Math Unraveled

[MylesBAstor]

http://phys.org/news/2013-02-human-fourier-uncertainty-principle.html

 

[DonH50]

DSP will do that too. The brain is a marvelous thing, however, and I was shocked to discover a few years ago on this very forum just how small a time difference we can distinguish.

 

[jkeny]

Yes I posted about this back in Oct 2012 http://www.whatsbestforum.com/showt...-and-acoustics&p=141145&viewfull=1#post141145 but maybe here is a better place for discussion.
An number of possible issues worth discussing from this paper:
- the apparent almost exclusive focus on freq analysis seen in audio measurements
- the use of linear measurements in evaluating non-linear audio playback systems
- the use of single tone, linear signals as input signals for these measurements

 

[LL21]

Whats the takeaway here for laymen? I read the article and believe the takeaway is that while good engineering fundamentals are important...because there is so much we do not know scientifically about how we hear...the ultimate test is still just listening to equipment...because we do not know everything we need to measure in order to get great reproduction.

 

[DonH50]

Measurements are hard, and understanding them is harder. It is also very hard to construct a test that isolates what we are trying to measure, and the observer can corrupt the observation.

Less cliche, an observation from one of my college socialology classes: There were studies long ago showing driver's ed made better drivers and it was often mandated and taught in schools. Later studies showed the students taking driver's ed were often upper students and are better drivers with or without the driver's ed training. They weren't better because of the training, or not much, but were better for many, many other reasons.

That paper, or earlier versions of it, have been hotly debated for similar reasons.

 

[LL21]

Measurements are hard, and understanding them is harder. It is also very hard to construct a test that isolates what we are trying to measure, and the observer can corrupt the observation.

Less cliche, an observation from one of my college socialology classes: There were studies long ago showing driver's ed made better drivers and it was often mandated and taught in schools. Later studies showed the students taking driver's ed were often upper students and are better drivers with or without the driver's ed training. They weren't better because of the training, or not much, but were better for many, many other reasons.

That paper, or earlier versions of it, have been hotly debated for similar reasons.

Thanks, Don. Perhaps i should ask a follow on question...as someone who seems very technically capable, do you think the industry knows all of the right measurements to take in order to get the best sound, and the priorities in which those measurements should be judged in order to make the best sound? Or do you think we are still trying to understand the human ear, auditory brain response, etc...as this article seems to suggest?

 

[DonH50]

That's a loaded question... I think we have the equipment, and certainly the dynamic range and resolution is there, but IMO the really, REALLY hard part is figuring out what to measure and how to correlate them to what we hear. As for the industry, only a very few things are actually measured and make it into a data sheet. I do not know what is going in universities but I think there is a lot more research that could be done on the measurement side, understand the ear/brain/hearing system, and psychoacoustics in general. Put another way, we have the tools, but knowing how to use them and how to relate them to what we hear, there's a gap.

 

[LL21]

That's a loaded question... I think we have the equipment, and certainly the dynamic range and resolution is there, but IMO the really, REALLY hard part is figuring out what to measure and how to correlate them to what we hear. As for the industry, only a very thing things are actually measured and make it into a data sheet. I do not know what is going in universities but I think there is a lot more research that could be done on the measurement side, understand the ear/brain/hearing system, and psychoacoustics in general. Put another way, we have the tools, but knowing how to use them and how to relate them to what we hear, there's a gap.

GOt it. Thanks.

 

[mep]

I couldn't agree more with what Don said. I certainly believe the test equipment exists to measure anything we would possibly want to measure. The trick is knowing what to measure with the test gear we have that correlates with what our ears hear. We aren't there yet. We have a decent foundation, but we don't have the entire building constructed yet.

 

[LL21]

I couldn't agree more with what Don said. I certainly believe the test equipment exists to measure anything we would possibly want to measure. The trick is knowing what to measure with the test gear we have that correlates with what our ears hear. We aren't there yet. We have a decent foundation, but we don't have the entire building constructed yet.

Thanks...helpful further clarification. That has been my sense from reading these forums. Thanks.

 

[jkeny]

Could anyone tell me what might be a good measuring technique for noise modulation in S-D DACs as reported by Opus? Not challenging what has been said, simply trying to understand where we are with measuring abilities, currently.

 

[JackD201]

I suppose now that studies like these are starting to surface, we can be "allowed" to trust our ears.

 

[LL21]

I suppose now that studies like these are starting to surface, we can be "allowed" to trust our ears.

 

[treitz3]

Well now, there's a concept.

Tom

 

[bbb]

Less cliche, an observation from one of my college socialology classes: There were studies long ago showing driver's ed made better drivers and it was often mandated and taught in schools. Later studies showed the students taking driver's ed were often upper students and are better drivers with or without the driver's ed training. They weren't better because of the training, or not much, but were better for many, many other reasons.

Very similar to the findings contained in the book, Bounce by Matthew Syed. He writes that talent is overrated because successful people often are those who has put in the most effort. Purposeful practice of a minimum of 10,000 hrs is required before someone begin to excel at his/her particular fields and this includes sportsperson, musicians etc. This also applies to the so called golden ears. The key word here is purposeful practice. Just listening a lot is not purposeful practice. Many people claim to hear/not hear something but do they actually have the skills required to hear/not hear it in the first place. This is a major reason why we always have no confirmation in any test, dbt or otherwise because the participants do not have the same level of competency but somehow everyone will always claim to be a pro. Noticed the tests in the papers are aced by a musician and a sound engineer? And people whose fields of expertise related to music scores best? this is no anomaly, it is because they had trained to hear better. Now before someone jumps in and say, "but but but there are also musicians who fails in test", yes but they might not necessarily trained to hear what was being tested. Purposeful practice is key.

There are many other subjects in the book that is so relevant to our hobby.

 

[MylesBAstor]

Very similar to the findings contained in the book, Bounce by Matthew Syed. He writes that talent is overrated because successful people often are those who has put in the most effort. Purposeful practice of a minimum of 10,000 hrs is required before someone begin to excel at his/her particular fields and this includes sportsperson, musicians etc. This also applies to the so called golden ears. The key word here is purposeful practice. Just listening a lot is not purposeful practice. Many people claim to hear/not hear something but do they actually have the skills required to hear/not hear it in the first place. This is a major reason why we always have no confirmation in any test, dbt or otherwise because the participants do not have the same level of competency but somehow everyone will always claim to be a pro. Noticed the tests in the papers are aced by a musician and a sound engineer? And people whose fields of expertise related to music scores best? this is no anomaly, it is because they had trained to hear better. Now before someone jumps in and say, "but but but there are also musicians who fails in test", yes but they might not necessarily trained to hear what was being tested. Purposeful practice is key.

There are many other subjects in the book that is so relevant to our hobby.

Doesn't that fly in the face of Olive's results that there's no hearing differences between trained and untrained listeners?

 

[FrantzM]

Doesn't that fly in the face of Olive's results that there's no hearing differences between trained and untrained listeners?

The same Olive who some times ago, wrote about how to listen? in this very website (WBF)? If that was his conclusions, why would he write about how to listen? I find this very odd or is it taken out of context, if indeed he wrote such?

 

[MylesBAstor]

The same Olive who some times ago, wrote about how to listen? in this very website (WBF)? If that was his conclusions, why would he write about how to listen? I find this very odd or is it taken out of context, if indeed he wrote such?

Yes check out his results. Amir emphasized this data set. And it's not taken out of context. The context was that there was no difference between trained and untrained listeners in discriminating between speakers.

 

[FrantzM]

Yes check out his results. Amir emphasized this data set. And it's not taken out of context. The context was that there was no difference between trained and untrained listeners in discriminating between speakers.

So from the results that listeners preferences didn't vary whether listeners were trained or not you can state that Olive position is that training makes no diference on one hearing acuity when he , Olive and Amir stated many tmes that in order for people to discern certain nuances they must be trained ..

For those interested you can find it CLICK HERE

To save you going there this is the introduction to this blog piece, with apoligies to Sean Olive:

Trained listeners with normal hearing are used at Harman International for all standard listening tests related to research and competitive benchmarking of consumer, professional and automotive audio products. This article explains why we use trained listeners,

I can't see how this conflict with the post by bbb

Very similar to the findings contained in the book, Bounce by Matthew Syed. He writes that talent is overrated because successful people often are those who has put in the most effort. Purposeful practice of a minimum of 10,000 hrs is required before someone begin to excel at his/her particular fields and this includes sportsperson, musicians etc. This also applies to the so called golden ears. The key word here is purposeful practice. Just listening a lot is not purposeful practice. Many people claim to hear/not hear something but do they actually have the skills required to hear/not hear it in the first place. This is a major reason why we always have no confirmation in any test, dbt or otherwise because the participants do not have the same level of competency but somehow everyone will always claim to be a pro. Noticed the tests in the papers are aced by a musician and a sound engineer? And people whose fields of expertise related to music scores best? this is no anomaly, it is because they had trained to hear better. Now before someone jumps in and say, "but but but there are also musicians who fails in test", yes but they might not necessarily trained to hear what was being tested. Purposeful practice is key.

THis seems to corroborate the used of trained listeners thus training aka Purposeful practice in Sean Olive blog. AMir meeds no defense from me but those have been his positions as well ...

emphasis is mine.

 

[jkeny]

We have all the measurment capability we need, what we don't have, in plain old stereo, is a proper replication system of what our ears hear. There are no unknown mesurments, there is however, a severly limited replication system. And nobody has hearing better than the existing measureing equipment, nobody. And, like was said, simple tests used in specifications in the industry are not the only measurment systems that are and can be used.

VW will not pull a railroad car loaded with rocks. POS has its limits, and our measurments are far past its capabilities.

I'm not sure what this post actually means. If we have all the measurements needed can anyone point to a known set of measurements that characterise how a device will sound? It seems to me to be too easy to say that we have all the measurements needed but yet can't point to how they correlate to what we hear (even if you think the exist stereo reproduction system is POS). Are we expected to operate on a belief system that all measurements are available & that it's just that no-one is actually doing them?