I suppose for this thread this is considered trivia:Folk music (worldwide) is a good way for us conditioned to western classical music (whether we know it or not, I include pop music in that!) to expand our horizons musically (obviously)
I have a pet theory that the dreaded 7th harmonic is part of the same raw deal
Piano hammers strike the string at 1/7th the total string length from one end to reduce the volume of the 7th harmonic (except for the higher notes where the 7th is above hearing).
There are certainly lots of detailed aspects of hearing that are not understood and as time goes on they will become ever more detailed and obscure. That's what research does. All of the things listed above are certainly reasonable and believable, but I don't see any of it affecting the way loudspeakers are designed or evaluated. (Perhaps #2 which would imply that more nonlinear distortion would lower the sensitivity of a low frequency tone.)
Lidia and I have never concerned our self's with "thresholds, especially not for pure tones. They just don't tell us much about how a loudspeaker will perform using music at levels way beyond thresholds. What we want is a scale that tells use how the effects can be judged on a continuum of values, not simply the threshold. The scale may even get the threshold wrong, but still be quite adequate for music at normal levels. That is why we have always used musical signals and not special laboratory generated signals. Those kinds of signals make the research easier, but the applicability comes into question.
I do not see any of these more obscure details having any effect on my beliefs or how I design loudspeakers. It's the major effects like masking, which is not a threshold situation but is exhibited throughout our hearing dynamic, that matter. I find the new realization that different aberrations have different scales of perception at different SPL levels. This has absolutely nothing to do with thresholds and yet, to me, is the most significant result to come about for loudspeakers in a long time. If I were still doing research that is where I would dig for gold.
In PinYin possibly, but in the classical script there should not have been any ambiguity.
I am very fond of Chinese Opera having seen about half dozen over the years - everywhere from Beijing to Shanghai. The "monkey king" is my favorite.
Classical script tell you the exact word, but learning to pronounce it requires PinYin in China, quite different from the way we learn in Taiwan. The character in China are also simplified versions. It is not so easy for me to recognize all myself, but learning though. Definitely can write faster.
Apple has made some improvements in the PinYin keyboard so that it automatically can suggest Chinese or English, it was a gift to China in iOS 10 release, and after the fact that I suggested they need to improve the method of switching between the languages.
I think TIM cook could go into politics after his Apple carrier, from the way he interacts with situations.
Apple has made some improvements in the PinYin keyboard so that it automatically can suggest Chinese or English, it was a gift to China in iOS 10 release, and after the fact that I suggested they need to improve the method of switching between the languages.
I think TIM cook could go into politics after his Apple carrier, from the way he interacts with situations.
Last edited:
I introduced the concept in the other thread that complex tones are not limited to the thresholds that are normally cited when people talk about audibility. I find that people often look at an FFT & conclude such & such a harmonic frequency is below audibility or the even more gross misconception - it will be masked in music. Both concepts are simplistic & ignore how auditory perception works as I pointed out with links to some of the research. So , yes, I find it is relevant to all audio reproduction
With respect to comodultaion masking release (CMR), no it's not about modulating noise which unmasks in every noise dip - the research uses noise because it is convenient but CMR is about the fact that a tone which is inaudible alone becomes audible when comodulated noise is overlaid on it. It's also about a much wider aspect & that is that any other frequency tone which is comodulating will also have the same effect i.e reducing the threshold of audibility of a target sound which would normally be inaudible. This is not just a new research topic - it has been known about from the early days of auditory perception research & came about because it is a natural phenomena that is observed in nature - lots of sounds in nature feature complex sounds t different frequencies which are comodulated
With respect to comodultaion masking release (CMR), no it's not about modulating noise which unmasks in every noise dip - the research uses noise because it is convenient but CMR is about the fact that a tone which is inaudible alone becomes audible when comodulated noise is overlaid on it. It's also about a much wider aspect & that is that any other frequency tone which is comodulating will also have the same effect i.e reducing the threshold of audibility of a target sound which would normally be inaudible. This is not just a new research topic - it has been known about from the early days of auditory perception research & came about because it is a natural phenomena that is observed in nature - lots of sounds in nature feature complex sounds t different frequencies which are comodulated
Last edited:
If what you say were true then perceptual coding would not work as it does not use any of the features of CMR (or at least none that I know of.) And yet we all know that they work quite well even inaudible if done correctly. So I do have to discount that CMR is as significant as you imply.
My wife has her PhD in audiology and Psycho-acoustics and hardly knew anything about CMR so its clearly not mainstream even if it has been known. Blauert never even mentions it in his book. Moore does mention it (about 9 pages of his 500 page book) claiming that it was only discovered in 1984 (not the "early days of auditory perception research" which go back to the 20's and 30's at Bell Labs with Fletcher) and he only ever mentions modulated noise, not tones. So while it obviously is a real effect and that people do study it it is not the dominate effect that spectral and temporal masking are.
I'm sorry but I remain unconvinced that this has anything to do with loudspeakers and their perception.
Doesn't this comodulation thing apply to the lower end of the dynamic range ? The masking effects (referred to by Earl) that are affecting the audibiltity of harmonics are influencing the other end of the dynamic range, making low level harmonics less audible (or even inaudible) at higher SPLs.
Since loudspeaker distortion is usually veeeeeeery low at low volumes this comodualtion might possibly be completely irrelevant to listening via loudspeakers at reasonable levels.
Regards
Charles
Since loudspeaker distortion is usually veeeeeeery low at low volumes this comodualtion might possibly be completely irrelevant to listening via loudspeakers at reasonable levels.
Regards
Charles
This is not just some arcane phenomenon that just occurs in human hearing, if follows from theory. The addition of noise is used to push low level signals above the noise floor in for example neural network processing.
The concept is fairly simple to grasp. Imagine a signal that is below the threshold of detection: it will not be detected. Now, add noise, just enough to push that signal above that threshold. The result will be that enough of the signal will be pushed above that threshold in order for it to become detectable.
Our human sensor pod is quite noisy, a nice design feature. Everybody has tinnitus, only the extent varies, and not everybody gives it a name. But all ears produce noise. Eyes the same thing. Best time of day to notice is on a clear day at sunset or, if you are so inclined, sunrise. Your mind makes a construct of a solid gradient from dark blue to light blue with perhaps some red. Look a bit better, and see how speckled and non-uniform the image in reality is. That is your own hardware bumping up the ISO, creating noise, without which it would have been dark already.
I'm no expert in compression codecs using perceptual techniques but there certainly has been research into CMR use in such codecs & the conclusion that it may well be an oversight in certain situations. It can be argued that it's immaterial based on what you say above but it might also be argued that much higher compression may be possible using CMR in the codecs. In other words the perceptual model used in such codecs may be considered incomplete.
Yes, Fletcher's 1940s experiments identified CMR as a phenomena - noise/tones - it's still the same underlying phenomena
You may be correct that it is immaterial to loudspeakers in their current design?
I don't believe that there is a specific loudness range associated with the CMR phenomena
No, if I understand what you're saying, it's not like you are describing it - it involves noise or tones at a different frequency which are amplitude modulated in correlation with the target tone & thus lower the target tone's threshold of audibility
Hmm, I don't think I've ever noticed the speckling in natural vision??
Firstly, the research is still trying to unravel how CMR works. Your theory does not work when CMR involves a tone at a completely different frequency to the masked tone but which is comodulating with the masked tone
I don't want to take up bandwidth here with OT but I don't understand how a tone at another frequency can push a masked signal at a target frequency above threshold?
There is a correlation mechanism stipulated to be possible at the neurological level but it's not a simple mechanism
Yes, that's the main takeaway - complex tones presented binaurally are perhaps perceived differently to what the thresholds suggest !!
Spectral masking effects allowed for Perceptual Codecs (PC) to have a data reduction of about 60-70%, temporal masking allowed for another 20-30%, I do not doubt that using CMR might offer another 2-3% reduction. That's just about how large the effect seems to be. It's there, it's measurable, but it is not very significant.
But according to Moore the term CMR did not come about until the 80's, so how could Fletcher have used it in the 40's? And if he did call the same phenomena under a different name then how is it that Moore missed this incredibly important fact, since he makes no mention of it?
Has it been measured at 2-3% data reduction? I haven't seen this as I haven't seen many use perceptual models which include CMR in their compression codecs
Here is one such company & AES presentation presenting a new compression codec using it, however
"A New Broadcast Quality Low Bit Rate Audio
Coding Scheme Utilizing Novel Bandwidth
Extension Tools"
Some excerpts from it
What is being said is that a panel of expert listeners confirms the model but unfortunately no statement about how often the conditions occur in music signals that give rise to CMR? make of this what you will.
Fletcher noted this effect in 1940 & how the threshold reduced as the bandwidth of noise increased.
Sorry but can't say why Moore or others fail to mention Fletcher's discovery of this phenomena but Fletcher's work in this area can be read?