Of course we need to make sure that whenever we gain something, we are not losing something more important.
That's why I'm trying to simplify to "what's most important." So that it can be understood what needs to be right and what would be nice to have right to get to a better design. IOW, get your polars in order no matter what the design. 😉 Spectral balance of the direct and reverberant field are where it seems to start for recording to sound right. I'd bet CSD is useful is well for us who can't read an impulse response in great detail (that would be me). All I know is that a shorter tail means less delayed energy and a steeper slope at the start tells how much HF is coming out of the speaker. I can tell you that my speakers with a compact impulse response not only sounds more detailed to my ear than a long tailed one, but also more dynamic. Of course it doesn't mean that necessarily the cause even though it does seem reasonable. I can't say it has affected my perceived spectral balance in any recording, but I can "see" where it might do such a thing. Keep in mind that the statistics of one are useless especially when sited and the metrics are known and thus I hate stating them. However, It would be interesting if some other might have had a similar experience or if there is any studies out that that correlates to my perception. Soongsc's experience with similar phenomenon seem different from my own--or at least what he has discussed to far.
In any case, I'll take a compact IR to go next on my list. I wish I could do detailed CSD plots.
Dan
In any case, I'll take a compact IR to go next on my list. I wish I could do detailed CSD plots.
Dan
Dan, those are very interesting observations about the impulse response. I should spend some time with my different speakers and try and see how different the impulses are.
Why can't you do detailed CSD plots? IF you are using REW, it has them. No?
If you can do an impulse measurement then you save the wave file and import it into ARTA.
Why can't you do detailed CSD plots? IF you are using REW, it has them. No?
If you can do an impulse measurement then you save the wave file and import it into ARTA.
The time scale on REW is too large for the plots Soongsc is talking about. To be honest though, I believe the new version can do better ones. I need to check it out. Currently I only have access to a Macbook and can only use REW or Fuzzmeasure(the free version unless I'm feeling spendy one day). I need to get a netbook. A man just needs a MS product around the house unfortunately. It would be interesting to do CSDs from different angles as well as various time periods. I've never seen them. But a compact impulse on all angles sounds good for sure and it's been added to my list.
Dan
Dan
Just run Windows on your MacBook. Sun VirtualBox is free if you don't want to boot natively,
Or better yet, be nice if some of these would work in Crossover Mac (ie WINE)
dave
...
Besides a uniform or at least "smooth changing" dispersion with frequency,
i regard CSD's highly. To me there is a clear correlation between sound
quality and CSD although the importance of fast decay lowers IMO at
the ends of the audio spectrum, especially at the upper end.
In my experience a bit ringing above say 8 Khz is not that critical
than in the presence region from 2..6 Khz. Also the bass to midrange
region should be as "clean as possible".
For the frequencies below the Schröder frequency of the room,
things seem a bit tricky.
The more balanced the room modes are, the more a speakers
misbehavior gets into focus.
But i feel even in a mediocre room it is audible whether a speaker
produces a "smeared" bass or not. I know that some do not share
this opinion.
I did some "burst decays" whith ARTA under suboptimal conditions,
there are some reflections visible in the measurements.
But the burst decays shown clearly correlate with the sonical quality
of the three speakers (quality not meant necessarily in terms of
"good" and "bad").
I the burst decays the "time" axis is normalized to periods and
not in absolute time. This means compared to a CSD the decay
at higher frequencies is made more stretched and visible.
Kind Regards
Attachments
Last edited:
Yes, the decay characteristics of loudspeakers will definitely affect the perception of Domingo's sound. Trained opera singers have a formant in the region of @3kHz and a subsidiary one at about double the frequency or a bit higher.
There is no mystery here. 😉
See this article:
The Singer's Formant and Speaker's Ring Resonance: A Long-Term Average Spectrum Analysis
[snip]
[QUOTEIn this study, the authors found energy at around 8 kHz in most trained singers' singing voices and soprano' speaking voices. Although we did not measure the value in this study, there was the possibility that another energy peak could be theoretically present near the 15 kHz. However, the energy level around the 8-9 kHz region is about 20-30 dB below that of singer's formant region, suggesting that it may be beyond the range of perceived sound since the sensitivity of the human auditory system is also about 20 dB lower around the 8 kHz region instead of the 3 kHz region. However, with other recording and amplification systems, where high frequency emphasis is given, a formant of 8-9 kHz could reach perceptual significance (6). It is thought that this second singers' formant may contribute to their rich and ringing voice.[/QUOTE]
The reproduced sound of these singers can pose problems for loudspeakers (and for recording engineers). The live performance SPLs some of the men can range from 70 to a 100 dB at one meter. So, if your average listening level is 80 db these guys can definitely come out and bite you, and your loudspeaker.
There is no mystery here. 😉
See this article:
The Singer's Formant and Speaker's Ring Resonance: A Long-Term Average Spectrum Analysis
[snip]
[QUOTEIn this study, the authors found energy at around 8 kHz in most trained singers' singing voices and soprano' speaking voices. Although we did not measure the value in this study, there was the possibility that another energy peak could be theoretically present near the 15 kHz. However, the energy level around the 8-9 kHz region is about 20-30 dB below that of singer's formant region, suggesting that it may be beyond the range of perceived sound since the sensitivity of the human auditory system is also about 20 dB lower around the 8 kHz region instead of the 3 kHz region. However, with other recording and amplification systems, where high frequency emphasis is given, a formant of 8-9 kHz could reach perceptual significance (6). It is thought that this second singers' formant may contribute to their rich and ringing voice.[/QUOTE]
The reproduced sound of these singers can pose problems for loudspeakers (and for recording engineers). The live performance SPLs some of the men can range from 70 to a 100 dB at one meter. So, if your average listening level is 80 db these guys can definitely come out and bite you, and your loudspeaker.
There is no single factor that is most important. It's like trying to say, input impedance is more important than output driving capability of an amplifier. I don't think it's realistic to even try. User have their own criteria what is important to them, designers decide what's important in thier own view. Data is not useful in making the final decision, but is useful in to areas:
1. Help finding more significant problem areas if you first identify there is a problem audibly.
2. Help determine compatibility.
Only sales persons will try to mislead customers to believe one set of data is more important than the other. I don't even think that's ethical.
Distortion at the amp output when loaded with the speaker is also very important. If a speakers can be designed so that it does not worsen the amplifier distortion figures when connected to an amplifier, then it's designed with good interface characteristics.
If you use BootCamp you should have no issues... you are in essence, running a high end Windows machine,
With a virtualizer like VirtualBox (Fusion or Parrallels) depends how well the software is written, I'll let you know when i get an Intel Mac (my old ones, just like the Energizer Bunny, keep going & going & going).
If the "soundcard" is an external Firewire device there should be no issues. USB is a bit more problematic. Accessing the internal Mac I/O should be OK but it may have limitations at the ends of the Frequency Range
BTW: i just upgraded to Fussmeasure 3, haven't done any measures with it yet.
dave
Thanks FrankWW! Very cool stuff. Where were you earlier in this thread? I'm going to read the article completely when I get a chance.
Soongsc, I don't see how designing a speaker to conform to what research indicates is just marketing. In other fields they call it R&D. I do think getting a great speaker will require more than 1 metric as I'm sure you know by now.
Thanks Dave. I think you're right. Why buy a cheap machine to have widows when you can just put it on a better machine? Can I still use my old purchased copy of XP? I put it on another machine that is now too broke to be worth fixing.
Dan
Soongsc, I don't see how designing a speaker to conform to what research indicates is just marketing. In other fields they call it R&D. I do think getting a great speaker will require more than 1 metric as I'm sure you know by now.
Thanks Dave. I think you're right. Why buy a cheap machine to have widows when you can just put it on a better machine? Can I still use my old purchased copy of XP? I put it on another machine that is now too broke to be worth fixing.
Dan
Last edited:
I think Sumaudioguy raised the issue about dynamics of speakers. Can't quite remember what that post talked about. But I am quite interested in various aspects of that subject. We know that power compression is a factor if we play at certain levels for extended periods of time, driver linearity might cause compression, which is also identified by change in harmonic content, and can occur at the very small signal level and the excursion limits as well.]
How can we objectively quantify the dynamics capability?
Dan, I just think that advertising "the most important" thing is like forming a cult around it. And I will butt it whenever I see it. Sorry about that. I have both my engineering hat and audiophile hat on when I say this.
How can we objectively quantify the dynamics capability?
Dan, I just think that advertising "the most important" thing is like forming a cult around it. And I will butt it whenever I see it. Sorry about that. I have both my engineering hat and audiophile hat on when I say this.
Last edited:
I think's it is more of an issue than many think. I have only seen power compression numbers published on a handful of commercial speaker systems. The JBL Array 1400 being one of them. They rate them at less than .5db of compression at 100db. That's with about 10-12watts of input. This system is not using inexpensive low power drivers.
I have a clone of them and they sound great all around and effortless as far as dynamcs are concerned.
Rob🙂
Last edited:
A simple dynamic range test is to use a spectrum analyzer and two oscillators. Set one oscillator at say 440Hz and the other at 466, A and A#. Two peaks on the spectrum should be visible if the tones are equal magnitude. Next turn down one tone until it disappears in the noise. At 95dBa many times this tone vanishes when it is but 20dB down. That difference in level is the dynamic range of that speaker at that frequency at that SPL. Obviously the dynamic range will decrease with increasing level and decrease with decreasing level away from the SPL and frequency where the dynamic range of the speakers is a maximum.
Power compression and power non-liearity... The attached article from post #887 on July 2nd is short and worth reading by anyone interested in dynamics and power compression.
As for listening "experience" versus "measured" results.... I believe everyone should keep in mind than no matter how good a speaker measures if listeners find the speaker to sound bad then the speaker serves no purpose as the speaker is useless and annoying to listen to. If a speaker sounds really good but has very bad measurements then this begs the question "am I being fooled?" IMO a speaker needs to sound good and measure reasonably well to qualify as the genuine article.
I have seen many speakers that measure very well according to the manufacturer and yet SLS and are unlistenable for undetermined reason. I kind of thought part of the point of this thread was to come to some useful correlation between what sounds good and measures well. Certainly measurement and sound experience correlation is what might guide us. Claiming a single parameter is by far the most important has always proved wrong in other fields of endeavor were science meets experience. I mean there are people that like the way a BMW drives and rides yet I hate them for those very reasons. This is why way back in post #88 I think, breaking the listening audience into different groups based on hall seating positions and focus on satisfying a particular group only makes sense. There is no "one size fits all" in shoes, clothes, cars, or sound reproduction goals. The sound heard by the conductor at the podium is what I like. No compression driver system has every taken me there.
Power compression and power non-liearity... The attached article from post #887 on July 2nd is short and worth reading by anyone interested in dynamics and power compression.
As for listening "experience" versus "measured" results.... I believe everyone should keep in mind than no matter how good a speaker measures if listeners find the speaker to sound bad then the speaker serves no purpose as the speaker is useless and annoying to listen to. If a speaker sounds really good but has very bad measurements then this begs the question "am I being fooled?" IMO a speaker needs to sound good and measure reasonably well to qualify as the genuine article.
I have seen many speakers that measure very well according to the manufacturer and yet SLS and are unlistenable for undetermined reason. I kind of thought part of the point of this thread was to come to some useful correlation between what sounds good and measures well. Certainly measurement and sound experience correlation is what might guide us. Claiming a single parameter is by far the most important has always proved wrong in other fields of endeavor were science meets experience. I mean there are people that like the way a BMW drives and rides yet I hate them for those very reasons. This is why way back in post #88 I think, breaking the listening audience into different groups based on hall seating positions and focus on satisfying a particular group only makes sense. There is no "one size fits all" in shoes, clothes, cars, or sound reproduction goals. The sound heard by the conductor at the podium is what I like. No compression driver system has every taken me there.
Everyone should read this link: Audio Musings by Sean Olive: Part 3 - Relationship between Loudspeaker Measurements and Listener Preferences and then Dr. Toole's book. It would help remove us from the merry-go-round discussion. Sort of like building loudspeakers to specific metrics can remove us from the circle of confusion.
SUM, thanks for the dynamic range test. Cool stuff. BTW, I like the BMWs drive but there are better ways to test its actual performance. However, if you don't enjoy driving it, there's no point in owning it.
Dan
SUM, thanks for the dynamic range test. Cool stuff. BTW, I like the BMWs drive but there are better ways to test its actual performance. However, if you don't enjoy driving it, there's no point in owning it.
Dan
It only helps remove those who follow and believe what Harman International is "selling". I only say "selling" because it is what others think and they are a for profit company anways. FWIW, I do follow and agree with most of Sean's research. There has been several discussions with him over on AVS.
I agree, I think we can savely conclude that there just isnt a good enough correlation from sound experience back to specific measured differences. There is a stubborn nature about the audio world in terms of always pushing Science to the back seat. There isnt a discussion/debate in any forum that does not have two POVs.
1. Let your ears decide what is good.
2. If we can not show the measured difference then it didn't really happen.
Which leads me to this other POV. Why do we care what others choose? If we have the ability to build our own systems then we have hopefully almost complete control over the final result. What is great for me is not great for others, what is great for others does not suit my tastes.
The point about "one shoe does not fit all" is a prefect summary and kind of what I have been posting many times in this thread.
There will never be a general consensus on what is truely meaningful because there are so many truely unique requirements out there.
THere are different driver choices for different applications. I have yet to read about a signal driver choice that does it all perfectly. IMO Ribbons are incredible and for near field or short listening distances I would pick them over Compression drivers.
Last edited:
Yes, but without standards recording/mixing/mastering, design of loudspeakers and room treatments are just a shot in the dark.
Unfortunately there are no useful standards. What should the reflection pattern of a listening room look like? What speaker directivity is desirable? What is practical?
There are some standards in the film industry but nothing near the level of standardization we should aim for. And as long as audiophiles mix up the art of music creation with the solely technical aspects of sound reproduction, nothing will ever change.
Last edited:
Isn't this quite similar to the nature of measuring the Modulation Transfer Function?
- Status
- Not open for further replies.