion speaker project

Post by MikeC » Thu Jan 07, 2016 9:23 pm

glyphin --

This discussion w/ you made me wonder about my SPL comments. The last time I had the SLM set up in the LR was like a year ago, and max SPL was not what I was checking. So...

This evening, I decided to do some checks on my listening habits, and I'm afraid I've exaggerated the volume levels. Here's what I measured today:

1. High volume -- I enjoy this best, but can't or don't do it for that long, perhaps 90 minutes at most. The peaks don't quite reach 90 dB @ 2m.

2. Medium -- my wife prefers it here. Typically, max peaks are around 75~80 dB.

3. Fairly low -- still too loud for dinner conversation with guests, with peaks not reaching 60 dB. I checked my amp AC power draw here and found it never went above 30W.

At all levels, clarity is pretty amazing, w/ great detail, but music recorded loud definitely doesn't sound natural when played softly.

It would be interesting to try a class A amp on these speakers, but w/ 8 ch (or 6 min), it's no easy thing to do. The fact of multi-band electronic xover spkr systems is that the load on each amp is way easier than any passive xover system with a single amp/ch per side. My amps only deal with 20~100Hz, 100Hz~1500Hz, and 1500Hz on up. Very narrow band, with no inductors or caps between the amp outputs and the drivers. Active spkr systems tend to minimize the differences between amps, afaik, because the load is so simple.

glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Re: Linkwitz Orion speaker project

Post by glyphin » Thu Jan 07, 2016 10:52 pm

MikeC wrote:It would be interesting to try a class A amp on these speakers, but w/ 8 ch (or 6 min), it's no easy thing to do. The fact of multi-band electronic xover spkr systems is that the load on each amp is way easier than any passive xover system with a single amp/ch per side. My amps only deal with 20~100Hz, 100Hz~1500Hz, and 1500Hz on up. Very narrow band, with no inductors or caps between the amp outputs and the drivers. Active spkr systems tend to minimize the differences between amps, afaik, because the load is so simple.
Indeed, active speakers are where the benefits of class A would be most apparent. The narrow bands of frequencies going to each driver, combined with the steep slope of the crossovers, severely limits the amount of power needed to drive each driver to a given SPL. So in effect your amps are gaining 2-4x the effective power compared to if they had to drive a full range system. This means you could get away with just a handful of watts per driver and still have plenty of headroom. This is equivalent to typical pairings of single-ended triode amps with high efficiency single driver speakers.

For example, the Seas W22EX-01 is 88 dB/1w/1m. It's only being asked to drive less than half the frequency range, so you're already gaining a doubling in efficiency. This makes the driver 3 dB more efficient, so 91 dB/w/m. In order to reach peaks of 75 dB at 4m, you would need only 87 dB at 1m from the driver, so the driver would need less than half a watt at its peak. This opens up many lower powered/lower cost class A amp options. I'm currently intrigued by the Berning MicroZOTL, which cranks out a full watt per channel of power (plenty for headphones), and is reasonably power efficient using an switching power supply and no output transformer - idles at 33 watts per 2 channels and costs $1100. Another affordable option at $1k is the Decware Zen triode amp with 2 watts per channel - double the power! but also double the power consumption. A transistor amp that has a good reputation for 'first watt' performance is the Gainclone, which you can DIY for as little as a couple hundred bucks in parts. It's a class AB amp but the LM1875 spec sheet shows falling distortion below 1 watt. I built the LM3875 and paired it with my 95 dB open baffle crossoverless line arrays and enjoyed it quite a bit.

glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Re: Linkwitz Orion speaker project

Post by glyphin » Thu Jan 07, 2016 11:05 pm

MikeC wrote: At all levels, clarity is pretty amazing, w/ great detail, but music recorded loud definitely doesn't sound natural when played softly.
Aside from my contention that lower level music listening benefits more from class A amplification and is where time/phase coherent speakers can shine, one of the basic psychoacoustic effects is the Fletcher-Munson equal loudness contours. At lower volumes more bass SPLs are necessary relative to midrange SPLs to be perceived as equally loud. Conversely, your 'music recorded loud' 'when played softly' will sound midrange dominant, losing perceived detail in bass and treble. This is what the old 'bass boost' function on receivers was for. You can try controlling your volume per pair of drivers separately as a crude kind of equalizer - have bass drivers turned up a little higher relative to midrange when overall volumes are low.

HammerSandwich
*Lifetime Patron*
Posts: 1288
Joined: Sat Oct 25, 2003 3:21 pm
Location: 15143, USA
Contact:

Re: Linkwitz Orion speaker project

Post by HammerSandwich » Fri Jan 08, 2016 9:16 am

MikeC wrote:1. High volume -- I enjoy this best, but can't or don't do it for that long, perhaps 90 minutes at most. The peaks don't quite reach 90 dB @ 2m.
Good thing you're working on those 400W amps! :D
glyphin wrote:For example, the Seas W22EX-01 is 88 dB/1w/1m. It's only being asked to drive less than half the frequency range, so you're already gaining a doubling in efficiency. This makes the driver 3 dB more efficient, so 91 dB/w/m.
Where can I find the math on this? My brain's hung up on amplitude being a function of voltage.

glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Re: Linkwitz Orion speaker project

Post by glyphin » Fri Jan 08, 2016 9:59 am

Here's a good article with a chart:
http://www.gcaudio.com/resources/howtos ... dness.html

10 dB change is only a perceived doubling in volume - such is the logarithmic nature of dB. So an amp with 10x the power only lets you get twice as loud. And 3 dB is the generally accepted 'just noticeable difference' for normal people who aren't sound engineers.

SPL decay over distance is more complicated and depends on the room. I used this calculator:
http://www.sengpielaudio.com/calculator-distance.htm

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Linkwitz Orion speaker project

Post by MikeC » Fri Jan 08, 2016 10:27 am

HammerSandwich wrote:
MikeC wrote:1. High volume -- I enjoy this best, but can't or don't do it for that long, perhaps 90 minutes at most. The peaks don't quite reach 90 dB @ 2m.
Good thing you're working on those 400W amps! :D
glyphin wrote:For example, the Seas W22EX-01 is 88 dB/1w/1m. It's only being asked to drive less than half the frequency range, so you're already gaining a doubling in efficiency. This makes the driver 3 dB more efficient, so 91 dB/w/m.
Where can I find the math on this? My brain's hung up on amplitude being a function of voltage.
lol!

I'm working towards a house mod that will take out the dividing wall between the LR/DR, which is around 30-33' long, and the kitchen, which is about 20' long. That will mean a ~50' open path from the speakers to the back french doors, which opens to a big patio where we like to hang out whenever weather is good. That's what the high volume capability is really for. :lol:

10 dB would be quite a reach -- would be great if I could get it. Twice is loud is about what I would love to get. I'm not sure if I'll get it, at least not with all music. The bass drivers will probably hit their excursion limits before 400W.

Anyway, I might be able to find out whether the upper volume limit to comfortable listening in my system is the speakers, the amps, the room or me. :lol:

Regarding spkr efficiency, it's usually specified as SPL@1m per 1W (usually 2.86V) of input at 1kHz... which is a narrow definition; white noise would be a better source, and the correct voltage to reach 1W with whatever driver would be better, too. The standard applies for a speaker which measures 8 ohms at the 1kHz freq... but the same voltage is often used for 4 ohm spkrs, which exaggerates the sensitivity even more. An audio buddy, another Mike, and I have played around with enough drivers to come to the same conclusion: Manufacturers' SPL sensitivity specs are usually optimistic by 2-3 dB. Glyphin's comment is incorrect, btw -- sensitivity does not work like that.

And no, I will not build 6 or 8 single-ended triode amps to test your theory, Glyphin. I'm trying to REDUCE the footprint of my gear, not increase it!

I'm at a loss about how to explain to Glyphin my perspective on quiet music sounding wrong. I'm not sure whether others also don't get it, as there haven't been any other comments. Let me try again:

It really has NOTHING to do with equipment. When I wrote "music recorded loud definitely doesn't sound natural when played softly", I did not mean that MY system sounds especially unnatural at low volume. I meant that ALL music that was played loud when recorded sounds unnatural when played softly, regardless of system, even if reproduced perfectly at say 30 dB lower volume. In other words, SPL is one of the essential keys to authentic high fidelity audio reproduction.

Now, we don't always want such authenticity; our LRs cannot always be turned into a jazz club or music hall -- there are other family members, neighbors, etc who'd object to such a thing in their midst. But to say you have high fidelity if everything is reproduced perfectly at 30~40 dB below the actual recorded level isn't right. That's more like a perfect miniature model, which is far less involving, interesting or compelling than the original. Again, not everyone is interested in trying to reproduce the original, but I am -- it is my hobby. I'm sure I'm far from achieving it, I'm convinced it won't happen routinely in my lifetime, but I get close enough sometimes with the right source material that I keep tinkering & trying, which is more than half the fun! 8) :wink:

Here's an example of what I mean -- I had Spotify on pretty quietly (max 60 dB peaks, methinks) the other day, letting a semi random stream on my HD account (320 kbps, which is most often indistinguishable from 16/44.1 WAV - CD quality). It's one of the ways I seek out new music/musicians. A male singer who sounded vaguely familiar came on performing what could have been an early Led Zeppelin blues/rock piece. He was really belting it out. Turns out it was Tom Jones, who in his mid-'70s, has remade himself into blues/R&B type of singer, quite appealing to me -- tho I didn't think much of the music that made him famous. I could picture him doing his emotional full-tilt delivery -- but there's such a disconnect between that intensity and reproduction at 60 dB peak levels.... that I had to turn it up. This music had to be turned up to sound right.

In the same way, I would not like to listen to Beethoven's 9th limited to 60 dB peaks. Or Steps Ahead's electrifying Radioactive (from Modern Times album). Or any number of amazing musical performances that would be diminished or compromised by playing at low volume.

glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Re: Linkwitz Orion speaker project

Post by glyphin » Sat Jan 09, 2016 5:50 pm

Sounds like Mike needs speakers that can rock out. Nothing wrong with that - lots of people like to rock out. But to say that all music sounds unnatural unless it's played at live levels is pushing a rather absolutist definition of 'natural' on everyone. Essentially, Mike is claiming his subjective preference is an objective truth. It's not an objective truth. Many people derive full enjoyment from their music without making their walls shake, or listening from four rooms away. Sure ideally, we would all love to have an audio reproduction that recreates the sound field of the recording venue with exactly the same SPLs and same time/phase relationship of waveforms. But audio reproduction in the real world is very far from this ideal. There are tradeoffs in audio system design to allow for high SPLs, and there are plenty of people like myself who prefer not to make those tradeoffs around phase and small signal distortion.

Mike emphasizes his point:
I have NO interest whatsoever in listening to music at low levels. Sure, I can have things on in the background at low level, but then I hardly care what it sounds like, because I'm not paying attention. My system sounds about the same as any high quality hifi system at such volume. Perfectly listenable and ultimately unrealistic & unnatural.
Mike believes he has no interest in listening to music at low levels because the SPLs aren't high enough, because SPLs lower than the source recording can never sound engaging. I'm offering an alternate explanation that Mike has no interest in listening to music at low levels because the tradeoffs in his system design make it sound unengaging at low SPLs - namely, the detrimental effects of phase distortion from high order crossovers and higher amplifier distortion relative to signal due to the circuit topology involved. Conversely, I assert that there are plenty of audio systems that do sound engaging at low levels - those with minimal phase speaker crossovers and class A single ended amplifiers - at least to some people. This is not a novel assertion in the audio world, and if Mike can't acknowledge that these preferences exist, then there's not much more to say.

My comments about sensitivity were a simplification for the sake of the calculation for necessary amplifier power for a given SPL at the listening position. If you prefer, you could say that the amplifiers are effectively twice as powerful - a 100 wpch amp into an actively filtered driver passing only half the frequency range will be able to generate the same SPLs as a 200 wpch amp driving the full frequency range. Satisfied, Mike?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Linkwitz Orion speaker project

Post by MikeC » Sat Jan 09, 2016 6:05 pm

glyphin -- it's like you didn't read a word of my last post.

I'm not trying to push anything on anyone. I'm saying amplitude -- along with everything else -- is part & parcel of audio high fidelity, although it's tough to achieve for many reasons, and not always desirable.

This exchange is getting silly. enuf already. :roll:

And no, playing in a narrower bandwidth does not magically increase speaker sensitivity. This is nonsense. Even the standard sensitivity test is done with a narrow band signal!

glyphin
Posts: 21
Joined: Fri Nov 08, 2002 1:08 pm

Re: Linkwitz Orion speaker project

Post by glyphin » Sun Jan 10, 2016 12:21 am

Yes, clearly you have your own interpretation of what is audio high fidelity and you're unwilling to entertain different conceptions of what could be high fidelity.

You're so busy trying to denounce me as wrong that you did not read my last post about achievable SPL with a given amount of amplifier power with bandwidth-restricted signal vs full range signal. This is a poor discussion.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Loudness etc: Tempest in a Teapot

Post by MikeC » Sun Jan 10, 2016 9:40 am

Woah! Take a step back...
Yes, clearly you have your own interpretation of what is audio high fidelity and you're unwilling to entertain different conceptions of what could be high fidelity.
I've never insisted my definition of HiFi is the only one; imo, it's the same as any other musicophile might have, with an added emphasis on amplitude fidelity -- and I certainly never excluded precise reproduction at lower levels. And in my long post above, I made full disclosure of the various levels I listen at, and they certainly include lots of low level listening -- of music generally recorded at quieter levels -- which I enjoy.

You've been insisting I like my loud music loud because my equipment distorts more at low levels. How can you say this with any conviction without having listened to my system -- at either high or low level? Lots of assumptions on your part. It seems to me that for some reason, you need to find flaws in my system that compel me to turn the volume up.
You're so busy trying to denounce me as wrong that you did not read my last post about achievable SPL with a given amount of amplifier power with bandwidth-restricted signal vs full range signal. This is a poor discussion.
Again, you take this so personally -- denounce you as wrong? Really? The last long post I made was trying to express my reasoning for wanting to hear loud music at loud levels.

And it's true not all music is loud, but a lot of it is pretty loud, including a lot of the music I like.

As for your original statement on speaker efficiency/sensitivity, "It's only being asked to drive less than half the frequency range, so you're already gaining a doubling in efficiency. This makes the driver 3 dB more efficient, so 91 dB/w/m." you're really wrong on that, at least the way you expressed it. Speaker sensitivity has a very simple definition which I won't repeat again, and though I think it's not very useful because it's a poorly defined spec, your claim that it is somehow "boosted" by 3 dB when operated narrow band made no sense at all.

Now your last words, "achievable SPL with a given amount of amplifier power with bandwidth-restricted signal vs full range signal" say something different altogether, which is that the amplifier should be able to play to a higher level before clipping when the bandwidth is limited -- and this is likely true with most music. I agree that my 40Wx8 amp powering each of the 8 drivers directly w/ limited bandwidth signals can play louder than a 160Wx2 amp into the same 8 drivers with passive xovers between them.

----------------------------

Finally, any discussion of music dynamics (which is one of main issues we've been circling, methinks) has to take into account the sources, by which I don't mean the source components but the actual recordings themselves -- which make more difference to audio high fidelity than virtually anything else in the playback system. Dynamic compression during the recording/mixing process is pretty much ubiquitous, there are umpteen sources that document this. It's worst in pop, top-40, etc and occurs least in classical music, especially smaller scale classical. There are pockets of hifi recordings in jazz, folk and perhaps traditional music done with minimalist techniques (2 top quality mics, high quality electronics, no eq, no compression) but these do not reflect the mainstream.

Perhaps you listen more to classical music, which tends to have less compression than other genres, and I listen mostly to jazz, blues, R&B and a smattering of pop -- which tend to have more compression. The music I listen to most often includes amplification in the original performances. This may be significant for the different playback level preferences we have.

OTOH, I have a bunch of really dynamic, minimal manipulation recordings of both jazz and blues, and it is with these recordings that I enjoy turning up the volume the most. The ease with which my current system scales up a projection of the live event is quite delightful with such recordings. It's not possible to say that I'm transported to the original venue, or that the musicians are here in my room, but subjectively, it seems like a mix of the two -- the acoustic of the original recording space are clearly audible, yet there's no question that I can hear the acoustics of my LR as well.

This is an excellent exposition on the topic of dynamics & compressions -- and includes a good audio demo: http://www.parallelhomeaudio.net/PAInDepth.html

mlee
Posts: 1
Joined: Sun Jan 10, 2016 11:02 am

Re: Loudness etc: Tempest in a Teapot

Post by mlee » Sun Jan 10, 2016 11:31 am

Hi All,

Was intrigued by this discussion title so I just signed up. I'll have to do an introduction post here shortly but wanted to comment on my experiences with the points brought up in this post.

Roughly 6 years ago I embarked on a mission to design a pair of speaker that played well at all volumes/SPL levels. Something that could do it all nicely. Rock out, play vocal jazz, small and large ensemble. Well at everything but Master of none as I don't think there is a speaker topology that supports perfect.

Try 1: Hi-fi drivers, seas, vifa ect. Mostly 5-6.5" drivers to tweeter and sub topologies. All sounded smooth, bad at low volumes, sometimes not integrated well at low volumes. Turned volume up and everything clicked better, but fell apart at high volumes and didn't have the dynamics I liked.

Try 2. Added 10" hi-fi woofer to above try 1 and went active. A bit better at low volumes but still most of the problems of try 1.

Try 3. A number of fullrange drivers (mostly Mark Audio and Fostex). Dynamically limited, dispersion problems, a sound I did not like but very very good with frequency limited vocal performances. no high SPL.

Try 4. Econowave style pro woofer 15" with horn top. Hard to integrate, a bit shaky in the 600 to 1100hz range but....Dynamics in spades, great at lower volume, great at higher volume. Good all rounder. Seems to reduce dependence on the Equal Loudness ear curves.

That said has anyone tried adaptive curve eq based on absolute SPL like Dolby Volume or Audessey?

Try 5. Open baffle, mocked several topologies including Lx521, regular open baffle, U and H-frame. Currently listening to MTM NEO10 and NEO3 top baffle with 15" H-Frame woofer on the bottom. Most of the Try4 goods but rock music is not as good. Seems impact and bass light but that is how metal/rock is recorded.

General Comments
What does all this boil down to for me? Sealed and ported load the room differently than Open baffle. To listen at lower volume without adaptive volume control Large surface area drivers reduce problems with low listening levels. I have a hard time to choose between, high efficiency pro woofers (3-way with horn) and 3-4 way open baffle as my the preferred speaker.

All of the above are independent of amplifier topology or me. I usually time align and stay between 2nd order to 4th order xover, BW or LR. Found true single order acoustic roll off almost impossible to practically do without a good dispersion 3" and 6" driver. Forget tweeters, they end up being 2nd order at least by the time crossover occurs.

I found myself not sensitive to phase problems but prefer dynamics in spades. Also found room problems and speaker topology to trump reasonably designed AMPs/topology by a large margin. This must mean I also am not as sensitive to the parameter THD as measure by todays standards. But I don't like higher order (above 5th order) distortion.

I'd be interested in comments on the adaptive volume since I don't think anyone disagrees with the Equal Loudness Curve.

Cheers,

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Loudness etc: Tempest in a Teapot

Post by MikeC » Sun Jan 10, 2016 5:33 pm

mlee wrote:Hi All,

Was intrigued by this discussion title so I just signed up. I'll have to do an introduction post here shortly but wanted to comment on my experiences with the points brought up in this post.
Welcome to the SPCR forum. :)
mlee wrote:Roughly 6 years ago I embarked on a mission to design a pair of speaker that played well at all volumes/SPL levels. Something that could do it all nicely. Rock out, play vocal jazz, small and large ensemble. Well at everything but Master of none as I don't think there is a speaker topology that supports perfect....
Quite the journey!
mlee wrote:That said has anyone tried adaptive curve eq based on absolute SPL like Dolby Volume or Audessey?
I haven't but perhaps someone else has and will comment.
mlee wrote:What does all this boil down to for me? Sealed and ported load the room differently than Open baffle. To listen at lower volume without adaptive volume control Large surface area drivers reduce problems with low listening levels. I have a hard time to choose between, high efficiency pro woofers (3-way with horn) and 3-4 way open baffle as my the preferred speaker.
My own experience with open baffle speakers before the Orions were electrostatics (Quads, Infinity, Martin Logan) & planars like Magnepans. I never warmed to them due to narrow dispersion, visually dominating presence, generally limited dynamics, often limited volume capability, and mediocre bass. The Orions are a different animal altogether, and I totally agree with mlee on the fundamentally different, more open sound of open baffle bass with large dynamic drivers.
mlee wrote:All of the above are independent of amplifier topology or me. I usually time align and stay between 2nd order to 4th order xover, BW or LR. Found true single order acoustic roll off almost impossible to practically do without a good dispersion 3" and 6" driver. Forget tweeters, they end up being 2nd order at least by the time crossover occurs.
This jibes with my speaker design/building 100%. Given realities with drivers & constraints of design, it's usually best to have the xover work with the natural rolloff of the driver, which below resonance is typically 12 dB/oct. The only way you can actually get true 1st order xover slopes (6 dB) is to put the xover points far away from the natural rolloff of the drivers -- far above the natural rolloff for tweeters, below for woofers, and both for mids -- but this means having incredibly linear broadband drivers, and you'd end up "wasting" the linear FR of each driver. In truth, the claims of 1st order xovers by manufacturers refers only to the xover circuit used (ie, single capacitor or inductor), not the actual rolloff curve of the drivers + xovers.
mlee wrote:I found myself not sensitive to phase problems but prefer dynamics in spades. Also found room problems and speaker topology to trump reasonably designed AMPs/topology by a large margin...
Also generally my experience -- I love the live "jump" one gets from good dynamics, as in the Orions -- but I'm not as clear on audibility of phase problems for me. Have to experiment more.

Reachable
Friend of SPCR
Posts: 396
Joined: Wed Nov 19, 2003 9:55 am
Location: Western Mass.

Re: Loudness etc: Tempest in a Teapot

Post by Reachable » Mon Jan 11, 2016 3:58 pm

When you listen to recorded music at a volume that's lower (or higher) than that of the actual performance there's a kind of "cognitive dissonance" present. Partly it's because you have a notion of what the actual loudness was, but it's also because the acoustic waveform of musical instruments is different when sounded at different intensities. That doesn't mean that the experience necessarily becomes less pleasant. Pop music, especially, is likely produced to sound just as good at lower volumes.

I've always wanted to like the Beethoven symphonies when played on my record player, but I've never been able to. Although the compositions are endlessly fascinating, played at modest living room volume or even on great headphones they sound to me like, well, tempests in teapots. Beethoven wrote the Fifth Symphony for you to sit in a concert hall and get your head blown off. I do imagine, though, with a great audio system at appropriate volume it would be terrific. On the other hand, don't try to listen to an album by Cream at the authentic volume or you'll damage your house, your hearing, and your reputation as a law-abiding citizen.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Re: Loudness etc: Tempest in a Teapot

Post by MikeC » Tue Jan 12, 2016 8:53 am

Reachable wrote:When you listen to recorded music at a volume that's lower (or higher) than that of the actual performance there's a kind of "cognitive dissonance" present. Partly it's because you have a notion of what the actual loudness was, but it's also because the acoustic waveform of musical instruments is different when sounded at different intensities. That doesn't mean that the experience necessarily becomes less pleasant. Pop music, especially, is likely produced to sound just as good at lower volumes.
cognitive dissonance -- good one! 8)
the acoustic waveform of musical instruments is different when sounded at different intensities -- absolutely!!!
Reachable wrote:I've always wanted to like the Beethoven symphonies when played on my record player, but I've never been able to. Although the compositions are endlessly fascinating, played at modest living room volume or even on great headphones they sound to me like, well, tempests in teapots.
IIRC, classical music lovers the first to embrace the CD fully, no longer... having long works interrupted by having to flip or change records in the middle, being bothered by surface noise on quiet passage with a single solo violin, putting up with inner groove distortion, or tracking distortion on loud orchestral crescendos.
Reachable wrote:Beethoven wrote the Fifth Symphony for you to sit in a concert hall and get your head blown off. I do imagine, though, with a great audio system at appropriate volume it would be terrific.
I'm sure this is one of S. Linkwitz's pastimes given his classic music bent, and his 6 x 180W amp.
Reachable wrote:On the other hand, don't try to listen to an album by Cream at the authentic volume or you'll damage your house, your hearing, and your reputation as a law-abiding citizen.
Sitting very close to the speakers might help achieve this... or robust headphones. I recall many a post-concert experience in my youth, departing the venue with ringing in my ears. The loudest concert ever -- quite memorable because the ringing lasted a couple days -- was Jorge Santana's band playing in a small hall of some Boston-area college. It was so uncomfortably loud that I actually left my friends before the concert was over. Maybe Jorge felt the need to prove something given his superstar brother.

NeilBlanchard
Moderator
Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth
Contact:

Re: Loudness etc: Tempest in a Teapot

Post by NeilBlanchard » Thu Jan 14, 2016 9:51 am

I have been wearing ear plugs at most amplified concerts - you actually can hear better, and avoid the ringing ears, too.

I am going to have to work on my woofer - and I need to use the system more. Another angle on the amp/speaker/loudness discussion is speaker efficiency. My Snell Acoustics Type E/II's were rated at ~93-94dB/watt, so they always provided LOTS of volume with my B&K ST-140 amp. I made a few minor mods to it (trying to improve the grounding scheme), and it seems to work really well with my Audible Illusions Modulus 2C. The latter has zero negative feedback, and the low level resolution is excellent.

I listen to a broad range of music, and I ended up exposing problems in my Snell's playing (approximately) live levels of piano music - I had to make a large brace for the back of the cabinet between the rear tweeter and the input posts. But I DJ'ed several dances for friends and never had an issue even after several hours of house shaking volumes.

xan_user
*Lifetime Patron*
Posts: 2269
Joined: Sun May 21, 2006 9:09 am
Location: Northern California.

Re: Loudness etc: Tempest in a Teapot

Post by xan_user » Thu Jan 14, 2016 5:35 pm

NeilBlanchard wrote:I have been wearing ear plugs at most amplified concerts - you actually can hear better, and avoid the ringing ears, too.
look at the people working at the shows and playing onstage. 99%+ have hearing protection.

shoot, i even make a habit to wear plugs when im working in the server room. (wait, that should be our #SPCR silencing tip...19 cent earplugs... ) :mrgreen:

LostHighway
Posts: 38
Joined: Sat Oct 24, 2015 2:32 pm

Re: Loudness etc: Tempest in a Teapot

Post by LostHighway » Fri Jan 15, 2016 2:53 am

xan_user wrote:
NeilBlanchard wrote:I have been wearing ear plugs at most amplified concerts - you actually can hear better, and avoid the ringing ears, too.
look at the people working at the shows and playing onstage. 99%+ have hearing protection.

shoot, i even make a habit to wear plugs when im working in the server room. (wait, that should be our #SPCR silencing tip...19 cent earplugs... ) :mrgreen:
Amen to that! People tend to forget that the risk factors for hearing damage are not just amplitude but time (duration of exposure) and probably some individual differences in physiological resilience too. According to the CDC and NIOSH for every 3dB over 85dB the exposure duration for potential risk of damage is cut in half. IIRC the recommended maximum exposure time for average levels of 95dB is just under an hour. I've generally worn hearing protection if I anticipated average SPLs over 90dB for ages save for a few young and foolish years in my twenties.

I haven't measured my home listening levels for over a decade but unless they have shifted with time my average levels have tended to primarily range from the low-to-upper 70 dBs with peaks, save for those of very short duration (< a second or two), in the mid 80 dBs to maybe low 90s. Average levels above the low 90 dB range for in home listening I have almost zero interest in and average listening levels above the low to mid 80 dB range would be very, very rare for me.. As of my mid-thirties I could still easily hear the 19 kHz pilot tone for FM on a good system provided there wasn't a ton of other ambient noise. I'm sure that is no longer the case but I still try to take care of my hearing.

Post Reply