Perhaps "argue nonsensically with TMM that there is more to audio distortion than linear, non-linear and stochastic noise" would be more appropriate.
There is a massive 'con' in my reasoning for Remote Subjective Blind Auditioning on the previous page which i glossed over. Surprised that no one has yet pointed it out since it's a big and obvious one.
There is a massive 'con' in my reasoning for Remote Subjective Blind Auditioning on the previous page which i glossed over. Surprised that no one has yet pointed it out since it's a big and obvious one.
Music to my ears 🙂 all in all that helps remove an uncertainty from the obtained results (Dave likes to quote probabilities). So besides the fact I personally don't like Zoom anything, there is no criticism from me in the relative validity of your tests XRK. Besides....if anyone can get two measures on two separate days reading within 1dB, then there is little left to moan about 😉
Which are? Noise and distortion is divided into 3 categories:
1. Non-linear Distortion (Harmonic, Intermodulation)
2. Linear Distortion(Frequency response, Energy storage (Resonances), Reflections, Delay)
3. Stochastic Noise (Background noise, 'hiss')
There is nothing more than that scientifically. All 3 can be objectively measured.
One 'noise' that I have noticed audibly,but found difficult to isolate in measurement graphs is the radiation from the surround. In one driver I had it was really very obvious in audition, but not apparent in impedance or easily seen (by me) in frequency response.
Another example I recall is the resonance in Alpair 5 in the 400-800 Hz range. Small kink.in impedance, but little to show it in frequency response. But audibly it was very apparent and in measures by another person very apparent in THD.
Not edge resonance (I don't think) but the contribution to output that the surround makes.
I'm guessing its a noise rather than a harmonic effect (but I really don't know enough to be sure, and have found it difficult to see in data)
I'm sure it does show in data, but how would it be found?
Another example I recall is the resonance in Alpair 5 in the 400-800 Hz range. Small kink.in impedance, but little to show it in frequency response. But audibly it was very apparent and in measures by another person very apparent in THD.
Not edge resonance (I don't think) but the contribution to output that the surround makes.
I'm guessing its a noise rather than a harmonic effect (but I really don't know enough to be sure, and have found it difficult to see in data)
I'm sure it does show in data, but how would it be found?
I am (we are) equally interested in difference between bad and good, which xrk971 tests clearly showed so far. Mark Audio Alpair comes to my mind...
It would be nice if the test also show difference between good and great, by I (we) will be very satisfied with the results discerning good from bad only. We need this kind of test, because manufacturer graphs are not always reflection of the truth. Again, Mark Audio Alpair comes to my mind...
All 'noises' caused by speakers are either linear or non-linear in nature. Sometimes both.
These are Zaph-audio's measurements of the Alpair 6 which has two resonance issues, a major one at 1.9kHz and a minor one at 800Hz. Issues show up in all his measurements at those frequencies.
(Linear distortion) A dip/peak combination in the frequency response:
(Linear distortion) Small blip in the impedance:
(Non-Linear distortion) Increase in harmonic distortion:
(Linear distortion) A ridge in the cumulative spectral decay:
Without being an expert at designing speaker drivers, I'd take a guess that it is caused by an impedance mismatch where the edge of the cone transistions sharply to the reverse half roll of the surround. The issue can be helped by applying a flexible doping compound to the surround at the fold in the surround, which is why you see a lot of drivers with a sticky compound applied to either the front, back or both sides of the surround at that point.
These are Zaph-audio's measurements of the Alpair 6 which has two resonance issues, a major one at 1.9kHz and a minor one at 800Hz. Issues show up in all his measurements at those frequencies.
(Linear distortion) A dip/peak combination in the frequency response:
(Linear distortion) Small blip in the impedance:
(Non-Linear distortion) Increase in harmonic distortion:
(Linear distortion) A ridge in the cumulative spectral decay:
Without being an expert at designing speaker drivers, I'd take a guess that it is caused by an impedance mismatch where the edge of the cone transistions sharply to the reverse half roll of the surround. The issue can be helped by applying a flexible doping compound to the surround at the fold in the surround, which is why you see a lot of drivers with a sticky compound applied to either the front, back or both sides of the surround at that point.
Is that why all my pro audio cloth surround drivers have that sticky coating? Picks up dust and lint like crazy.
Yes and no, its the goop that they put at the transition point of the cone and the surround, check out all the AR drivers from the 70s and early 80s they all have a fillet of sticky stuff where the half roll starts on the cone edge, its like the shock absorber inside the "spring" of the surrounds suspension(auto/car analogy) which helps to stop the sound wave bouncing back into the cone
definitely.
having played with MA audio. they were measuring far from their published measurments.
thanks a lot for this test xrk971
Id personally like to have also subjective observation with each drivers with a xo say of 200hz and 2 khz. and also fullranged....
Last edited:
You're welcome! Why would you like to hear a full range xo at 2kHz? Just to check its ability as a tweeter? It can be done but then requires it to be used in a 3 way or a 2 way with a good wide range woofer. I suppose the Dayton RS225 could work in this instance. Easier would be to put it in my B&O xo testbed which has a 12in woofer, and room for 3in to 5in drivers.
I am preparing for round 3 and trying to come up with a universal test rig that will provide better sound quality.
that would be very cool.
1: test the drivers fullrange
2- test them with a 200hz xo in a fast system.
3- test them in a midrange situation only with a 200hz xo and a 2khz or even 1.6khz cut off point. simply to see their potential as a dedicated midrange.
I'm personally a big believe in low crossover point to a tweeter that can handle low xo point as the ear is the most sensible between 2khz and 5khz
maybe im repeating myself and others have asked you to do this!
Last edited:
Forgot you a subjective camp guy ? normal arguing there is more under the surface to defend when a MA driver with many times worse response than X's Zoom recorder is on the table. At a diy sector comparison if Zoom recorder has a peak somewhere is not so big a problem when all drivers sat in same chain only variable was DUT itself.
That is one driver. And we know it has a wonky response. The designer told us that and why. That it (actually a test unit CHN70eN not a stock driver) does other things very well seems to be ignored -- or lost. It just happened to be one of the other drivers kicking around when we 1st listened to the TC9.
That it has a wonky response has nothing to do with the issues with the test. The Zoom will affect the response of all the drivers.
dave
That it has a wonky response has nothing to do with the issues with the test. The Zoom will affect the response of all the drivers.
dave
Surprised you say "that is one driver" because yes at round 1 it was, then what about round 2 is A7.3 not wonky. A7.3 was beforehand known to look wonky as it did there reflected from other shared measurements but hard to see in datasheet.
Also the comment designer told us, yes afterwards consumers had their new toys, but not reflected in datasheet.
Agree it probably could do other things well but guess we enter a three way system then.
- Status
- Not open for further replies.