I'm on the same way, but I've already moved beyond the hypothesis phase approaching a proof
While I've been studying psychoacoustics in deeper levels of auditory chain I've about to come to the conclusion
modulation perception is more important for music listening than for example frequency response anomalies and distortion. Music is a process where carrier signals are AM and FM modulated by information. In our perception we demodulate the AM and FM information from the carriers. If there are external disturbancies modifying the modulations (hint: MTF) it will testify itself in various forms of perceptual changes. Some of them may be beneficial depending on the original source signal, and some of them may be bad.
- Elias
I think the notion that hearing is a time-dependent process doesn't need any proof. The problem is relating perception to measurements. How does a new metric help? It might be useful but the proof can't be the metric itself.
Now I am completly confused. So in an anechoic chamber dynamics is "as good as it gets" and all speakers perform about the same? Or at least as good as they can be.
This would also make dynamics a direct tradeoff to spaciousness - can't have both. Now its starting to sound weak because I have experienced both.
That is because HT makes sense. One special single seat room is a waste when one can have it all.
maybe we don't use the right terms.. dynamics increase with a dry room because it goes deeper in the quiet end and hence a larger scale, but loudness is also increased by reverberation, which also kills microdynamics, so it can get loud, but not dynamic because of a reduced scale? We have all tried a speaker in a pure reverberant filed at one stage, haven't we? I just sounds painfully loud..
This is simplistic, but I can see and understand why you are taking that train, with everybody else.
Yes.
I've experienced it in very reverberant rooms too but only if sounds before a very dynamic sound event were really soft.
The tensor tympani and stapedius might play a major role. They provide a high degree of dampening without the auditory event to become louder or softer.
Hi Markus
I think one has to look at the situation something like this. One can have a very pleasing sound and essentially no intelligibility. A sustained major chord played in a reverberant room will generally be pleasing and yet has no dynamic information in it. What sounds good, sounds good.
On the other hand, what I am talking about is information conveyed in an audio signal. With voice, where our hearing and acuity seems to be focused, (assuming an identical frequency response etc) the information in a voice is conveyed in this dynamics of the sound.
Your ability to understand random words is related to how well that dynamic is preserved by the time it reaches your ears. In a large space, by the time you can’t understand 100% of the words, what is reaching your ears has much less dynamic information than the signal had.
That Speech transmission index idea (the percentage of random words that can be understood) has been recently distilled into a more accurate measurement called STIpa which itself is a series of MTF measurements in the range they found were most important to voice. Here they measure the modulation depth in 7 bands from 125Hz to 8Khz in frequency.
So, you have a situation where decreased MTF measurements cause lower stipa scores because less information from the source and more corruption is reaching your brain.
I believe that the MTF’s higher up in frequency (than of most concern for voice) are also an indicator of how much of the signals information is arriving at the microphone.
The issue many seem to be focused on is dynamics and yet the ability to reproduce / preserve and deliver to the listening position a signal that is dynamically accurate or faithful (to the input) is what MTF’s are all about.
There are a myriad of things which can harm this measure too (which makes it fun).
See STIpa here;
Speech transmission index - Wikipedia, the free encyclopedia
I have been wrong plenty of times but I really think this is at least tickling the tigers tail but I don’t have time to do much more than what I have done (our big fat trade show is coming up in a couple months).
Again too, my concern here is not a pleasing sound (one must have nice frequency and phase etc) but the stereo image which is like a window into another space and not a curtain of sound in front of you.
In that case for a real stereo image, a mono voice signal, makes a strong center phantom image like a speaker or person was there, much stronger than the source images on the r and l. Many things can stop that from happening or make the l and r speakers more noticeable than the phantom but I am pretty sure (all other things like response etc being equal) what the MTF’s show is the presence of those “things”.
If anyone has ARTA, take some comparative MTF measurements.
See if a speaker doesn’t measure better up close than at the LP to start, at the LP, you have added the room reflections etc, just one of the things which can lower the score.
Best,
Tom
Back to work.
I think one has to look at the situation something like this. One can have a very pleasing sound and essentially no intelligibility. A sustained major chord played in a reverberant room will generally be pleasing and yet has no dynamic information in it. What sounds good, sounds good.
On the other hand, what I am talking about is information conveyed in an audio signal. With voice, where our hearing and acuity seems to be focused, (assuming an identical frequency response etc) the information in a voice is conveyed in this dynamics of the sound.
Your ability to understand random words is related to how well that dynamic is preserved by the time it reaches your ears. In a large space, by the time you can’t understand 100% of the words, what is reaching your ears has much less dynamic information than the signal had.
That Speech transmission index idea (the percentage of random words that can be understood) has been recently distilled into a more accurate measurement called STIpa which itself is a series of MTF measurements in the range they found were most important to voice. Here they measure the modulation depth in 7 bands from 125Hz to 8Khz in frequency.
So, you have a situation where decreased MTF measurements cause lower stipa scores because less information from the source and more corruption is reaching your brain.
I believe that the MTF’s higher up in frequency (than of most concern for voice) are also an indicator of how much of the signals information is arriving at the microphone.
The issue many seem to be focused on is dynamics and yet the ability to reproduce / preserve and deliver to the listening position a signal that is dynamically accurate or faithful (to the input) is what MTF’s are all about.
There are a myriad of things which can harm this measure too (which makes it fun).
See STIpa here;
Speech transmission index - Wikipedia, the free encyclopedia
I have been wrong plenty of times but I really think this is at least tickling the tigers tail but I don’t have time to do much more than what I have done (our big fat trade show is coming up in a couple months).
Again too, my concern here is not a pleasing sound (one must have nice frequency and phase etc) but the stereo image which is like a window into another space and not a curtain of sound in front of you.
In that case for a real stereo image, a mono voice signal, makes a strong center phantom image like a speaker or person was there, much stronger than the source images on the r and l. Many things can stop that from happening or make the l and r speakers more noticeable than the phantom but I am pretty sure (all other things like response etc being equal) what the MTF’s show is the presence of those “things”.
If anyone has ARTA, take some comparative MTF measurements.
See if a speaker doesn’t measure better up close than at the LP to start, at the LP, you have added the room reflections etc, just one of the things which can lower the score.
Best,
Tom
Back to work.
I'll leave that to Eickmeier and his hopefully refeatured appearing in this marvelous disscussion forum !
Perspective: contributing to a thread is one thing, posting a thread is another (particularly your own design) - perhaps there might be modesty barrier?
Tom,
I will make some ARTA measurements tonight. I have a JBL2445J on a large tractrix horn and a B&W DM602 for comparison. One is very directional, the other has wide dispersion (in the mids). See post http://www.diyaudio.com/forums/mult...ns-beaten-behringer-what-120.html#post3394911 for some distortion measurements of these two speakers.
It appears that what you are saying and what keyser is saying are related.
I will make some ARTA measurements tonight. I have a JBL2445J on a large tractrix horn and a B&W DM602 for comparison. One is very directional, the other has wide dispersion (in the mids). See post http://www.diyaudio.com/forums/mult...ns-beaten-behringer-what-120.html#post3394911 for some distortion measurements of these two speakers.
It appears that what you are saying and what keyser is saying are related.
I share that reluctance. A single measure is useful to the extent that it can be mapped to a single, or at least a definably small number, of causes or perceived effects. If acoustic MTF maps to frequency response, or frequency response plus distortion, then it gives the same sort of information we already have . . . just in a different form. But just as the MTF of a lens doesn't tell you much about bokeh or chromatic aberation, I don't see acoustic MTF telling much about either "dynamics" or "acoustic scene".
There are just too many variables for a single measure to encompass in a generally useful way . . .
This seems to argue that "speakers" don't influence dynamics. Is that what you believe?
No. There's a multitude of factors that influence "dynamics". The recording, the speaker and the speaker-room interaction.
I am not "reluctant" because I don't think that it would show useful information about "dynamics" - because I agree with Tom, I think that it would. I would be reluctant because I don't know how to relate the metric to speaker parameters. Without this later step, its just a nice metric for comparisons, but not too useful for development. Of course the relationships could always be accomplished with a serious design of experiments on the various parameters, but I am not capable of doing that, and it would be a massive undertaking for anyone to do. The bottom line is, that's why so much of this is unproven hypothesis, because making it science is massively time consuming and expensive. Who is going to do that work?
OK, getting pretty hazy again. What in the "speaker" do you consider to be a "factor"?
Gedlee,
I was away for a while and this thread went crazy. I am still reading to catch up to all that was said and have to leave again so hopefully my content won't be really out of context here. I was saying not that you would record the piano through an original set of speakers and them play it back through another set of speakers with a similar tonality, but that you had recorded direct to disk and skipped any speaker interactions at all. Then I would want the speakers to have the same tonality that the piano player was use to hearing his piano over to create the sound that he would expect. Does that make sense. You could have just played it back on his normal speaker system and he should have identified that as what he normally hears, this would be the objective with a direct recording, DI, and playback.
I was away for a while and this thread went crazy. I am still reading to catch up to all that was said and have to leave again so hopefully my content won't be really out of context here. I was saying not that you would record the piano through an original set of speakers and them play it back through another set of speakers with a similar tonality, but that you had recorded direct to disk and skipped any speaker interactions at all. Then I would want the speakers to have the same tonality that the piano player was use to hearing his piano over to create the sound that he would expect. Does that make sense. You could have just played it back on his normal speaker system and he should have identified that as what he normally hears, this would be the objective with a direct recording, DI, and playback.
Not sure why you're asking. The discussion started when I agreed to a hypothesis that dynamics isn't just a speaker/transducer problem (this is what you're asking me right now) but is significantly influenced by the indirect sound field (which is a topic I'm interested in).
Well then . . . what does it measure (and how does one make the measurement), what does "dynamics" mean, and how (by what mechanism) do the two relate? Maybe there is something there . . . maybe it's just another of those numbers that turns out to be little more than a marketing gimick. The lack of any plain-and-simple definition and explanation is both confounding, and off-putting . . .
Except that the piano player only ever played acoustic pianos, like Steinway, etc. He only played with large orchestras and for recordings, so his familiarity with electric pianos was nil. He said that our Yamaha electric piano had a very good "feel", far better than he expected, he was surprised. Feel of the keyboard was all that mattered to him, and he said that was fine.
I'll ask him if I can post the recording. I am not sure what he will say.
He has an impressive show on Youtube Atamian Plays "The Rite of Spring" for Solo Piano, Live, Part I - YouTube. Worth checking out. Great performance average recording.
Hi here k: If you have attened a sympphony concert of the "1812 Overature" performed with 105mm howitzers, you can have a dynamic experience live outdoors, thus minimizing reflected sound. ( on the Mall in DC and Fort Monroe, Va.) I agree with your comment about quiet before the dynamic sound, such as a bass drum hit appearing out of nowhere in the first movement of "Rite of Spring". I've experienced this at Tanglewood (B.S., Ozawa, outdoors in a tent structure) and also in concert hall performance, even my dog takes notice of this when played at home. I feel that dyamics plays a larger part in music, than just the kaabangs...Ahma Jamal's trios wih piano-bass-guitar (or later performances with drum set), a Issac Stern violin solo, Adrian Rollini bass sax work etc, are also dynamic. So how can we quantiy what we experience and apply to loudspeakers? ...regards, Michael