GoldenSounds passes apparently ABX test for DACs (NOT Really)

[Geert]

So I'm not going to sit through that video, but does Sharur claim to hear 20khz+ as well?

He claims to hear an image shift. Maybe that shift is the result from a difference in sensitivity between the 2 ears for very high frequencies. It would mean only people with good 20kHz sensitivity left in 1 ear can pass the test.

 

[Geert]

It sounds to me , you wanted him to fail the test, by altering the test criteria.

I totally don't, that's your interpretation. And to be clear, when I say 'a technique is used to identify differences', that's not suggesting he's cheating.

There are so many instances on ASR, referring to Blind tests, and it being the only reliable way to determine audiblity. Now that it has been done with a result we didn't expect (or like), . . .

I'm a very rational person, so I don't have a problem with any outcome of a test. But I do have a problem with jumping to conclusions and spinning a narrative.

 

[goat76]

In bold is exactly the point I tried to make. Conclusions made from using special techniques to identify differences are irrelevant to the normal listener.

That will probably differ depending on the audio file used for the listening test, what it contains, and what the particular listener focuses on when trying the identify possible differences. A short time window may be less important and the differences may be more pronounced for another listener, depending on what he focuses on and what music file is used for the listening test.

Again, the question here is not if the differences are relevant for the normal listener or not, that is another question, the main question is if human hearing can detect differences in sound, and for that to be proven, it's enough even if there are just a small portion of the people in the world who can hear it.

I did the ABX test twice and couldn't hear the difference between the sound files GoldenSound used for his test, but that doesn't mean he can't so I chose to believe him until someone proves he cheated on his test. However, I realized I must take the test again as Foobar2000 was not set to ASIO output, so my system was resampling both files to 96 kHz.

 

[MaxwellsEq]

This is why I said, what do you expect.
It sounds to me , you wanted him to fail the test, by altering the test criteria.
There are so many instances on ASR, referring to Blind tests, and it being the only reliable way to determine audiblity. Now that it has been done with a result we didn't expect (or like), . . .

I know what you mean, but catching up on this thread after a couple of days away I don't pick up on an "ASR groupthink". I get the sense that people are trying to understand. I get a sense that most people seem interested in finding out what enables an outlier to be spotted like this and to determine whether a generalized rule can be captured.

There's a bit of back and forth about "tells" and whether that's an appropriate methodology. Personally I think it is. Because if there is a tell, we will understand more about detecting differences.

 

[VientoB]

The result is also going to differ depending on what headphones or speakers are used, as different frequency responses will change the effect.

 

[Tell]

This is what he has on repeat.
View attachment 367685

And he identifies the difference each time within 2 seconds, that's why I say he's listening for kind of a tell. In the video he explains the difference at that point as a shift in imaging (sound more drawn to the left if I recall correctly).

I would not trust that guy AT ALL, his videos in the past have been so bad, stupid and weird that I wouldn't be surprised in the slightest if he cheated this somehow.

 

[VientoB]

I would not trust that guy AT ALL, his videos in the past have been so bad, stupid and weird that I wouldn't be surprised in the slightest if he cheated this somehow.

What have been bad about his videos in the past? This video was released on the Headphone Show channel, and most of those guys are sceptical of differences between DACs and amps.

 

[Tell]

What have been bad about his videos in the past? This video was released on the Headphone Show channel, and most of those guys are sceptical of differences between DACs and amps.

I was talking about Sharurs video that was screenshotted in the post I commented on.

 

[oleg87]

What have been bad about his videos in the past? This video was released on the Headphone Show channel, and most of those guys are sceptical of differences between DACs and amps.

Sharur is a well known troll/attention seeker.

 

[VientoB]

Sharur is a well known troll/attention seeker.

Ah I thought you were talking about GoldenSound.

 

[Geert]

the main question is if human hearing can detect differences in sound, and for that to be proven, it's enough even if there are just a small portion of the people in the world who can hear it.

Of course, but if you follow some subjective forums then you know that results from tests like this are often being generalised. The title of the GS video is all they need.

I did the ABX test twice and couldn't hear the difference between the sound files GoldenSound used for his test, but that doesn't mean he can't so I chose to believe him until someone proves he cheated on his test.

Since I've given a reasonable hypothesis of how he could pass the test I'm certainly not the one suspecting him of cheating:

He claims to hear an image shift. Maybe that shift is the result from a difference in sensitivity between the 2 ears for very high frequencies. It would mean only people with good 20kHz sensitivity left in 1 ear can pass the test.

 

[markus]

The test stimulus was a shaped 'chirp' at the frequency of interest. I imagine that it would be experienced as more of a 'click' with a certain 'tonal balance' than anything else.

Not a click. Test stimulus was a pulsed tone (sinewave). Sounds like this:

1000Hz-pulse-tone.wav

drive.google.com

Makes it easier to detect than a steady-state sinewave.

 

[markus]

I don't believe we can conclude that from the study, because of: "Note that thresholds for all age groups plateau at 105 dB SPL, essentially because of a ceiling effect imposed by the maximum output of the hardware. This is considered “no response” as it represents the output limit of our equipment and does not represent true thresholds".

In the table on page 9 you see that in the 95% of people in each age group there are people who score that fake threshold of 105dB at high frequencies, actually meaning they couldn't here the test signal with the limits imposed by the equipment.

Yes, the graph gets misinterpreted by people here. Around 30+ you usually don't hear 20kHz anymore, regardless of sound pressure level. Receptors are gone.

 

[ads_cft222]

This is not over . A test comparing a topping and holo will be posted later . He wanted to post this video first

 

[solderdude]

"stereo difference" (Stereo difference half speed.zip) is available for download

Click to download the transfer "stereo difference" (1 files, 118 MB) with TransferNow

www.transfernow.net

Here is the stereo difference file and half speed stereo.

Listening to both files with the volume turned up louder than I would really ever want to be listening to I heard .. nothing... not surprisingly.
In order to get the difference file to reach near 0dBFS I had to increase the gain by 40dB, turn the volume up to very loud levels (when listening to music).
This is zoomed in a lot (in time scale) and was the biggest peak I could find. Normalized it to -6dB (to prevent clipping)

Also not with the half speed version at already 'loud' levels (if music were playing).
I had to increase gain 39dB to finally hear some ticks (at 10kHz) in the half speed version.

Needless to say that the differences heard in this file cannot possibly explain the differences audiophiles seem to hear between say a Chord DAC and any 'regular' DAC.
Question remains how Cameron got 18/20...

Also the conclusion that well measuring DACs can sound different enough to reach audible levels is really far fetched and might just be the transducers acting weird.
I wonder if Cameron could pull that off with another good and well extended headphone.

 

[Geert]

I wonder if Cameron could pull that off with another good and well extended headphone.

And another DAC. We're now looking at the source files if 'm not mistaken. It would be nice to see a recording of the output of the DAC to see if it introduced any artifacts. But I guess GS is happy with the results as they are.

 

[Ken Tajalli]

I know what you mean, but catching up on this thread after a couple of days away I don't pick up on an "ASR groupthink". I get the sense that people are trying to understand. I get a sense that most people seem interested in finding out what enables an outlier to be spotted like this and to determine whether a generalized rule can be captured.

There's a bit of back and forth about "tells" and whether that's an appropriate methodology. Personally I think it is. Because if there is a tell, we will understand more about detecting differences.

OK. I accept that there are "tells" , perhaps plural.
A Tell states, that there are differences. How else can one Tell them apart.
However, Cameron (thanx @solderdude ) muddied the water by stating that in a quick fire, he can detect supersonics. And now, some are not even looking for other tells. "It must be that then!"
But it is illogical to accept that. He listened to music, not test tones at full volume.
Let's look past that supersonic anomaly. not much else has been suggested as plausible.

 

[Geert]

How else can one Tell them apart.

Audiophiles wife's tend to hear it from the kitchen with their head under the extractor hood with a song they last heard 2 weeks ago. No tells in repeat mode needed. They don't get the credits they deserve.

 

[solderdude]

And another DAC. We're now looking at the source files if 'm not mistaken. It would be nice to see a recording of the output of the DAC to see if it introduced any artifacts. But I guess GS is happy with the results as they are.

Looking at the DAC I don't expect any weirdness at 176kHz sample rate. I would suspect the Susvara (in combination with his ears) a bit more than some minutia coming from an amplifier or DAC or re-sampler. I did not see any weirdness from 15kHz up to 30kHz in my Susvara measurement but that's just one sample and simple measurements.
Transducers usually do have the biggest impact to sound quality and distortion.
Note that both Sharur and Cameron had to find a 'tell' in the song and needed to concentrate on that in order to tell the files apart.
I am sure when just listening to file A and then listening to file B and would have say which was what they might not be able do any better than chance.
It is not that one of them sounded 'obviously' different.

 

[Ken Tajalli]

Needless to say that the differences heard in this file cannot possibly explain the differences audiophiles seem to hear between say a Chord DAC and any 'regular' DAC.
Question remains how Cameron got 18/20...
Also the conclusion that well measuring DACs can sound different enough to reach audible levels is really far fetched and might just be the transducers acting weird.
I wonder if Cameron could pull that off with another good and well extended headphone.

I think you got it there! let's not complicate it.
If he can do the same with say a Chord TT2 and the top Topping DAC.
- The Chord DAC does have a similar filter to what he used here.
- The Topping has a similar Filter to the other filter he used here.
For as long as it is blind and level matched (no louder than -1-2dB). We also need something to nulify the TT2 delay. We'll give him a choice of headphones that both devices can drive happily, to choose from.

He's got the right ABX gadget too, he can use it this time.