Don't know what to say to that... hmmm thanks 🙂
Yes it was difficult until I picked that "window" within the track. It seemed to have the sort of content where differences might show, and they did but the differences are minute in the extreme. And of course when there is a difference you have the next question "so which do you prefer ?" and that is to close to call on these.
Do you find these "internet" ABX tests show differences that you yourself might hear at home or do you think some of the finer detail and differences do get lost in the AD -DA and WAV process.
Yes it was difficult until I picked that "window" within the track. It seemed to have the sort of content where differences might show, and they did but the differences are minute in the extreme. And of course when there is a difference you have the next question "so which do you prefer ?" and that is to close to call on these.
Do you find these "internet" ABX tests show differences that you yourself might hear at home or do you think some of the finer detail and differences do get lost in the AD -DA and WAV process.
Though some of the minor differences might be lost, I have found that ABX comparator is able to say very reliably that there is a difference audible. Or that the difference is not reliably audible. I have learned a lot during all these tests and I have learned that preferences are not very important. Different people may have very different preferences, but to me it is important if they can hear the difference or not. This ability is also very individual.
Still haven't got the ABX thing happening, there is just something about foobar I "don't like", just tiny little idiosyncrasies that add up to make using it not my first choice.
For me, when I was listening in Nero, mango won out over kiwi. This echoes an earlier assessment of blue vs. green, where the first time through I preferred the direct path, but in a different listening the preference swapped. Each time there were definite differences noted, but the 'preference' decision comes down to how conditioned the playback path is, what specifically one is tuning into at the time to note variation, and even how much a row one had with others in the house an hour earlier, 😉. So, the differences are quite subtle ...
For me, when I was listening in Nero, mango won out over kiwi. This echoes an earlier assessment of blue vs. green, where the first time through I preferred the direct path, but in a different listening the preference swapped. Each time there were definite differences noted, but the 'preference' decision comes down to how conditioned the playback path is, what specifically one is tuning into at the time to note variation, and even how much a row one had with others in the house an hour earlier, 😉. So, the differences are quite subtle ...
Interesting that more people don't try this ... just using a cheapo label recording of Brahms Trios for Piano, Violin and Cello this morning to warm up the PC gear - and it's so easy to hear the progression from the start: first up, tiny, screechy violin, bassless piano, kiddy sound, wouldn't pay 2 bob for it ... slowly, slowly, the tonality builds ... the strings come to life, a sheen and shimmer starts to fill the air, the piano develops depth and authority - I go over and increase the volume by quite a few notches, this is steadily turning into music, a real life and energy are now in the air. But, at the beginning it just sounded like junk radio sound, nothing to do with "high fidelity" ...
Last edited:
I did the ABX test before tearing apart the waveforms. Rapid result from Mooly was good indication that a positive result was achievable, unlike the other too close to readily call comparisons.
I did exactly what Mooly did, I kept switching back and forth until I discerned a difference, and focussed on sense of change with changes between tracks.
Even with foreknowledge, either the switching between tracks causes sensation of change, or it doesn't.
Really ought to see how Mooly does with 183 samples trimmed off of mango. I didn't spend much time after two back to back fails, after having completed three sets of trials with big positives. Could be expectation bias; expecting to fail.
Please don't lose focus on what I have mentioned several times, even if you don't believe it.
1) Many experienced listeners can pick up difference in sound by focusing on a very narrow part of the sound (including a small digital glitch). So what?
2) Knowing there is a difference (this way) will at best lead us to a USELESS conclusion that everyone can have different preferences or tastes.
3) So, is there any other thing that is more than just "differences"? Yes, MUSICALITY/SONIC and FATIGUE. If you close your mind and think that class-B amp is as good as class-A amp, or class-D is better than others, may be nothing can change that. And one thing is for sure that you will never be able to design a product that will be successful in the audiophile market.
So, don't give up (on trying to get a useful result).
Related to blind test, let's move on, progressing to the "harder" listening practice. We don't listen for just "differences". But we feel for the fatigue and musicality. What is the point? The point is it is then less subjective. Because you can't say that you prefer a fatiguing sound, can you? (only later we can progress to knowing how to design circuit with/without those variables).
Two different sounds (of different songs) will surely different in ABX perspective. But the fatigue and musicality can be similar. This is what happen when people nominate for "the best sound of the show". We can use different songs as I have suggested. Why must be different song? To prove that there EXIST other variables that can be the same or different instead of just casual "differences".
Will this be difficult? Well, do you know anything of the real scientists spirit? They are not after any Nobel prizes and they are not afraid of failures. In this case, "just do it". Hopefully there will appear someone who can "solve" it.
Frank, adaptation of human hearing is continuous. 10 minutes may condition mind to all sorts of states. Nothing like some good foreplay to zero in on sensitivity, or tune out many other details as unnecessary.
With kiwi/mango, forcing mind to assert a continuing difference requires changing back and forth every few seconds. Only holding onto difference for minutes or longer leads to preference. Strong preference is readily triggered by previous preference forming events.
If forced to listen to 15-30 seconds with fade in, fade out and then be able to repeat listening, or switch to other version, most would be reduced to random status in identifying these tracks.
The fatigue and musicality aspect is somewhat different, in that in many cases it is cumulative - in that sense switching between A and B doesn't work: if A is more 'tiring" than B then after a period of doing back and forth one feels that one is getting tired - but was it because of A, or B, or perhaps both of them 'did it', together?
The great difficulty in all this is, that the mind works hard to compensate for little problem areas it detects in the sound - it wants to 'nicefy' it, a lot of the time! But, the fatigue element is the giveaway that mental energy is being expended to achieve this - you are dealing with 'poorer' sound in that particular case ...
Yes, listening 'adaption' is often proposed as an explanation for this sort of thing. However, there are all sorts of ways of skirting this - otherwise, no-one would be able to assess anything in the sound field - a Stadivarious would end up sounding the same, playing a certain piece of music, as a "roughie" picked up from a pawn shop. A good techniques is to put on a piece of complex, aggressive music at high volume, and attempt to concentrate strongly in a direction completely unrelated to music - your brain will very rapidily tell you the true status ...
Preference comes about after noting changes - for average quality playback I would go for that version revealing most detail, with accompanying treble clarity .
Edit: Bizarre, for a moment diyAudio considered "assess" to be a 'bad' word
Last edited:
Jay;
Success in audiophile market has next to nothing to do with having a product; only ability to sell product.
Music and musicality go with musical instruments and musicians.
Reproduction is recreation via a representation where transduction occurs from mechanical energy to electrical energy and back to mechanical energy. Accuracy can be assessed in purely mathematical realm. The better the correlation, the higher the fidelity.
With recorded signals, any portion of the signal may be listened to as periodic occurrence and allows hearing perception to be used as differential processor for two systems via ABX switching system. With practice fairly fine differences in timing and amplitude become discernible.
When listening for pleasure, fatigue is loss of pleasure. When listening for differences, it is loss of ability to discriminate. In both cases fatigue is when listening is no longer enjoyable or productive.
I know the importance of marketing and selling skill but product is important too. Please mention one good example where one bad product has gained success in audiophile market (and a non-successful good product for comparison). Do it in a "fair" way (isn't fair for example, if you mention a good product that I have never heard about, no distribution channel, etc).
Of course, fatigue and musicality are two different aspects. Mooly has stated that sonic (which usually go hand in hand with musicality) differences of the mango and kiwi was so small that the final sonic will be more determined by amplifier-speaker quality. I also think so for sonic, but for fatigue I still think the GIGO applies.
It doesn't mean that you have to do it that way. It is not a new thing actually. Have you ever heard statements like this: "I have built better and more expensive amps but I don't know why I have always go back to my X amp". I have examined this for years. And I trained myself for this almost everyday when I work with my speakers or amps.
I have mentioned several times that my ultimate test is to listen to the system while going to sleep. Why?
1) When we wake up from sleep, we have the most fit physical condition, the best ability to hear/feel differences.
2) At night, in the morning, external/environment noise is so very low. Electrical condition is so clean from interference.
3) We are tired/sleepy and we want to sleep. So any small disturbance such as slight peak and fatigue will makes us go crazy.
So there are 3 different results (from the worst):
1) You want to turn it off because the sound disturb you. You may go to sleep but then wake up in the middle of the night and turn it off. Often you cannot sleep well, you will feel distressed in the morning.
2) You don't feel anything. The sound system goes unnoticed.
3) In the morning when you "open" your consciousness (before your eyes) you will hear relaxing beautiful music, even if it is a Metallica.
The "ABX" I proposed is to ensure that the listener can pick differences in fatigue or musicality. Not really ABX but the third presentation should goes through the same path with one of the 2 other files. More files to increase validity. But each with different songs so no cheating is possible (so any positive result will prove more "claims").
Last edited:
Square wave performance of any capacitive coupled circuit is not exemplary. Here is 100Hz square wave with 3.3Hz Butterworth filter:
This is before mass/spring system of transducer further twists waveform. Square wave tests of best headphones/ear buds tend to look even worse. Is it audible? Fairly easy with ABX.
Asymmetric waveforms, such as sawtooth waveforms are more subtle looking in differences, but provide two sets of harmonics with complex phase structure; these waveforms sound different when polarity is inverted, or when reversed in time, or inverted and reversed.
Human voice, bowed instruments, woodwinds, all tend to have significant asymmetry of waveforms for any sustained note. Direct sound from such sources is very distinctive. These distinctions are rapidly degraded by room reflections. For live sound this is one of the distinctive differences that convey proximity of source, and sound of room/space that sound is produced in. For reproduction, controlled reflections can add or remove contrast in a frequency dependent way.
Good headphones/ear buds provide more consistent reference for making fine differentiations.
Speaker systems/headphones that have been equalized to linear phase response (with flat frequency response) across full listening spectrum provide most uniform reference.
You mean the two tracks aren't quite running together...
That wouldn't make any difference to me because I pick out and listen for a difference occurring in just one tiny part of that clip such as the attack/decay of a particular percussion instrument etc and try and keep focussed on just that sound. I then play the other version and see if it sounds as I remembered it or is it "different" in some way such as attack or decay. The fact the tracks don't run together doesn't matter because I find switching between X or Y (for me) doesn't work as well.
Do you expect everything else in the recording and playback chain, except for this CR input coupling 1st order HP, was DC coupled and able to transfer 0Hz, i.e. DC signal? I guess you do not. Please check headphone 30Hz square response. I am sure you also know how speaker square response look like. Your comments are misleading. There is nothing like square waveforms in music signals. There is no microphone/preamp system that would capture DC and there is no way how to produce positive or negative DC air pressure in the non-sealed environment.
Attachments
Last edited:
Yes, that's what he means. He made a multitrack comparison (I did later too) and found 183 samples difference, which is 3.8 milliseconds exactly. No one else did that. He is the one who makes analysis rather than ABX listening comparisons. My fault I did not make multitrack. Anyway, I consider all other results except for his as valid. There have not been many screenshots here at all, so I do not care.
Goes without saying if you find one sample has an identifiable glitch but these don't. I'm listening purely to the music and can pick the difference out from just that.
Again, goes without saying, but when you hear a difference with the music, with the sonics, and its absolutely minute then it does become difficult if not impossible to say which is preferable. On the other hand, the tests are useful for helping design circuitry down to a level where it becomes impossible to tell X from Y. If you get an inconclusive result from all participants then the circuitry has to be pretty blameless and transparent.
Musicality is something I have banged on about for years, and in some ways goes against what I have just said above with the true test of a complete audio system being how much you enjoy listening to it, how much it makes you want to explore the music... and this is where it becomes as much an art as a science. For small signal circuitry I think accuracy is important, its in the poweramp/speaker/speaker interface areas that great audible differences occur.
It is true that some listeners (probably me included) find that listening to what is imperfect circuitry or circuitry that for example interacts with a speaker less than perfectly can be preferable.
Consider this, and would this work in any way... take two amplifiers, a blameless and say a tube amp both connected to real speakers and playing music at say 1 watt average level. If you made a WAV file of the output of the tube amp (taken from across the speakers) and played it through the blameless... what would you hear ? And do the same for a resistive load to.
Mooly, I have analyzed the difference file of the part of the music that you correctly suggested as a hint for finding of audible difference.
Please find the fft attached, together with the sound file in zip. (no glitch there)
Attachments
- Status
- Not open for further replies.