The 3-ways only I have always been moved by a single theme, but I think it is more because of the interpretation.
Nessum dorma by Franco Corelli, well recorded in 1966. The best Nd, better than the super known Luciano Pavarotti.
nessun dorma franco corelli 1966 - Google Search
The Youtube videos have very bad sound.
Nessum dorma by Franco Corelli, well recorded in 1966. The best Nd, better than the super known Luciano Pavarotti.
nessun dorma franco corelli 1966 - Google Search
The Youtube videos have very bad sound.
Let me try to explain a few things such that you will see that nothing here is really confusing or difficult to understand...
There are more than one way how human can be pleased. When we look at paintings, I guess that colorful pictures are more pleasing to the eyes than black and white.
Now when we push a needle to our skin, we feel pain instead of joy.
In music we can find similar things to the above. Distortion is like needle. Whatever the order, whatever the profile, it is never 'pleasing' to the ears.
But we can as well appreciate sound the way we appreciate paintings (what is wrong with that?).
I understand that ivanlukic (and PMA too) tends to look at the distortion as a needle. We don't usually apply such sensing in ABX test. I believe that PMA understands that it takes completely different skillsets to perform in ABX than to feel whether a sound is pleasing or not.
I'm guessing that those who can hear unpleasantness in sound do have more sensitive ears that I believe that it is just a matter of time and practice and experience that they will be able to learn the skillset to hear sound in microscopic or short scale/duration like in an ABX.
From my site, 2008
[Spanish] Musica de Cine EMI Classics (1991) -> Franco Corelli: “Nessun dorma” del Turandot de Puccini (1966) | Nauscopio Scipiorum
to English with Google:
[Spanish] Musica de Cine EMI Classics (1991) -> Franco Corelli: “Nessun dorma” del Turandot de Puccini (1966) | Nauscopio Scipiorum
An externally hosted image should be here but it was not working when we last tested it.
to English with Google:
Last edited:
I still don't understand how it is possible to listen to a sample just once, yet also choose the point where the distortions are highest. Do you need to get someone else to listen for you, and choose the right point in the sample? Maybe your computer can do it?
It has been known for decades that many people prefer sound with a little low order distortion present. This will be harmonics plus intermods. The fact that they may dislike the distortion on its own, if they heard it on its own, does not negate this observation. Why there is this preference is unclear; the usual explanation about instrument overtones is weak, because overtones are not necessarily harmonics. I suspect it may be that most people never hear relatively undistorted sound reproduction, so just prefer what is familiar to them. When FM radio first appeared a lot of listeners complained about the thin shrill sound; they preferred the 'warmth' of AM (e.g. restricted frequency response, extra noise, significant distortion especially on peaks).
It has been known for decades that many people prefer sound with a little low order distortion present. This will be harmonics plus intermods. The fact that they may dislike the distortion on its own, if they heard it on its own, does not negate this observation. Why there is this preference is unclear; the usual explanation about instrument overtones is weak, because overtones are not necessarily harmonics. I suspect it may be that most people never hear relatively undistorted sound reproduction, so just prefer what is familiar to them. When FM radio first appeared a lot of listeners complained about the thin shrill sound; they preferred the 'warmth' of AM (e.g. restricted frequency response, extra noise, significant distortion especially on peaks).
I was asked, by my colleagues at our local forum, to prepare a new test with more contemporary recording that would in 24-bit resolution. So I took their recommendation and prepared the same test chain with new files, that may be downloaded from
Dropbox - skot.zip
This time, I am not indicating which file is which, it is up to you, dear listeners. And tell me which one you prefer and please attach an ABX protocol, in case you have any.
Dropbox - skot.zip
This time, I am not indicating which file is which, it is up to you, dear listeners. And tell me which one you prefer and please attach an ABX protocol, in case you have any.
Well this test seems at least to show that H2 up to -50 dB is almost inaudible.
Don't you think -50dB of H5 or H7 certainly wouldn't yield the same subjective result? In that case the distortion profile has an inflluence on the outcome.
For me, the test still confirmed that chasing lower distortion than -60dB for H2 (and monotonically falling levels of higher harmonics) has no sense other than for academic reasons in case of diy, or for certifications and sales brochures from a company standpoint.
I agree with you that -50dB of H5 or H7 would probably give a different result, but, an audio component having -50dB of H5 or H7 distortion components would be considered as defective by myself. I have not considered something like that in my contemplation.
In addition, H5 and H7 of -50dB without, at same time, H3 of some -30dB or more are, in technical circuit practice, rather impossible. So -50dB of H7 would automatically indicate to very high H3 and this both would be most probably easily audible.
I think you understand that to some point he's just guessing. He is against the ABX so is fair to say that he does not really understand what to do to succeed in an ABX.
Listening a sample just once is difficult but will yield useful result if you believe that distortion is painful. Listening to short range, even if you can hear differences easily, doesn't mean that you will know what you prefer.
Looking for a point with higher distortion is almost impossible. If the difference in dB is subjectively high, looking for a peak/clip is probably possible, for example when the singer sings high notes.
First question is in what way they prefer the sound? Like people prefer the Rembrandt over Picasso? if so, nothing to argue.
I don't know the usual explanation, but i think this is a critical point. When an instrument relies on overtones, especially with the non harmonic ones, to sound good, it is then important that the amplifier should be able to reproduce all these correctly. Vivaldi will not be able to play good music with cheap violin.
I don't know how FM were when they first appeared in your area, were they so good? When the environment is ideal, FM will outperform AM, but how ideal is it for the majority? The broadcaster, the cheap receiver, the distance, illegal broadcast, etc. Sometimes the signal changes from stereo to mono to stereo... frequency deviation as no channel separation between stations...
No. You are wrong. When I downloaded the foobar plug in for ABX and saw what methodology the test use I did not even try the ABX test. I immediately understood that the test is useless. But you are right that I am against the test.
All I did was to play samples once. I was not guessing. I clearly heard the distortion in sample1. I repeat: the test is purpose made to prove that PUTs can not discriminate between samples.
Last edited:
We have a member, at our local forum, who achieved a 20/30 result and posted a valid test protocol. This, IMO, may be considered as success. Probability that he was guessing is <5%. This is his protocol:
Code:
foo_abx 2.0 report
foobar2000 v1.3.6
2018-11-16 22:04:36
File A: koza.flac
SHA1: 44fdba15dd770f9dbc09ba0cbfc7fc131a11e1a3
File B: ovce.flac
SHA1: 01086c25cd350d102906e8b1e8c36a4daa8210ef
Output:
WASAPI (event) : Reproduktory (ASUS Xonar Essence ST Audio Device), 24-bit
Crossfading: NO
22:04:36 : Test started.
22:05:43 : 01/01
22:05:58 : 02/02
22:06:14 : 03/03
22:06:32 : 03/04
22:07:02 : 03/05
22:09:14 : 03/06
22:09:33 : 04/07
22:09:52 : 04/08
22:10:50 : 05/09
22:11:09 : 05/10
22:11:53 : 06/11
22:12:08 : 07/12
22:12:18 : 08/13
22:12:34 : 09/14
22:13:02 : 10/15
22:13:15 : 10/16
22:13:26 : 10/17
22:13:58 : 11/18
22:14:10 : 12/19
22:15:45 : 13/20
22:16:38 : 14/21
22:16:52 : 14/22
22:17:08 : 14/23
22:17:29 : 15/24
22:19:02 : 16/25
22:20:23 : 17/26
22:21:57 : 17/27
22:22:11 : 18/28
22:22:34 : 19/29
22:22:51 : 20/30
22:22:51 : Test finished.
----------
Total: 20/30
Probability that you were guessing: 4.9%
-- signature --
c6084dd3becbe0b197657e65a35f818d50968595Like all other methods, it is whether you can interpret the result.
Then it should be easy for you to know which one is which? Whether people will deliberately, or mistakenly draw a wrong conclusion, that's different issue.
Thanks for the new files, Pavel.
Yesterday, I tried a rather quick ABX of the "trees" files and failed as already in a sighted compare I couldn't find a section that has differences prominent enough to withstand the "ear fatigue" if you have only a weak indication ("stomach feeling") and try to concentrate on that in quick A/B of a short snippet. From previous vast experience with ABXing (testing absolute polarity, allpass/crossover phase shifts and such) I know that I can't get any useful results unless I have trained a lot to quickly pinpoint a difference once I've "learned" the characteristic details to listen for.
Looks like the person with the 20 of 30 score found some of these hot-spots where he or she felt confident enough to use a short snippet in repetitive quick A/B to id the files, probably listening in a very anlytical/technical fashion, even with some fatigue that for sure has kicked in. At least the time stamps do suggest that.
That's the whole deal with ABX, you got to master the "repeat, repeat, decide" fatigue in some way (you have to play A and X at least one time each), and the less confident you are about differences the more training, time and patience it takes to find something with ABX even when "stomach feeling" in sighted casual listening is strong. Prolonged more or less unsuccesful trials with same piece of music also quickly tends to spoil the usage of that piece forever, the moment you start to play it the bad feelings evoked again kill any joy of listening, for me at least. This is a known deficit of ABX one has to be aware of.
I'll try the new files later and see where I get to. The 20/30 result is encouraging at any rate.
Yesterday, I tried a rather quick ABX of the "trees" files and failed as already in a sighted compare I couldn't find a section that has differences prominent enough to withstand the "ear fatigue" if you have only a weak indication ("stomach feeling") and try to concentrate on that in quick A/B of a short snippet. From previous vast experience with ABXing (testing absolute polarity, allpass/crossover phase shifts and such) I know that I can't get any useful results unless I have trained a lot to quickly pinpoint a difference once I've "learned" the characteristic details to listen for.
Looks like the person with the 20 of 30 score found some of these hot-spots where he or she felt confident enough to use a short snippet in repetitive quick A/B to id the files, probably listening in a very anlytical/technical fashion, even with some fatigue that for sure has kicked in. At least the time stamps do suggest that.
That's the whole deal with ABX, you got to master the "repeat, repeat, decide" fatigue in some way (you have to play A and X at least one time each), and the less confident you are about differences the more training, time and patience it takes to find something with ABX even when "stomach feeling" in sighted casual listening is strong. Prolonged more or less unsuccesful trials with same piece of music also quickly tends to spoil the usage of that piece forever, the moment you start to play it the bad feelings evoked again kill any joy of listening, for me at least. This is a known deficit of ABX one has to be aware of.
I'll try the new files later and see where I get to. The 20/30 result is encouraging at any rate.
Last edited:
Depends on whether you are into audio or hi-fi. Audio is about making a sound I like; hi-fi is about making a sound which is a reasonably faithful reproduction of the original sound. People often confuse these two quite different goals.johnego said:
Yes. The BBC always transmitted good quality sound. In those days it had good sound engineers. The main 'problem' with FM back then was that it was so good that it was too different from what people had been used to.
Broadcaster good. Early receivers with FM were mostly at the better end of the price/quality range; cheap receivers were AM only. No illegal broadcasts on FM back then. Few stations, so huge channel separation.
Last edited:
Thank you Maty, I like to listen to singing on my speakers and is the benchmark for sound. I only listen to vinyl for serious music.
I will repeat myself, nothing is further from emotion in music than POLY cones. Kevlar is alright to very good, especially listening almost full range the 805D from B&W is the best modern speaker, listened with Class D separates (around 4K for the power amp and 3K for the Pre). Carbon/woven carbon are also almost 0 emotion cones.
Saba, and other types of alnico are alright, not better much than carbon or new paper cones.
The best, by far is the old full-range Phillips 5/or6 inch driver listening quasi-open baffle (remove the back panel, less bass but sound becomes 5 times more pure of any resonance). You have to listen with a tube 845 amplifier to get the best, guarantee you will cry listening to Italian soloists. This is a serious departure from hifi... with a bandwidth of maybe 100 - 5000...
Last edited:
Are you sure you are in the right thread?
Witch orchestral music, the best is a BIG room and open bafles. The problem of open bafles, usually, are the other musics. Maybe the open bafles made/kit by Danny Richie...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.