• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Archimago has a new DAC listening test

Well Cameron (goldensound) and Sharur (did I say that right?)
invalidated that statement.
Exactly. I wonder how many people has reached to the same lazy conclusion you did from that "demonstration". I don't know what their intent was. If it was to create more confusion and allow people to sell DACs with ridiculous claims - that's mission accomplished.
 
what Cameron and Sharur showed is that they can successfully differentiate between 2 different filters
And even that is highly uncertain.
 
Well well well
The usual suspects ! :)
Full Audible transparency is contradiction in terms, as the Audible assumes individual hearing abilities.
FWIW, and for those who just take one sentence out of a whole thread to jump on:
I also said, small variations are irrelevant. Even to the likes of Cameron.
As it happens, in Archie's experiment, there are two recon filters . the Apple one and the Linn one. and audible differences (if any) seem to be coming from those.
Better wait for full report from Archie.
Talk about lazy assumptions!
We weren't just talking about DACs. @board mentioned all components specially DACs.
Bottom line is, there are always tiny differences, whether they are audible depends on listeners and our preconceptions of what constitutes audible.
For most parts it works and won't matter.
Just a few weeks ago, a tiny +19/20 khz cut was considered inaudible.
Don't look for absolutes, practically transparent, I won't argue with, but Full?
 
Full Audible transparency is contradiction in terms, as the Audible assumes individual hearing abilities.
I know that I'm also just zooming in on one sentence now, but it does after all pretty much sum up your post:
Yes, there are individual hearing abilities, and it's also possible to train your hearing. I think most people here acknowledge that, and if they don't, they should. I'm able to pass ABX tests that my mum can't. You can perhaps pass ABX tests that I can't.

However, as far as I know "full audible transparency" does exist for certain things, since there are things that no one in the entire world would be able to hear, such as for example a nasty artifact in a DAC at -140 dB while music is playing with peaks close to 0 dB, even if the music is really loud.
Where the line is between completely inaudible for everyone, inaudible for most people all of the time, and inaudible for most people most of the time is of course a bit blurry, I acknowledge that.
Just to clarify what I said in an earlier post: DACs with distortion at e.g. -120 dB will be more transparent than an amp with distortion at -60 dB. I was talking about threshold of audibility, so in decibles, and I didn't mean that certain types of products (e.g. DACs or amps) will always be fully audibly transparent. Since it's more likely that an amp changes the frequency response than a DAC does, amps as a category will be less audibly transparent than DACs.
Did you by any chance listen to the two files that someone posted, run through two amplifiers - one with more distortion than the other, and if so could you hear a difference?


At xiph.org there's a file which has a 1 kHz tone at -105 dB. When I listened to this I had to turn the volume on my amp almost all the way to the max before I could hear anything, and all the way to the max to hear it at even quite a low volume.
If I had played music at that volume, my speakers would quite surely have blown up!
I can send you a link to this if you like.
I could not imagine that you, Ken, would in any way be able to hear that tone, or a nasty constant buzzing artifact, if it was inserted at that level under regular music with peaks close to 0 dB.
I've already tried something like this myself, taking a really good recording of Mozart's "Lacrimosa" and mixing it with Dark Funeral, which is fast, pounding black metal, at -80 dB, and I couldn't hear the black metal under the Mozart. If you like, I can send you this.

None of this is meant as a personal attack on you, but unfortunately I find that the people who talk the most about individual hearing abitlities and how they've spent years training their hearing to perfection (and maybe they truly have), and often also say that they don't care how something measures, they only care about how something sounds, are often the ones who are the least willing to actually listen, and only listen, and thereby back up their claims with an ABX test - instead prefering to go on and on and on, arguing about how they don't need to prove anything to anyone because they know what they heard, and it was so goddamn obvious - so of course they should be able to ace the ABX test in a few minutes. Yet, they don't.
 
Last edited:
I've already tried something like this myself, taking a really good recording of Mozart's "Lacrimosa" and mixing it with Dark Funeral, which is fast, pounding black metal, at -80 dB, and I couldn't hear the black metal under the Mozart.
Bill Waslo had a Sousa march buried 65 dB below Brahms's Lullaby. They were posted for years and no-one ever did an ABX that showed successful detection.

Here's the page with his files, but I can't guarantee they're still hosted...
 
Bill Waslo had a Sousa march buried 65 dB below Brahms's Lullaby. They were posted for years and no-one ever did an ABX that showed successful detection.

Here's the page with his files, but I can't guarantee they're still hosted...
The files are still there. Audio DiffMaker software is needed to play them. It works.
 
The files are still there. Audio DiffMaker software is needed to play them. It works.
The Sousa march is still there, if we make a diff file. It only shows how pointless are all those -120dB SINAD wars here, if -65dB uncorrelated signal is completely inaudible.
 
Not entirely pointless as they explicitly reveal good engineering.
Keith
 
I've already tried something like this myself, taking a really good recording of Mozart's "Lacrimosa" and mixing it with Dark Funeral, which is fast, pounding black metal, at -80 dB, and I couldn't hear the black metal under the Mozart. If you like, I can send you this.
Ethan Winer has something similar: http://ethanwiner.com/aes/
Presented at Audio Myths Workshop, 32:09:
 
Bill Waslo had a Sousa march buried 65 dB below Brahms's Lullaby. They were posted for years and no-one ever did an ABX that showed successful detection.

Here's the page with his files, but I can't guarantee they're still hosted...
I haven't tried that yet, as I need to install Audio DiffMaker first, but from my own experiment I saw that it depends very much on what type of music you mix and the volume level at which you mix it. I took some of the most brutal music I had and mixed it with some of the lightest and quietest/most dynamic music I had, and I could hear the Dark Funeral song underneath the Mozart if the volume level of Dark Funeral was loud enough.
If instead of Mozart I had used The Beatles or Rolling Stones, I could probably get away with having Dark Funeral at a much higher volume level.
 
Last edited:
I know that I'm also just zooming in on one sentence now, but it does after all pretty much sum up your post:
Yes, there are individual hearing abilities, and it's also possible to train your hearing. I think most people here acknowledge that, and if they don't, they should. I'm able to pass ABX tests that my mum can't. You can perhaps pass ABX tests that I can't.

However, as far as I know "full audible transparency" does exist for certain things, since there are things that no one in the entire world would be able to hear, such as for example a nasty artifact in a DAC at -140 dB while music is playing with peaks close to 0 dB, even if the music is really loud.
Where the line is between completely inaudible for everyone, inaudible for most people all of the time, and inaudible for most people most of the time is of course a bit blurry, I acknowledge that.
Just to clarify what I said in an earlier post: DACs with distortion at e.g. -120 dB will be more transparent than an amp with distortion at -60 dB. I was talking about threshold of audibility, so in decibles, and I didn't mean that certain types of products (e.g. DACs or amps) will always be fully audibly transparent. Since it's more likely that an amp changes the frequency response than a DAC does, amps as a category will be less audibly transparent than DACs.
Did you by any chance listen to the two files that someone posted, run through two amplifiers - one with more distortion than the other, and if so could you hear a difference?


At xiph.org there's a file which has a 1 kHz tone at -105 dB. When I listened to this I had to turn the volume on my amp almost all the way to the max before I could hear anything, and all the way to the max to hear it at even quite a low volume.
If I had played music at that volume, my speakers would quite surely have blown up!
I can send you a link to this if you like.
I could not imagine that you, Ken, would in any way be able to hear that tone, or a nasty constant buzzing artifact, if it was inserted at that level under regular music with peaks close to 0 dB.
I've already tried something like this myself, taking a really good recording of Mozart's "Lacrimosa" and mixing it with Dark Funeral, which is fast, pounding black metal, at -80 dB, and I couldn't hear the black metal under the Mozart. If you like, I can send you this.

None of this is meant as a personal attack on you, but unfortunately I find that the people who talk the most about individual hearing abitlities and how they've spent years training their hearing to perfection (and maybe they truly have), and often also say that they don't care how something measures, they only care about how something sounds, are often the ones who are the least willing to actually listen, and only listen, and thereby back up their claims with an ABX test - instead prefering to go on and on and on, arguing about how they don't need to prove anything to anyone because they know what they heard, and it was so goddamn obvious - so of course they should be able to ace the ABX test in a few minutes. Yet, they don't.
Well bless you and your civil, polite tone. I do appreciate that. Heaven knows, I have been the other way at times.
Regarding the word Full , that is my only objection, as I don't believe in absolutes, otherwise it is all ... partridge on a pear tree ... to me .
My hearing is not perfect, and I am fully versed on hearing limitations.
The point I was trying to make (unsuccessfully) was:
Don't keep stressing on full transparency! what we believed to be transparent, irrelevant, inaudible ..... was proven to be otherwise.
If one person on planet can hear it, then the notion is Practically transparent, which I always believed in.
 
what we believed to be transparent, irrelevant, inaudible ..... was proven to be otherwise.
That's the claim, anyway. "Proven" is a leap too far.
 
That's the claim, anyway. "Proven" is a leap too far.
Tell me, if the results agreed with your expectations, would you have tried to pick holes at the experiment? suggesting it could be gamed? trying to do it again with witmesses, under strictest of conditions you could imagine?
Or, would you have accepted it as proof, that you were right and quoted it at every opportunity?
I wonder!
For me it was proof!
that one in 10000, could detect such discrepancies.
Would it matter? does it change my opinion on hifi equipments, playing real music?
not in the slightest.
It was an interesting experiment, on wonders of audiblity.
 
Tell me, if the results agreed with your expectations, would you have tried to pick holes at the experiment? suggesting it could be gamed? trying to do it again with witmesses, under strictest of conditions you could imagine?
Or, would you have accepted it as proof, that you were right and quoted it at every opportunity?
I wonder!
For me it was proof!
that one in 10000, could detect such discrepancies.
Would it matter? does it change my opinion on hifi equipments, playing real music?
not in the slightest.
It was an interesting experiment, on wonders of audiblity.
Extraordinary claims demand extraordinary proof.
 
I haven't tried that yet, as I need to install Audio DiffMaker first
I can easily post the DYF files converted to standard wav. Will you make an ABX then and post the result? Or is it just another way of excuses?
 
I can easily post the DYF files converted to standard wav. Will you make an ABX then and post the result? Or is it just another way of excuses?
???
What makes you think I'm making excuses?
 
???
What makes you think I'm making excuses?
OK, you do not. So here are the files attached (it was easier to extract them than I expected). Looking forward your ABX result. Of course anyone else would be welcome as well. Do not hesitate to use Paul's @pkane Deltawave to check there is really the Sousa march background in one of the files.
 

Attachments

  • Sousamarch_test.zip
    2.5 MB · Views: 32
Tell me, if the results agreed with your expectations, would you have tried to pick holes at the experiment? suggesting it could be gamed? trying to do it again with witmesses, under strictest of conditions you could imagine?
Or, would you have accepted it as proof, that you were right and quoted it at every opportunity?
I wonder!
For me it was proof!
that one in 10000, could detect such discrepancies.
Would it matter? does it change my opinion on hifi equipments, playing real music?
not in the slightest.
It was an interesting experiment, on wonders of audiblity.
Thanks for your compliment :).
I always try to keep a civil tone, but unfortunately I find many people in the audio world too often go into insults or screaming matches, and too often because they make assumptions about what the opposing side actually believes instead of just asking them.

Anyway, I understand your rationale in the quoted post, and it really is the right scientific mindset to avoid words like "always" and "never", since they can hardly ever be used.
However, I have to ask:
If it was possible to create an amplifier where the noise floor or distortion components were -200 dB below full scale (or even -180 dB or -170 dB), do you believe that there might be at least one person on planet Earth who might be able to hear that?
If the answer is yes, I think this is where you differ from most people on this forum.
Although it's best to always keep an open mind to any and all possibilites, many experiments have been conducted through the years, and as we've been talking about in this thread, often the threshold for audibility is less than -100 dB below full scale.
And I suppose this is also what you're saying - that for most people most of the time in most situations they can't hear artifacts at -80 dB or whatever, but it is possible that there is someone out there who is able to do it.
 
Back
Top Bottom