Alternative method for measuring distortion

[rajapruk]

Even if the df-metric overemphasizes phase errors in comparison to what is considered hearable, I don’t care!
I want my audio electronics to be as perfect as possible, not corrupting signal in any way. I want to be able to choose electronics after this df-measure rather than SINAD, I think.
@amirm could it be included in your electronics review as an extra part, like a second opinion?

 

[rajapruk]

To me it is very interesting that you point out the Forsell dac, Serge.
I kind of know the Swedish guy designing this one (I think). At least he did one Forsell dac.
He Is a guy who goes his own way, and always in the right direction (I beleive).

 

[Serge Smirnoff]

I kind of know the Swedish guy designing this one (I think). At least he did one Forsell dac.
He Is a guy who goes his own way, and always in the right direction (to me).

If you crisscross with him by chance ... show him this thread, may be he be inspired by the idea ))

 

[pkane]

May be we should stop all research and measurements? What the audio science and this forum are for then? Good questions BTW. We can ask @pkane: why you develop your application, what will be the result of research of audibility of distortion? Some audio metric? What is your goal, pkane?

DeltaWave purpose is to help measure differences between devices. It wasn't designed to come up with a single number to quantify this difference. It was designed to measure distortions using real music files and to allow them to be studied and compared. Aside from RMS difference and A-weighted difference, DW computes correlated depth, group delay, variable group delay, jitter, frequency spectrum, cepstrum, linearity, etc. It lest you separate files into frequency bands to be compared separately. But these are all still 'engineering' metrics.

The next step was always to determine whether these differences are audible or not. DW allows you to listen to the difference, but more than that, the DISTORT app allows you to test various metrics to determine your own, personal audibility threshold. Including things like jitter, harmonic distortion, noise floor differences, etc. DISTORT isn't done yet, as I plan to add more distortions related to phase and frequency response.

 

[rajapruk]

If you crisscross with him by chance ... show him this thread, may be he be inspired by the idea ))

I have sent him the link.

 

[Serge Smirnoff]

I have sent him the link.

Great. Thanks.

 

[solderdude]

May be we should stop all research and measurements? What the audio science and this forum are for then?

I don't get your conclusion that it is all pointless.
Sites like ASR can show measured performance, broken down in aspects of equipment and now also speakers.
Forums are to discuss and learn.
You get a chance to communicate with others that have similar and dissimilar interests and thoughts.
How can one not learn from this ?
What could possibly be gained by stopping measurements and audibility studies ?

 

[solderdude]

Do you think that it is not possible to manufacture such devices or that this will be economically inefficient? What do you want to save/economize?

I already gave examples of cheap equipment that already make low distortion devices and there are plenty expensive ones that do that as well.

Why are you afraid to set really high (overkill) standard?

I am not in the least. Everyone is free to choose what overkill is needed and in what area.
I am in favor of proper engineering and not over-engineering for the sake of over-engineering or to increase prices to absurd levels.

... prefer to set lower standard but perform a huge amount of really complicated and expensive research of audibility of distortion?

I am not preferring to set any low standard at all. It would be fun to set some audibility levels of certain aspects. This research has already be done... extensively. It is known what is audible and what not, but it is complex to get the full picture. One can relax audibility thresholds depending on certain factors as well.

I like your idea of a single metric with cheap test equipment and software doing the analysis. I just think you have a long way to go before I (and some others) start to take it seriously.

 

[rajapruk]

Can amplifiers be tested with this type of messurements?

Cables?

Speakers?

 

[solderdude]

Can amplifiers be tested with this type of messurements?

Cables?

Speakers?

amplifiers and cables yes, speakers and headphones no.
At least it won't give any meaningful results in the latter 2 cases.
for cables I would say a direct null would provide more reliable info.

 

[rajapruk]

Great. Thanks.

Was it Forsell or Forssell dac? I think I made an misstake here.

The guy I know made an dac called Forsell (with a single S).

There is also an american studio brand Forssell:
https://www.forsselltech.com/#!/products

 

[Serge Smirnoff]

The next step was always to determine whether these differences are audible or not. DW allows you to listen to the difference, but more than that, the DISTORT app allows you to test various metrics to determine your own, personal audibility threshold. Including things like jitter, harmonic distortion, noise floor differences, etc. DISTORT isn't done yet, as I plan to add more distortions related to phase and frequency response.

Is DISTORT intended mostly for personal researching/playing with thresholds or you have plans to aggregate those findings in order to discover some figures that will be common for broader audience?

 

[Serge Smirnoff]

I don't get your conclusion that it is all pointless.
Sites like ASR can show measured performance, broken down in aspects of equipment and now also speakers.
Forums are to discuss and learn.
You get a chance to communicate with others that have similar and dissimilar interests and thoughts.
How can one not learn from this ?
What could possibly be gained by stopping measurements and audibility studies ?

I asked these provocative questions in the context of your:

Only very few people really care for perfect waveform reproduction. For them devices already exist in all price ranges.
The rest really, really does not give a damn about any measurements.

I'm trying to understand - is ASR forum a place where people just exchange their knowledge/information/experience or they also want to, say, "make the audio world better"? In other words, what are prevailing intentions/sentiments of ASR community - to influence or to observe/review?

 

[Serge Smirnoff]

I am in favor of proper engineering and not over-engineering for the sake of over-engineering or to increase prices to absurd levels.

I am not preferring to set any low standard at all. It would be fun to set some audibility levels of certain aspects. This research has already be done... extensively. It is known what is audible and what not, but it is complex to get the full picture. One can relax audibility thresholds depending on certain factors as well.

I understand the psychological difficulties of accepting over-engineering. But here is my reasoning. I tried to say above that thresholds of audibility of various distortions depend not only on characteristics of hearing but also on listening tastes/habits of people. Those thresholds are not solid, they change from individual to individual and also can change in time for an individual. And this is the area of cognitive psychology, not even psychoacoustics. I'm pretty sure (but can mistake of course) that in this situation an objective audio metric, which have good/reliable correlation with perception is impossible in principal. It will be toooo hard to "get the full picture". In this particular situation the over-engineering can be really helpful and grounded. The more so, in this particular situation over-engineering will lead to dropping the prices on the audio market, not increasing them (commoditization). Education is another story but I also have some ideas for the purpose.

 

[solderdude]

what are prevailing intentions/sentiments of ASR community - to influence or to observe/review?

That could well be quite different for community members.

I can only tell you what my intentions and sentiments are.
To learn and to help other community members and casual readers and to discuss relevant and less relevant matters I take an interest to.
Of course all from my P.O.V.

 

[solderdude]

in this situation an objective audio metric, which have good/reliable correlation with perception is impossible in principal

Exactly my point. The desire to have a metric that can tell people HOW it will sound to them does not exist. One can have a metric(s) that tells something about signal integrity/quality.

Over engineering is not needed at all. Take the Atom and Heresy for instance.. nothing over engineered and will have good metrics. If it doesn't something is wrong with the way the metrics are obtained or the DUT is broken in one or more ways.

 

[Serge Smirnoff]

amplifiers and cables yes, speakers and headphones no.
At least it won't give any meaningful results in the latter 2 cases.
for cables I would say a direct null would provide more reliable info.

I agree in general. With the caveat. A metric dealing with accuracy of waveforms can be used/accustomed for speakers/headphones, but it will be less effective. I think it would be more effective just to have the same/similar transducers on both sides - in studio and at listener's premises. Then both creators and listeners will have similar listening experience, which is the primary goal of any sound reproduction I think.

 

[solderdude]

The acoustical output of a speaker, certainly at the listening position will look nothing like the original waveform.
The differences will be big enough to get DF numbers around -6dB or so yet sound fantastic.
Same is true for headphones.

 

[Serge Smirnoff]

That could well be quite different for community members.

I can only tell you what my intentions and sentiments are.
To learn and to help other community members and casual readers and to discuss relevant and less relevant matters I take an interest to.
Of course all from my P.O.V.

Thanks, I've heard you.
How do you think, if I suggest mass testing of modern portable audio devices with df-metric for the purpose of creating alternative view of the situation on this market, will this be interesting for ASR community? I'm asking because I'm new here and can not read even a representative portion of the threads/posts in order to understand if people are ready to act, not only to discuss.

 

[Serge Smirnoff]

The acoustical output of a speaker, certainly at the listening position will look nothing like the original waveform.
The differences will be big enough to get DF numbers around -6dB or so yet sound fantastic.
Same is true for headphones.

Yes, I know about all those complications and do not have illusions or high expectations about that )).