[U-3C]

Hey guys! I hope it's okay if I post this here.

I plan to set up a small experiment where I will play the same music file to different people, but claim that one is a heavily compressed file, while the other one is one of those ridiculous 300 mb highest quality audio files that cost a kidney to purchase legally (at least for a broke college student like me. Thanks, Head-Fi, for helping my wallet lose so much weight. ;v.

It's just a fun experiment, so nothing too serious, and I hope all the volunteers can get a laugh out of it. However, I do wish to seek some help on how I should set this test up.

Now, most of these volunteers are just friends who purchase headphones at local stores for the sake of listening to music on the go and do not really care about audio, but there are those who claim to be audiophiles. I had tbe explain to most of them what compression means, as well as promise a few I won't blow their headphones. There are some who claims to be audiophiles though, so I'm interested in putting them to the test.

o(0v0)b

I basically told them that I will play files of different compression rates, and I will try my best to bring up their expectations for the "high quality" music, vs the "lousy" 128 kbps mp3 file. I will be using a centrance dacport slim and a pair of q701 headphones, but I will also encourage the volunteers to bring their own gear. I will lie to them that the goal of the test is to see at what point the difference in compression stops becoming noticeable.

For half of the volunteers, mostly those that do not know about audio gear but uses gear from Bose, Beats, and the like, I will claim that my lime green q701 is just a cheap pair of "gamerz" headphone that I picked up on ebay for 30 bucks, and I will attach a fairly dirty cable. I will tell them that the USB dac/amp combo is just a cheap sound card, as my laptop is so old, the audio jack is broken, so the only way to even play audio is to use this ghetto external card. I will encourage them to use their own "higher quality" gear, hoping it will affect their subjective judgements.

For the remaining volunteers who might recognize my setup, I'll just show the Amazon page for the equipment I'm using (waaaaayyyy more expensive than the actual price, but maybe seeing the price will affect their judgements? >_<).

I'll also talk to them in advance, talking about the possible claims that people made regarding higher quality audio (4 bajillion kHz sampling rate at a bit depth of 512 bits and so on), but with an affirmative tone, hopefully so that they will genuinely think that there is a difference, even though I'm clicking the same file. I'll also be playing pop music, so even if I was indeed playing different files, it will be a lot harder to notice than say jazz or classical orchestra.

If I have time, I'll also tell them at the end the truth behind the test, maybe have a laugh, and seriously ask them to listen to highly compressed vs the lossless versions of well mastered songs and see at what point they can actually hear a difference.

...

Only I'll once again play the exact same song back and forth! \(;v/.

This time, I'll try to find music that will shine with the q701 a few times and hopefully wow a person or two, and guide their thinking, hopefully to give them the impression that they really are listening to high quality versions of songs, and then repeat, but claiming that the music is compressed.

I'm interested in whether or not I can get people to truly believe that there is a difference. I'm also interested in whether or not I can notice a difference between people who think that they are listening to a "hi-fi" setup vs those who believe I'm testing them on a cheap platform that can't even play music properly.

This is where I would like to ask for some help. Do you guys have any advice or suggestions on what I should say to get them to believe me?

I won't be doing a blind test as the point is to plant the concept of a false difference in listening experience, and if a volunteer notices it and suggests that I do a blind test...I risk revealing the truth. I need to make sure to distract them from it, especially since I know one volunteer who is extremely skeptical, just as or even more knowledgeable than me regarding audio equipment and is interested in seeing the test results. He already pointed out a few "flaws" when I asked him if he is interested in participating (hopefully he isn't reading this very post!!!)

Next, I really wish you guys can help me think of ways to apply the placebo affect as much as possible to the volunteers. I don't have too much time with each volunteer (around 5 to 10 minutes each), so I need to effectively play with their expectations between changing songs (if they plan to stay and listen to my headphones for 20 min, I can't stop them ), but I can't sound like a jerk. Any suggestions are welcome, as long as they are not offensive (remember that these people are all my friends).

Again, this is just for fun. The actual results do not really matter, so I don't plan to make it too strict like a scientific study. In the end, I hope I can get a good laugh with all of them instead of being sabotaged with payback pranks. (>;v>

Sorry for any bad English.

Thanks for reading, and cheers!

o(^~^)o♫

 

[VNandor]

How to make people think that lossless sounds better than lossy? Just use the moar must be better argument and you are good to go.
 
I'm quite sure some of the more active members here encountered people who made more scientific claims (probably wrong but who i am to judge?) about why lossless makes a difference. You could probably google it and find some pretty plausible explanations. (Plausible to people who have never even heard about what lossy compression is.)
 
You might also say that you can easily notice those differences however with that, you either make them to make the "if He hears it I can hear it for sure" assumption, or simply encourage them to lie so they don't look deaf and incapable (which are not the same at all).

 

[U-3C]

Thanks for your response!

 

[U-3C]

Small update on new version of test:

During the introduction for my volunteers, aside from talking about the theoretical advantages of higher bit depth, higher bit rates and higher sampling rates, etc. in an absolute tone, I will also show images such as this:

http://clas.mq.edu.au/speech/acoustics/frequency/graphics/800px-Pcm.gif

as well as quote the observed negative effects of switching from traditional analog media to digital media by famous artists such as Neil Young. I'll try to instill the concept of musical "soul" that is claimed to be subjectively found in traditional media, from real instruments to vinyl records.

For the experimenting section, I will first conduct a blind test and see if people can tell a difference. I assume no as I will be playing the same file. Then, I will announce which file I will play next, and guide the volunteers to listen to the file as much as they want. When they are satisfied, I will do the same for the "different" file. I hope that I will be able to make them hear a difference. I will also encourage them to try different headphones and guide their analysis based on what is going on.

Finally, I will reveal the truth and say that I want to do one final, serious listening test. Of course, the files will remain the same.

Any suggestions/advice?

Any thoughts are appreciated. If you have anything you want me to mention/try out, or if you think this is stupid, please tell me. More heads = more ideas, right?

Cheers!

o(^_^)o

 

[castleofargh]

the staircase graph is a long time winner. almost any manufacturer with nothing to sell will show it, be it highres audio or discrete stuff.
you can also mislead people on purpose, so that they will follow you instead of all having to create their own placebo. like talking about how the guitar has more micro details, or some cymbal being more realistic, or how the soundstage opens up...  at that point when listening we are looking for whatever you suggest and the brain won't stop at anything to make us succeed, or at least think we did.
I wouldn't suggest selling it with too many kinds of details, maybe just one or 2 particular points to look for, and the rest made up of vague ideas of overall improvement that are real(like lower digital noise or something).
 
if you can show a physical representation of 2 products, it would greatly help too. as once I have seen 2 different things, I no longer have any reason to contest that there are indeed 2 different products. maybe showing a DVD, blue ray, DSD. or if that's not possible, simply show a web page with the song sold at a ludicrous resolution for high money and say you got it there. anything helping the test subject to "see" different products.

 

[VNandor]

Slightly OT but I never understood what's wrong with the staircase image. I mean this is what actually happens isn't it? But after a certain point better sampling rate and bit depth doesn't make a difference because 44.1/16 is easily "good enough".

 

[RRod]

  Slightly OT but I never understood what's wrong with the staircase image. I mean this is what actually happens isn't it? But after a certain point better sampling rate and bit depth doesn't make a difference because 44.1/16 is easily "good enough".

 
It's what happens if you don't low-pass a zero-order hold. It's not what you would see on an oscilloscope if you hooked up a DAC trying for proper reconstruction.

 

[castleofargh]

  Slightly OT but I never understood what's wrong with the staircase image. I mean this is what actually happens isn't it? But after a certain point better sampling rate and bit depth doesn't make a difference because 44.1/16 is easily "good enough".

 
on the digital side of things, we deal with samples, so points would be a better representation. the problem with showing staircases and a sine is IMO how people tend to assume that the digital to analog process is only that, moving from point to point to draw the signal like we could do by hand, and then we're done. the kind of logic where Nyquist doesn't come into play and where a guy seeing 4 or 5 points to draw a 10khz sine will inevitably think "this is crap, it can't reproduce the real sound".  because that's the intuitive conclusion when we don't know much about digital audio. except it's been demonstrated to be false both in theory and in practice. so IMO it's a graph that can only be shown to an average crowd with the intended purpose of fooling them.
it gives people the idea that they understand digital to analog conversion when they totally forget about stuff like a low pass filter making straight vertical lines an impossible form at the output of the DAC.
 

Spoiler: exotic exception

excluding some NOS, no filter, discrete DAC manufacturer who decided to "demonstrate" the quality of his DAC by showing the output of square wave signals. but that's another level of deception, where showing good square waves is suppose to "demonstrate" the ability to reconstruct good sine waves. in fact, the absence of both low pass and oversampling will forbid to reconstruct good sine waves as well as a DAC that can't do square waves.

 

[U-3C]

Okay, thanks for the responses!

I guess I'll pretend I'm doing an abx test, where I play both files as I explain/mislead the volunteers, telling them what the differences are amd ask for constant feedback. Then I will do a blind test. The main section that I will be collecting data from is the part where they know which is which. This should be a bit better than the original test as it should erase the obvious plot hole regarding why I'm not doing a blind test.

At certain times, I will tell them to focus on a particular section, for example: "at this part of the uncompressed version, close your eyes and try to locate the instruments..."

I guess I will take a file and encode it to wav or something, so the file sizes are different (hopefully <2 mb vs >300 mb) and demonstrate both to the volunteers as visual proof that one file definitely has something that the second file is lacking.

I guess I can also ask the confidence that the volunteers have regarding the blind test results. I know most people don't care that much about audio, they might get the idea that they are just guessing? For the few that do care (including one who claims to be an audiophile), I wonder how their confident they are with their judgments.

I'm not sure if I can get enough "specialists" for this test though. Maybe I can increase the confidence of those who play instruments enough so that they believe they have golden ears vs the average Joe?

The confidence part is based on this video:

https://m.youtube.com/watch?v=v0XQd9zYMc0

 

[VNandor]

Does it have to be a "blind" test? A lot of people rely on sighted evaluation so maybe they do too. You could just let them switch and see which is which and ask if they hear a difference.

 

[U-3C]

Does it have to be a "blind" test? A lot of people rely on sighted evaluation so maybe they do too. You could just let them switch and see which is which and ask if they hear a difference.

Since this is kind of like a social experiment, I want to avoid them asking the question of why I'm not doing a blind test, as that should be the most effective way of seeing if they can actually hear a difference. As a result, I want to fake a blind test so nobody figures out that I may be lying to them.

(>^_^)>

 

[serman005]

Since this is kind of like a social experiment, I want to avoid them asking the question of why I'm not doing a blind test, as that should be the most effective way of seeing if they can actually hear a difference. As a result, I want to fake a blind test so nobody figures out that I may be lying to them.

I like your plan (diabolical...).

 

[1TrickPony]

.

 

[U-3C]

I like your plan (diabolical...).

\(> v <)/

 

[U-3C]

 
I like your plan (diabolical...).

Feel free to conduct this unethical human experimentation on your friends and foes and friends again!