|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
98.223.15.64
In Reply to: RE: High efficiency speaker with great reproduction of speech posted by beppe61 on April 07, 2014 at 01:10:11
Intelligibility has a specific meaning and can be measured, unlike fidelity is not a subjective thing. Originally measured by reciting a hundred or thousands of random words and then listeners record what words they heard. It was found MANY things were involved or govern intelligibility.
In commercial sound where all the acoustic problems are worse and the required sound level higher, there was an early formula called the Hopkins Stryker equation which had some predictive ability.
That can be seen here in the original H.S. equation, note that all things being equal, the more sources you have the worse intelligibility is, the less directivity the worse it is, the less absorption, the worse it is and if the speakers are not aimed at the listeners, the worse it is.
http://acousticworx.com/sound%20system%20design%20hopkins%20stryker%20formula.html
This was superseded by the modified equation and then the ALCON and STI measurements.
Now the current ‘best” language independent predictor of intelligibility which has legal standing is the STIpa measurement.
This measurement is based on the modulation transfer functions at a number of different frequencies weighted to a voice spectrum.
The MTF measurement in acoustics is one which quantifies how much of the information in the signal reaches the microphone. Here is an example of an MTF as used in optics.
http://www.normankoren.com/Tutorials/MTF.html
To be clear while intelligibility can go hand in hand with pinpoint stereo imaging, preserving information as indicated in MTF’s is not a requirement to enjoy the sound.
For example, a choir is a large room may sound dreamy and a recording of a choir played in the same room may also sound dreamy even though there is zero intelligibility and no sense of stereo image.
While STIpa can be easily measured;
http://www.studiosixdigital.com/audiotools/sti-pa/
I am not sure it is useful in a living room, on the other hand, if one were comparing different loudspeakers in the exact same locations or listening position in a room or anechoically, then the MTF measurements at the heart of the STIpa are I think useful as a way to compare “how much of the recording gets to the listening position” with one speaker vs another. Arta has a MTF routine.
What shows up too is that the more it preserves the signal here, the more generations one can record and play back music using the loudspeaker and measurement mic in the loop, before subjectively falling apart.
Here, most loudspeakers sound surprisingly bad by generation one or two,few reach three, by far they are the weakest link if being faithful to the signal is the yardstick.
Best,
Tom Danley
Follow Ups:
Hello and thank you sincerely for your kind and valuable advice
This is very difficult to me to understand as i am extremely uneducated in physics and similar sciences.
I am sorry for this but i will try anyway because this topic interests me very very much.
But i do not fully agree on the fact that " Intelligibility ... unlike fidelity is not a subjective thing "
I think that intellegibility and fidelity are the same thing really.
Much more.
The fact that intellegibility can be measured and fidelity instead is a less defined parameter elevates in importance the intellegibility versus fidelity.
" The perfection is mathematical " there is nothing of subjective.
For me my next step will be to select some digital tracks with speech recorded with very high quality and to use them as a main tool for testing audio rig set-up.
I am completely sure that a system great in speech reproduction cannot fail when reproducing music.
A least in the voice bandwidth.
But i also think that to add some bass and highs is not that complicated.
The midrange is what makes or breaks sound.
Thanks a lot again.
Kind regards,
bg
Edits: 04/09/14
Hi
What I mean is that intelligibility can be measured, if you use random words (not sentences) and can only make out half of them, you have 50% and don’t have good intelligibility.
To have intelligibility, one must convey the information within the audio signal and a key part of that is preserving time.
On the other hand, what is enjoyable, what sounds good can’t be quantified anywhere nearly as easily, it is subjective based on how it sounds to the listener and like the choir example, one doesn’t need to preserve of convey the “information” to sound good.
One can go to a concert or large live event and not understand many of the words but still very much enjoy the show.
The best way I have found to subjectively quantify a loudspeakers faithfulness is a generation loss recording, just like with other parts in the chain, the more generations one can record and play before being unlistenable, the more faithful that element is to the input signal.
If you do this outdoors up in the air, the only thing you hear is what the loudspeaker does transmogrifying the signal.
The MTF measurements I mentioned are a measure of resolution, if you Google “modulation transfer function measurements” and examine the many examples of it’s use in optics, you can get an idea why that is the key to the STIpa measurements used to predict inteligibility.
In audio, the loudspeaker is modulated on and off at a given base frequency and if what reaches the mic is either full or zero modulation, then the the MTF is 100% at that modulation rate.
Keep in mind, this is what I see not a universal belief but it has been the direction I have gone at work in large scale sound to preserve intelligibility and fidelity.
If you have ever been to a large stadium or concert and heard a CD being played, you have a base line for what line arrays sound like. The larger they are, the more they disperse time information, the less intelligibility they have, the less they are able to produce a mono phantom image in a stereo configuration.
Here are a couple examples of what can be done in large scale where everything is harder, these systems can produce a stereo image, have very high intelligibility and articulation, both nonexistent when there are many arrivals in time like a typical array produces.
Again, if you have been to a large stadium before and heard the typical state of concert sound, pop on some headphones and compare that approach to these large stadiums using loudspeakers and layouts driven by preserving signal information..
https://www.dropbox.com/s/tnsw5mb4v5vdlwq/20120726122124.mts
https://www.dropbox.com/s/oyosfc3adc6j1du/20130723135350.mts
We don’t advertise like the marketing driven companies but we do get press every so often;
http://www.psam.uk.com/danley-loudspeakers-infiltrate-sports-market
Best,
Tom
Hi and thanks a lot for the valuable reply but a large live event is very far from my situation (small to medium listenig room).
I am looking for a speaker that makes me wonder if i am listening to recorded human voice of real human voice.
And then the rest will follows automatically.
I really do not know if systems extremely good in open space can be scaled down for home listening.
I tend to think so anyway.
Thanks a lot again.
Kind regards,
bg
Edits: 04/12/14
You don't need to advertise...the people will find you on their own.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: