Jump to content
STC

Blind tests can be unreliable.

Rate this topic

Recommended Posts

Here are two audio files recorded using simple binaural microphones.  One file is with much higher room RT and another one by almost 6dB lower. The difference can be readily audible when heard in the listening room. However, I find using quick AB method usually would not reveal the difference readily when the change is from long RT to shorter RT. This temporary inability can be explained by how our ears adjust to filter out extra reverbs and takes time to adjust when a sudden reduced reverbs takes place.

 

Anyway, this is to show that there can be some unique occasions where ABX cannot prove that you could hear the difference.

 

I did a quick ABX and didn't score 100% although I believe I could after some familiarizing which would defeat the purpose. Here are the results:-

 

foo_abx 2.0.5 report
foobar2000 v1.4.2
2019-02-03 17:43:03

File A: H1.wav
SHA1: ef46572d386babadd4778a6c68bd45ae253dc94f
File B: H2.wav
SHA1: 194befda5dfb19e8b0038934bb3a089a489c1e70

Output:
DS : Primary Sound Driver
Crossfading: NO

17:43:03 : Test started.
17:43:44 : 01/01
17:44:01 : 02/02
17:44:20 : 02/03
17:44:31 : 03/04
17:44:49 : 04/05
17:45:08 : 04/06
17:45:27 : 04/07
17:45:55 : 05/08
17:46:11 : 06/09
17:46:22 : 07/10
17:46:38 : 07/11
17:46:53 : 07/12
17:47:13 : 08/13
17:47:26 : 09/14
17:47:40 : 10/15
17:47:52 : 11/16
17:47:52 : Test finished.

 ---------- 
Total: 11/16
Probability that you were guessing: 10.5%

 -- signature -- 
1e52e772f5c3f51a05af4556b5f149b54121fe7c

 

 

Difference between the two files.

 

 

 

 

H1.thumb.jpg.0ba0beb937b14a52aa8fe94565db918b.jpgH2.thumb.jpg.7d9848ba49c75dd043b0c045bcfe8b03.jpg

 

 

And samples here.

 

H1.mp3

 

H2.mp3

 

Let me know if you need the WAV.

Share this post


Link to post
Share on other sites
11 hours ago, STC said:

 However, I find using quick AB method usually would not reveal the difference readily when the change is from long RT to shorter RT. This temporary inability can be explained by how our ears adjust to filter out extra reverbs and takes time to adjust when a sudden reduced reverbs takes place..

 

And I agree entirely. The mind "fills the gaps" - it 'knows what it should sound like', and automatically compensates when something is momentarily missing. And for any length of time, in some situations - pipe organ makers for ages have exploited this human ability; the instrument 'fakes' a very low note being played, by using a combination of pipes to mimic the right acoustic signals - the brain "hears" a lower note happening.

Share this post


Link to post
Share on other sites

Not them. Only you who without even a proper playback hardware which you admitted deficient in every sense could recreate the real sound in head. 

Share this post


Link to post
Share on other sites

And you are caught up in the usual audiophile bind that only "proper playback hardware" could possibly deliver convincing sound - the ego trap. Only a Porsche could deliver Porsche performance type of thing ... well, if someone drove an older model of such they would find them very, very deficient in many areas, by modern standards - but brand charisma rings loud in many people's heads ... :).

 

All audio gear uses standard parts, with various levels of bling added to tease the buyer - the cheap stuff falls down because too many corners are cut - if one knows what corners to address, with a value for money approach, then the Phoenix of competent sound arises, yet again.

 

You see, Virginia, there is no magic in audio ... really ...

Share this post


Link to post
Share on other sites
5 hours ago, fas42 said:

the instrument 'fakes' a very low note being played, by using a combination of pipes to mimic the right acoustic signals - the brain "hears" a lower note happening.

 

Says he, who explicitly refuses to set up a system which could create some sense of the lower frequencies.

(I now see that STC already said similar)

Frank, maybe we start to see your game; you are creating all kinds of excuses to validate your ghetto blaster which isn't even used. You wouldn't dare to because it may invalidate your illusion. You make up more strong excuses first ad infinitum.

Share this post


Link to post
Share on other sites

Actually it could be Frank's "claim" that brings me to my resonse:

 

Just now, PeterSt said:

the brain "hears" a lower note happening.

 

Maybe with headphones. But I wouldn't dig it in-room. This is because the pipe organ (and this is not my own example) would be felt, once it is in the 32Hz region. No brain faking would do that. And Frank wouldn't be able to verify that (delusional him).

 

The reverberation could be working similar, but of coure, you, @STC, hinted at that already ...

 

17 hours ago, STC said:

The difference can be readily audible when heard in the listening room.

 

... although you did not explicitly say that with headphones it would be more difficult.

 

I actually only respond to this because I have a kind of "tuning" at hand which relatively exorbitantly is able to add reverb (hall) to about each song/track. As far as I can tell this is a positive feature but it can sound so overdone that I wonder where it actually comes from. If I could record it (OK, "would want to") then it would be so enormously obvious, you don't want to know. But would that work with headphones too ?

 

What I try to explicitly observe is to what degree the room helps. Is the hall related to my room itself ? I always think it is. But how could that be for real when the "tuning" is the other direction (more deadened) ?

Btw this is also why I personally can't and won't listen with closed eyes. ALL can happen then for illusion. Not so when I see the boundaries of the room and furniture which reflects.

 

I listened briefly to the 2 tracks but have an aversion to the means used. I just hear it. However, my "illusion" here is that I listened to all the example (of progress) videos (recordings) on your site, STC. And now I don't dig it any more.

And in the second I noticed the voice going off (wrongly processed, so to speak) at a first louder note. Didn't look back into the first one, because I know I would hear it there too, now. That's already why A-B does not work for me.

 

I think it is a good subject.

Share this post


Link to post
Share on other sites

My initial response to the thread as far as it is currently, would have been this (and then I got diverted):

 

When this would be about the level (like the 3dB of difference here), contrary to what everybody says, I would not have a problem with that. I may not even hear the difference (which is important, read on). I may need an absolute level in relation to reality (like a piano plays at 90dBSPL) but usually that works out too loud anyway. However:

 

When I set my bass 0.5dB off (and this is about the range up to ~250Hz) with the notice that the construction is such that this does not molest the cross over), all is right away "off" all over. And so it is about consistency. Again, for me. The whole thing being 3dB softer than the other I wouldn't care and probably would not notice (like it is hard to hear it through headphones in this case) but drawing all out of order by changing the bass with 0.5dB and I am out. Well, that is what I expect, generally.

I must add that I am not the only one, because all with the same system are also not able to down or up the bass section (of the speaker involved). And what could be important now is that the rating today is not +/- 3dB (which would be normal and implies a sheer 6dB of difference !) but +/- 0.5dB (1dB of difference). So I suppose that if I would go back to the common sense +/- 3dB, I would get crazy of inconsistency. This is just about how my reference was formed, by now.

 

Similar again could be the hall. I know how my room "halls". It does a little. But if I listen to the music as how it currently is reproduced ? it seems too much. It just can't be. But at least I can check it with some reference (the room itself). Not so with headphones ...

 

I hope I make sense somewhat.

Share this post


Link to post
Share on other sites
44 minutes ago, PeterSt said:

 

Says he, who explicitly refuses to set up a system which could create some sense of the lower frequencies.

(I now see that STC already said similar)

Frank, maybe we start to see your game; you are creating all kinds of excuses to validate your ghetto blaster which isn't even used. You wouldn't dare to because it may invalidate your illusion. You make up more strong excuses first ad infinitum.

 

http://www.mmk.ei.tum.de/fileadmin/w00bqn/www/Personen/Terhardt/ter/top/acbass.html

 

Having heard many systems with impressive subwoofers do a poor job of creating the sense of a pipe organ playing, I prefer my methods.

 

Share this post


Link to post
Share on other sites
8 hours ago, fas42 said:

Only a Porsche could deliver Porsche performance type of thing

 

Did Kingswood also believe this? No wonder they disappeared!

 

4 hours ago, fas42 said:

Having heard many systems with impressive subwoofers do a poor job of creating the sense of a pipe organ playing, I prefer my methods.

 

Don't bluff. Your Sharp cannot even touch 60Hz without loosing few dBs. Stop dreaming about pipe organ accurate reproduction via your speakers.

Share this post


Link to post
Share on other sites
5 hours ago, PeterSt said:

My initial response to the thread as far as it is currently, would have been this (and then I got diverted):

 

When this would be about the level (like the 3dB of difference here), contrary to what everybody says, I would not have a problem with that. I may not even hear the difference (which is important, read on). I may need an absolute level in relation to reality (like a piano plays at 90dBSPL) but usually that works out too loud anyway.

 

I don't think it is possible to level match binaural recordings. 

 

image.thumb.png.b2810f87f3759ebf00b2ddfe6ff08d0a.png

 

This is someone response in ears at different degrees. They are unique to each individual and cannot be reproduced accurately. So level matching is impossible...

 

Quote

 

 

 

Similar again could be the hall. I know how my room "halls". It does a little. But if I listen to the music as how it currently is reproduced ? it seems too much. It just can't be. But at least I can check it with some reference (the room itself). Not so with headphones ...

 

I hope I make sense somewhat.

 

Please don't judge the sound quality. All I wanted to emphasize is ears supress mechanism can make instantaneous blind test can give false negatives at times.

 

BTW, one of the track cannot be distinguished from the original in blind tests. However, if you take your own sweet time and just state your preference without AB comparison you will invariably pick one over the other.

 

Share this post


Link to post
Share on other sites
52 minutes ago, STC said:

Please don't judge the sound quality.

 

I wasn't talking about your two tracks. I was talking about my own means of reproduction lately (through speakers, and I listened to your tracks only through headphones).

 

53 minutes ago, STC said:

All I wanted to emphasize is ears supress mechanism can make instantaneous blind test can give false negatives at times.

 

I understood that. :)  What I tried to contribute is that nothing which is (somehow) fake(d) will adapt. With closed eyes it can, though. So it is all about consistency. If no low frequency is felt, no faked base tone will work out. If no 100% glass surrounds my room, no hard-hall will work out. And:

If this all does not work out, no A-B is possible to begin with. This, while a bit of level difference would not hurt me (says me).

 

I think I must let this go. It is too difficult to explain for me. O.o The subject is still nice.

 

Share this post


Link to post
Share on other sites
39 minutes ago, PeterSt said:

 

I wasn't talking about your two tracks. I was talking about my own means of reproduction lately (through speakers, and I listened to your tracks only through headphones).

 

Ok. Anyway, those were meant to be listened with headphones only.

 

39 minutes ago, PeterSt said:

 

 

I understood that. :)  What I tried to contribute is that nothing which is (somehow) fake(d) will adapt. With closed eyes it can, though. So it is all about consistency. If no low frequency is felt, no faked base tone will work out. If no 100% glass surrounds my room, no hard-hall will work out.

 

I am not very clear about this, but I believe you are trying to explain that we will able to know a fake sound no matter how well it is reproduced and this got something to do with low frequency. On this premise, I will share my thoughts on real and fake sound. To judge sound reproduction, I often use plain vocal. I am not sure how low frequency is relevant here. IMHO, I think we have reached high fidelity to reproduce the vocal as accurate as possible. The emperor penguin experiment showed that a mono sound from a single transducer was real enough to fool the chicks

 

For audiophiles, vocal and piano are usually the ultimate references for accuracy which can be anything. A real piano can sound different depending on where you place the microphone. You may prefer the sound of a Steinway, but in a recording, a Feurich could sound more accurate to your ears depending where the microphones were placed which depends on the speaker. You may also prefer the sound of Steinway with speakers A but probably prefer the sound of the Feurich with speakers B. There are too many variables for us to pass a judgment that a recording to be real enough and another is not. (provide both were made with utmost care as to their fidelity).

 

Firstly, I now believe that we have three different auditory scenes that we are familiar with. The first is the live sound. Natural sound which originates from  a single source. Then we have stereo which is a crude method to reproduce the sound field of the real performance. Lastly, we have headphones. These three are not the same but IMO we have learned to recognize them separately and identify them for what they are. What we are trying to do now is associating real sound within the realm of what you are familiar with. How real or fake they are is how well they fit into your preferred domain.

 

But we have now learned to associate a sound with three different auditory scenes. Real natural sound, stereo and headphones. The difference between fake and real depends on how to sound is perceived. A real sound originates from a single source. Cues to the pinna, the delays of left and right ears plus the reverberation AND your visual cue identify them to exist. In sound reproduction, via loudspeakers and headphones, we do not have a visual cue. Sound exists in space where the brain tells them that it should be there as sound emitting from the phantom center is fake. How are we ever going to overcome the fundamental difference that distinguishes fake and real? I am afraid the fakeness of sound reproduction is just not a matter of how accurate the original sound is reproduced but more importantly how well it stimulates the natural hearing HRTF fully.

 

 

@sementepost on Rumsey's talk made me think that we could be chasing the wrong thing after all. The uncanny valley mentioned there was something that I came across quite often and it could well be true. 

 

 

Share this post


Link to post
Share on other sites

 

12 hours ago, STC said:

 

 

Don't bluff. Your Sharp cannot even touch 60Hz without loosing few dBs. Stop dreaming about pipe organ accurate reproduction via your speakers.

 

Of course it doesn't reproduce low bass frequencies well! But those very low notes are a very, very minor part of what makes pipe organ recitals impressive - it's the rich interplay of the harmonics right through the spectrum, and the space in which the instrument is playing that creates the majesty, the "power" of the sound. Tackling this complexity of sound field is beyond the capability of most rigs - but when the system gets it right, you know it ...

Share this post


Link to post
Share on other sites
8 hours ago, STC said:

Firstly, I now believe that we have three different auditory scenes that we are familiar with. The first is the live sound. Natural sound which originates from  a single source. Then we have stereo which is a crude method to reproduce the sound field of the real performance. Lastly, we have headphones. These three are not the same but IMO we have learned to recognize them separately and identify them for what they are. What we are trying to do now is associating real sound within the realm of what you are familiar with. How real or fake they are is how well they fit into your preferred domain.

 

Conventional quality of stereo is "crude"; optimised stereo SQ presents an experience which matches the perceived qualities of natural sound - to all intents and purposes it's real; because there is nothing audible that hints that it's fake. The listening mind won't be fooled by anything else than a very high standard, which a standard ensemble of components is almost guaranteed not to meet.

 

8 hours ago, STC said:

But we have now learned to associate a sound with three different auditory scenes. Real natural sound, stereo and headphones. The difference between fake and real depends on how to sound is perceived. A real sound originates from a single source. Cues to the pinna, the delays of left and right ears plus the reverberation AND your visual cue identify them to exist. In sound reproduction, via loudspeakers and headphones, we do not have a visual cue. Sound exists in space where the brain tells them that it should be there as sound emitting from the phantom center is fake. How are we ever going to overcome the fundamental difference that distinguishes fake and real? I am afraid the fakeness of sound reproduction is just not a matter of how accurate the original sound is reproduced but more importantly how well it stimulates the natural hearing HRTF fully.

 

Use the force, Luke ... when enough boxes are ticked, the listening mind does the rest, completely automatically. It's a remarkable experience, hearing a setup cross the threshhold; and then, unfortunately, slipping back out of it ...

 

8 hours ago, STC said:

 

 

@sementepost on Rumsey's talk made me think that we could be chasing the wrong thing after all. The uncanny valley mentioned there was something that I came across quite often and it could well be true. 

 

 

 

I haven't watched yet, but I suspect nearly all high end systems lie within the "uncanny valley".

Share this post


Link to post
Share on other sites
52 minutes ago, fas42 said:

But those very low notes are a very, very minor part of what makes pipe organ recitals impressive

 

Can't you get out of here ?

Share this post


Link to post
Share on other sites
9 hours ago, STC said:

I am not sure how low frequency is relevant here.

 

It is not. At least not "generally". However, in say 50% of tracks/songs not high-passed for LP it is in there as part of what happened for real. If only the recording space was not deadened to death. But for example (and IMHO) a church' response is not about hall or reverberation as such but about LF roll which is sub 24 or so Hertz (OK, which *is* reverberation but which I envision can not be mimicked by DSP or something).

Anyway my subject is not about sub-low (although easy to make it that), but I reject that it is sub-relevant like Frank's stance. Actually it is crucial. Regarding your thread, STC, it may be important because the perceived reverb is not really that but is the LF roll which implies it by perception. This "roll" can be seen by observing the real time FFT of music with kick drums (or kettle drum in a church much better) which expresses as a kind of oscillation (back and forth front-backwall bouncing) going from right to left observing the FFT (higher to lower frequency) at fairly equal level until it quickly dies out. Just observe the roll "crawling" like you observe thunder in real life and how your mentioned RT is really in order but multiplied by several lengths of the space/room. Meaning: this goes lower than you'd think.

But use the real m2 of surface including the excursion to imply the pressure level because that can't be faked (in my thinking).

 

9 hours ago, STC said:

These three are not the same but IMO we have learned to recognize them separately and identify them for what they are.

 

Just a quote to emphasize a great post.

Share this post


Link to post
Share on other sites
15 minutes ago, PeterSt said:

 

Anyway my subject is not about sub-low (although easy to make it that), but I reject that it is sub-relevant like Frank's stance. Actually it is crucial.

 

I only got very good, personal feedback on how unimportant it is most of the time, subjectively, some years ago when the previous rig which did a have a proper, separate subwoofer as part of the package, lost its bass completely now and again. Plenty of experimenting going on, and some of the soldering had poor physical integrity - move something the wrong way, and a joint providing functionality let go.

 

Meaning zero subwoofer action ... and I didn't pick it. Sometimes 2 or 3 albums would go by, until something that really, really needed the bass foundation was put on - and then the penny dropped ...

Share this post


Link to post
Share on other sites
12 minutes ago, fas42 said:

Peter, you will have to do better than just trying to bully your way through.

 

Says you eh ?

 

There's this deficiency which is about all what happens to you must be imposed upon the other. My English is too poor to explain it better. However the educated will know what I mean, but some don't need to be an MD to recognize the issue.

One thing: the deficiency I am referring to is about the worst and the even worse part for the environment is that the recipient has grown a skin thicker than an elephant's and with that the recipient can not be reached.

Talking in riddles, I once made the mistake and I hope that the person involved even recognized my rudity at making up. Some things in life can not be dealt with properly. So Frank, I mean no harm. Maybe you can write this down so you can read back on it when needed (which may be every day).

A sincere Cheers and implied apology every day.

Share this post


Link to post
Share on other sites
2 minutes ago, fas42 said:

when the previous rig which did a have a proper, separate subwoofer as part of the package

 

FYI and FWIW: no separate subwoofers with sufficient quality exist IMHO. None. This may be exaggerated, but for goodwill I might claim this.

Now let's talk further from normal context. Not sure how. But there's a difference in rejecting subwoofers on one hand and ignoring the fact that sub-low representation surely can exist on the other.

IOW we agree long gone and you know it. Now accept and move on (elsewhere).

 

Peter out. :)

Share this post


Link to post
Share on other sites
15 minutes ago, PeterSt said:

Now let's talk further from normal context. Not sure how. But there's a difference in rejecting subwoofers on one hand and ignoring the fact that sub-low representation surely can exist on the other.

IOW we agree long gone and you know it. Now accept and move on (elsewhere).

 

Peter out. :)

 

Yes, the lowest frequencies can be accurately reproduced, :).

Share this post


Link to post
Share on other sites
1 hour ago, PeterSt said:

 

It is not. At least not "generally". However, in say 50% of tracks/songs not high-passed for LP it is in there as part of what happened for real. If only the recording space was not deadened to death. But for example (and IMHO) a church' response is not about hall or reverberation as such but about LF roll which is sub 24 or so Hertz (OK, which *is* reverberation but which I envision can not be mimicked by DSP or something).

 

Peter, no offence here but you have to give me some examples. As far as I know, church or hall response is all about the frequency roll-off. The ideal hall response is to have the roll-off equally for all frequencies. Physically this is impossible to achieve and most research accept a good hall's response as the ideal reverberation.

 

If there are no sub 24Hz frequency in the direct sound then you are unlikely to hear the harmonics of them in hall due to masking. I have a real sound samples of a concert hall and will check them. But I have my doubts.

 

A DSP can and cannot reproduce them. It depends on how you are going to use them. I know a so called audiophile recording which supposedly was made in a church contains artificial reverbs. No one could tell them apart.

 

Are we confusing reverbs and their purpose? I think we are probably talking at cross purposes.

 

 

1 hour ago, PeterSt said:

 

Anyway my subject is not about sub-low (although easy to make it that), but I reject that it is sub-relevant like Frank's stance. Actually it is crucial. Regarding your thread, STC, it may be important because the perceived reverb is not really that but is the LF roll which implies it by perception. This "roll" can be seen by observing the real time FFT of music with kick drums (or kettle drum in a church much better) which expresses as a kind of oscillation (back and forth front-backwall bouncing) going from right to left observing the FFT (higher to lower frequency) at fairly equal level until it quickly dies out. Just observe the roll "crawling" like you observe thunder in real life and how your mentioned RT is really in order but multiplied by several lengths of the space/room. Meaning: this goes lower than you'd think.

But use the real m2 of surface including the excursion to imply the pressure level because that can't be faked (in my thinking).

 

 

Just a quote to emphasize a great post.

 

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

×