AVS Forum banner

721 - 733 of 733 Posts

·
Registered
Joined
·
15,929 Posts
Discussion Starter #721 (Edited)
I personally think people, like that guy, who "ride the gain" and peg the volume to max or to the point that other higher level passages in the song would be clipping the amplifier are effectively cheating because the people who claim to hear differences between Hi-res vs. CD resolution:

- never describe this step being necessary to hear the differences they claim to hear in imaging, sound stage, warmth, details, etc.

- it can potentially fry the tweeters/headphones or the listener's hearing

- it is an artificial, contrived scenario which does not reflect real-world listening conditions

Many people don't seem to get that we aren't trying to determine if there is "a difference" between A and B; we know that there is because our analyzers can see it, like the peaks in the two images we see at the Dropbox download site of my two test files, A and B, I turned into an animated GIF:

Click image to expand.

What we are trying to determine is if there is an audible difference between A and B like the sighted listeners claim.


Yet another way to "cheat" would be to use an EQ to peg a single frequency to max, or say just "all the treble" if all one had were bass and treble knobs, and then suppress all other frequencies: "Narrow band listening". Again, this is an artificial, contrived scenario and it both never occurs in real-world listening nor is described as being used by the people who claim to hear differences under sighted conditions.
 
  • Like
Reactions: Soulburner

·
Registered
Joined
·
299 Posts
If you go here and fill out a form, just a name and an email, he'll give you access to the files:
He did (as I've underlined) mention using an analyzer though to find what to listen to:

"choose the one with the highest peak level which is given on JRiver DSP Studio Analyzer shown during playback.

... and analyzers are expressly forbidden. . . . Maybe "cheating" isn't quite the right word but he did use external gear. I'm not making a call on him but I'm not confident he's the guy (but I'm not sure) I'm trying to learn more about. I remember Waldrep saying some listener described a procedure which made him decide to drop him out of the competition. What did that guy do?
My post discussed both the guy who cheated and the guy who used his DAW only. The first bit was about the guy who cheated by using an analyzer:
"Comparison of A/B several times, taking a specific part with higher dynamics (15-20 seconds) and choose the one with the highest peak level which is given on JRiver DSP Studio Analyzer shown during playback. I used realtime switching between tracks comparing the peak level of 15-20 sec parts with higher dynamics. In fact, the determination of HR version by listening parts of the track and comparing A/B was confirmed or decided by the highest peak level if i could not hear any difference."
You asked if I was talking about the guy using his DAW, if that was the guy who cheated, it wasn't, it was the guy quoted above. In the post, I then started talking about the guy with his DAW, sorry about the confusion.
 

·
Registered
Joined
·
299 Posts
I personally think people, like that guy, who "ride the gain" and peg the volume to max or to the point that other higher level passages in the song would be clipping the amplifier are effectively cheating because the people who claim to hear differences between Hi-res vs. CD resolution:

- never describe this step being necessary to hear the differences they claim to hear in imaging, sound stage, warmth, details, etc.

- it can potentially fry the tweeters/headphones or the listener's hearing

- it is an artificial, contrived scenario which does not reflect real-world listening conditions

Many people don't seem to get that we aren't trying to determine if there is "a difference" between A and B; we know that there is because our analyzers can see it, like the peaks in the two images we see at the Dropbox download site of my two test files, A and B, I turned into an animated GIF:

Click image to expand.

What we are trying to determine is if there is an audible difference between A and B like the sighted listeners claim.


Yet another way to "cheat" would be to use an EQ to peg a single frequency to max, or say just "all the treble" if all one had were bass and treble knobs, and then suppress all other frequencies: "Narrow band listening". Again, this is an artificial, contrived scenario and it both never occurs in real-world listening nor is described as being used by the people who claim to hear differences under sighted conditions.
Yeah, I agree, that's what I meant saying "you're still led to the conclusion that the differences are essentially inaudible." It's because these contrived situations are so far removed from how you would ever listen to anything. It would be like taking an electron micrograph of a knife and claiming it can't be sharp, look how crappy that edge looks.

Brand new filet knife
3036110


I'd bet anyone claiming that knife isn't sharp would be as reluctant to slice their arm to test its sharpness as the 'hear drastic differences' are reluctant to do ABX tests.
 

·
Registered
Joined
·
15,929 Posts
Discussion Starter #724
Completely off topic, but I recently learned that the reason your disposable razor (shaving) blades go "dull" isn't because the sharp edge get rounded, instead of being a sharp "V", but rather it's because of microscopic chipping:
"
“Our initial thought was that this was a wear problem, that material was being removed from the razor,” Tasan says. “We were expecting to see that over time the blade gets rounder and rounder. We didn’t see it.”

Instead, he continues, “we saw fracturing and chipping of the blade that is forming this C-shaped crack.”

 

·
Registered
Joined
·
15,929 Posts
Discussion Starter #725 (Edited)
Good video of it happening here:
 

·
Registered
Joined
·
299 Posts
HA! Little you know, they go dull because you're not putting them under a pyramid, everybody knows that. My listening room is pyramid shaped,. that's why my speakers never sound dull.
 

·
Registered
Joined
·
299 Posts
Also is this challenge still up? Can I still download the files to compare? I couldn't find them.
I got stuck in an endess loop of "click here" which would just take me back to where I began.
FYI I requested the files same time I replied with corrected link and about an hour ago got his reply with the links to download, he also said there is still time to get into the survey but not much.
 

·
Registered
Joined
·
15,929 Posts
Discussion Starter #728 (Edited)
Debatable topics like these, "Do modern, competent DACs sound different?"/ "Does Hi-Res sound better than CD?", are difficult for even the experts to test because they simultaneously need, ideally, to either be an expert in, or have an expert consultation and peer review from several different fields:

  • high end audio gear expert (and they better have some understanding of ultrasonic reproduction for the playback chain setup for Hi-Res)
  • software expert to prepare the files so they are precisely time aligned and precisely level matched, and in case of the Hi-Res vs. CD you need to use a state of the art sampling rate convertor that introduces no aliasing noise nor latency
  • pyschoacoustics expert to determine what music material is suitable and likely to make a discernible difference within human hearing threshold limits
  • the recording engineers who made the recording because only they will likely know what the frequency response of the mics used were, for example
  • expert listeners with tested hearing
 
  • Like
Reactions: mechtheist

·
Registered
Joined
·
15,929 Posts
Discussion Starter #729 (Edited)
So this thread's ABX test was to determine if anyone could identify any audible deficiencies with a dirt cheap ($7.99 shipped) DAC and so far of all the people who've participated, none could detect any. This suggests that under these conditions the DAC is indeed "perfect" or "audibly transparent"; i.e., despite being extremely inexpensive it offers state of the art performance, at least audibly and within its limited feature set and flimsy construction. [I've never had one fail but reviews on the web say if you yank it out by the wire from your iphone too many times the wire can break.]

It is important to note however that this ABX test also establishes that the other devices in the chain used to create File B also inherently must be transparent, namely the $1 Monoprice 6ft. RCA cable and the free USB cable, or else people would have detected the flaws in them during the test, yet nobody did.

In another thread I described how by inserting an amplifier into the existing chain, namely my inexpensive Yamaha AVR, we can test if it too is transparent or not. It is driven under load (i.e., connected to either speakers or professional grade, 4-ohm test resistors during the creation of File B, which is how audio magazines do measurements of amp performance so they'll have standardized, consistent results from review to review). The speaker level output at the speaker input posts is tapped, and reduced down to line level, so that it can be re-digitized by my ADC and then recorded, via an "LOC", line out convertor.

Here is the signal flow diagram showing how Files A and B would be created for my "Perfect Amp Test". It is just like the Perfect DAC test which so far has found no listener who can discern any audible change to the signal passing through it. At the location marked with green text my cheap Yamaha AVR would be inserted with an LOC (line out convertor) attached to its speaker wire outs to attenuate the speaker level signal down to a voltage my digital recording gear can accommodate. I can either attach 8-ohm speakers, 4-ohm speakers, or professional grade, high power 4-ohm resistors* as the amp's test load, depending on what the listener wants, however the recording is done electrically, not acoustically. In a real life, in-person test I could instead attach the listener's actual speakers to my amp for its load, used for the test file creation stage, but obviously that can't be done over the web.
View attachment 3038891
Click to expand image.


*This is what professional magazine reviews use when measuring an amp's performance so they have uniform, standard results and can compare it to other amps they've tested.
Here is the ORIGINAL Perfect DAC test explained:


So the "Perfect Amp Test" would be the same but with the following additional steps in green text:

File A is: D --------------------------------------------------------------------->DAC in listener's system.

File B is: D -->inexpensive DAC-->Cheap Yamaha AVR (under load)--->LOC (line out convertor)--->RCA wire--->ADC --->USB--->laptop HDD--->DAC in listener's system.

Again, if Files A and B are found to be indistinguishable from each other in a listening test then all those added steps to create File B are perfect to the ear, i.e., they successfully resolve every single detail/nuance in the music and add no discernible distortion/noise/alteration. The amp would be shown to be audibly "perfect to the ear" or "transparent" under those conditions.

LOCs, line out convertors, are sometimes also called "voltage dividers". They convert a power amp's output signal, measured in watts. down to a smaller level that can be used for line level applications, including tape/digital recording. They aren't perfect and add distortions but I'm betting not enough to be audible.

It is important to note that although File B goes through all sorts of devices (which potentially can audibly degrade the signal) File A, the reference file, goes through none and will sound exactly as it did on the original CD. It I'm wrong and any or all of the added events in File B's creation add some audible noise, distortion, hum, etc. or fail to fully resolve the subtle nuances and details found in File A, then that should make the listener's task of trying to hear some difference between File A and File B even easier.

For simplicity I may switch to a different DAC for the "Perfect Amp Test" because this $7.99 Apple dongle DAC itself is no longer what is actually being tested and I find this little DAC, and iTunes in particular, rather cumbersome to work with.
The green text shows the insertion location for the DUT (device under test) in the signal path which is being tested for transparency, used to create the test File B:
3040061
 
  • Like
Reactions: mechtheist

·
Registered
Joined
·
15,929 Posts
Discussion Starter #730 (Edited)
Here's the newest least expensive USB DAC I've seen on Amazon with free shipping ($25 minimum purchase required). It's "Hi-Res"!

[She's wearing my Sony headphones.]

The DAC is less than a fancy coffee at Starbucks! :)
Example:
Pumpkin Spice Latte (Limited Time)Venti$5.25
 

·
Registered
Joined
·
15,929 Posts
Discussion Starter #732 (Edited)
Yes. That's the beautiful thing. All sorts of electrical things can be tested this way.

[I haven't owned a minidisc machine in decades and a VCR would have to be dug out of deep storage though.]
 

·
Registered
Joined
·
15,929 Posts
Discussion Starter #733
MiniDisc {ATRAC] usually sounded pretty good but it would throw a hissy fit trying to encode some material including this song I learned about from David Ranada:

He calls it a "codec killer". Its heavy stereo spread with the guitar hard left is troublesome for lower bit rate codecs.

I can hear the problem in this video too.
 
721 - 733 of 733 Posts
Top