Some real listening tests,
have been posted in the "everything else" forum. One is a test for passives and one a test of active circuitry. These are possibly the most "comprehensive" tests that have been run on the forum, comprehensive in the way that the files were produced and the way the test is run.
It is not a competition, the idea is to see if you can genuinely detect differences and what your impressions are.
Links are here together with a full explanation of the how the tests work in the Passives thread,
http://www.diyaudio.com/forums/ever...stening-test-part-1-passives.html#post3796009
http://www.diyaudio.com/forums/ever...test-part-2-active-circuitry.html#post3796014
have been posted in the "everything else" forum. One is a test for passives and one a test of active circuitry. These are possibly the most "comprehensive" tests that have been run on the forum, comprehensive in the way that the files were produced and the way the test is run.
It is not a competition, the idea is to see if you can genuinely detect differences and what your impressions are.
Links are here together with a full explanation of the how the tests work in the Passives thread,
http://www.diyaudio.com/forums/ever...stening-test-part-1-passives.html#post3796009
http://www.diyaudio.com/forums/ever...test-part-2-active-circuitry.html#post3796014
Thanks for the information, about this one i´ve never heard before.
There were similar observations reported by Newell and Holland:
http://www.edn.com/Pdf/ViewPdf?contentItemId=4015821
SY it gets kind of old to be told that I don't know what I am doing, that I am not current in engineering, etc.
By the way, is anyone going to the Analog Aficionados party 20 this year?
By the way, is anyone going to the Analog Aficionados party 20 this year?
It does sound interesting RM. I also think that both you and I have experienced speaker cable differences that were not easily explained.
It gets kind of old being told that we are only at stage "one", and then you spout 40 year old stuff like it's new to all of us.
I will note that I do enjoy reading everybody's posts on circuit topologies, strategies, noise calcs and equations, as I am not as well versed as many of you in those areas. It is always wonderful to learn new things from many who post here.
However, this post of yours is typical of you..
You believe the 'concept' of an IC making a difference, but did not engage the depth of the theory such that you understood the issue of ground loop and stray current paths. Instead, you continued to explain how the contact cleaner/cable metal non linearities were causing the problem. I pushed ground loop concepts on you, now we hear nothing about metal non linearities.
You believe the 'concept' of speaker cables making a difference, yet this discussion of details of test and measurement is "quibbling"?
You presented a 'concept' of skin depth vs gauge and frequency, yet the 'concept' was in error due to the use of the wrong equation.
Concepts without understandings is not science nor engineering. It is flailing.
jn
Depth of notch is starting reference point in exploring temporal coherence of speaker performance.
Short broadband bursts bandpass limited to 500-4kHz are highly revealing of speaker system imaging specificity. Divergence from coherent source is audible both in direct sound, and in reflected sound from room.
When I tune speaker to have null -24dB the sound is much different. Imaging detail is lost both in width and in depth.
Sorry Scott, you did go last year, didn't you? I put off jury duty until summer or I might have been there too.
JN I too double checked the page presented. I found it about 3 times LESS than the prediction of the skin depth. How can this be? Of course, I did NOT generate the table, I got it from a Google search.
JN I too double checked the page presented. I found it about 3 times LESS than the prediction of the skin depth. How can this be? Of course, I did NOT generate the table, I got it from a Google search.
Last edited:
A question where the party ? John take solace that you are one of only about 6 who's name is know outside of this board by anyone to whom they do not work directly with . Haters not so much . Speaker wire being for the most part un-shielded act as an antenna and a transmission line between the output and speaker. DCR being far less of a factor . I can not take the view point of anyone who steadfastly
argues that tin, copper and silver do not sound different when used in a cap if the magic polyprop is used for a dielectric . Further more that the quality of production above a certain level
does not effect the sound because we do not use the tool need to measure relevant data. I believe the use of the term dogma applies. I am not saying that audiophile cap are priced at other than what the market will bear . I just saying that how is it that different material can all sound the same because the wrong test where preformed and the score was the same for non relevant parameters.
Jakob2, this is a GREAT paper! Richard, be sure to read it too! Wow!
Barleywater, I am not criticizing your depth of notch. I must admit that I would not have thought it that important, but then I have this COMPLEX alignment that I have to do the with the SASHAS one day, probably for much the same reason.
Barleywater, I am not criticizing your depth of notch. I must admit that I would not have thought it that important, but then I have this COMPLEX alignment that I have to do the with the SASHAS one day, probably for much the same reason.
Those results are consistent with a loudspeaker being a somewhat non-linear load. This creates some distortion with just the amplifier output impedance (and/or short cable) and more distortion (but of exactly the same 'shape') with a longer cable.Jakob2 said:
The surprising thing is perhaps just how non-linear a speaker can be in impedance terms. To reduce the problem just use thicker cable, or a better speaker. Avoid SET, and other high impedance topologies, like the plague!!
I very much agree that the amplifier to speaker impedance ratio should be high . Set and other high impedance can work if the speaker is also of a high value and self controlled i.e. properly damped . There is no question that a SET driving a low impedance speaker will be all over the place and no where near flat . The love of that combination may be in it coloured sound . Many people some how prefer techincolor to real life in video and here to in stereo .
The paper is interesting, itself. Anyway, I have doubts. I have also done a lot of measurements of noise background on power amplifier output terminals and directly on speaker terminals, of course with speaker cables connected. I have never measured anomalies that can be seen in that paper. In fact there was nothing special, nothing unexpected in my measurements. Some RFI content of local AM transmitters, but only above 300kHz. Some content of TV transmitters, in expected level. But NOTHING in audio band or near audio band. When I see such results as in that paper I am skeptical and I am asking: was their test gear in a good condition? Did not they make some fatal setup mistake? Are the authors experienced enough to distinguish between measured phenomenon and measurement error? I assume, that if the plots as shown in the paper were usual, much more similar measurements from different authors would appear.
The exponential approximation equation is not very accurate for audio and typical speaker wire sizes. The bessels are needed to do it right. Bessels predict about 1/3rd the skinning the exponential predicts at 20 Khz and 1.5mm dia wires. There's your factor of 3.
Note also that the table you linked to is in essence worried about wire ampacity. As such, it really doesn't reflect speaker wire needs. Doubling the resistance of a 12 gauge wire will reduce it's ampacity significantly, from 20 amperes to 10, but for speaker wires, it reflects an increase of 1.7 milliohms per foot per conductor. If your running 10 amperes rms to an 8 ohm speaker in your front room, I think wire heating is one of your less significant problems..😱
I'd worry more about the non linearity of Rs in proximity effect.
jn
One thing I don't understand is how this talk goes on and on about every possible solution to making an amplifier that has little to no distortion and then the speakers are left out of this discussion besides how it screws things up? Before going down what I think Joachim calls the rabbit hole fix the impedance curve of the speaker with a real conjugate network and flatten the impedance curve. Then make a proper passive if you like or active x-over network. If you can't be bothered to do that then why fuss with 0.0001% amplifier distortion values, not missing the forest for the trees but missing the Earth for a puddle.
As for that notch and getting a perfect match at crossover that would only be for one exact position since the two devices are not on the same axis and usually not in alignment in depth either. You can bring that into a nice alignment for one exact position but nowhere else will that be true. The argument over on the speaker side is can we hear phase differences or not across the spectrum in human hearing? Many competing answers on that question, but just as you are chasing zero distortion number why not make it as good as possible?
As for that notch and getting a perfect match at crossover that would only be for one exact position since the two devices are not on the same axis and usually not in alignment in depth either. You can bring that into a nice alignment for one exact position but nowhere else will that be true. The argument over on the speaker side is can we hear phase differences or not across the spectrum in human hearing? Many competing answers on that question, but just as you are chasing zero distortion number why not make it as good as possible?
- Status
- Not open for further replies.