a sound experiment

[pogo97]

 

Did you change them round a little later than 35 votes?

 

 

I changed them at about 9:20 am est if that helps you make sense of it.

 

 

Right now, the results graph shows 12 votes for example 1 and 21 votes for example two. Because of the switcheroo (which I know is confusing, but necessary, I think) that really means 20 votes for current example 1 and 13 votes for example two. I figure I'll wait until there are a total of 70 votes (35 + 35) and then tell you the secret and try to interpret the results.

 

Thanks for contributing.

[pgunders]

I prefer the one that's example 2 at the moment. It sounds less compressed to me.

[program_insect]

I listened to both several times, and the only conclusion I can arrive at is that I definitely prefer the one that was sample one before it became sample two, at which point the one I prefered was actually what everyone else was calling sample one, which you will certainly agree was actually the first sample, called sample two.

[Nevandal]

Yeah...why'd you switch them?

 

Now we're all {censored}ing confused.

 

 

[-groovatious-]

Yeah...why'd you switch them? Now we're all {censored}ing confused.

 

Is CHAS back??

[Nevandal]

A couple of cusswords and sarcasm, and I'm the worst person in the world? Heh. Awesome.

 

 

EDIT: Not that CHAS is the worst person in the world. I don't even know CHAS

[pogo97]

Hi Nevandal. Here's why I switched them.

 

In any experiment with living subjects there is the effect you're trying to test for and there are other things that may skew the results, called 'confounds'. Confounds may lie in the instructions or in the subject pool or in a variety of other places such as--and this is why I switched them--in the order that the examples are presented. I know that when I hear a piece, I'm likely to enjoy it better the second time because I have a better idea what to expect and I can listen more closely to what's going on inside the music. Thus, in this experiment, if I left the order the same throughout, I could easily be testing whether people prefer any piece the second time around.

 

In a well-financed 'real' music perception experiment, I might control for this by having a second group of subjects hear truly identical versions and ask them to rank them and then apply that to the findings (there are statistical methods for doing stuff like that--utterly beyond me). But I decided, instead, to control for it by simply switching the order half-way through the experiment. There may be problems with that, too; I'm not an experimental psychologist (though I do some programming and administrative work in that field and took a graduate course in music perception/cognition at one point so I'm not utterly ignorant).

 

Anyway, this is a rough-and-ready experiment about a topic that I find interesting. We can discuss how relevant the results are when the results are complete and I tell you all what the difference really was. The comments so far have been fascinating, so I'll be very interested to hear from you then.

[Tony Scharf]

Wouldnt it have been better, however, to switch them without telling people they were switched? being aware of the switch will skew the results, simply by fact of peer influence. I mean, if you read that everyone else prefered the second example, wont that influence some people into choosing the first sample if they knew that they were replying after a switch? Conversely, some people will select the less popular example just to be contrary.

 

I think to really do this right, you should have switched the samples, but also asked people not to comment on their choices. that way, no one could see and be influenced by other respondents choices.

 

of course, you may just be f*cking with us all, and its the effect of switching the samples and telling us that your actually testing for

[pogo97]

If you read the instructions, they ask people not to read the comments before doing the poll.

[Tony Scharf]

 

If you read the instructions, they ask people not to read the comments before doing the poll.

 

 

but can you count on people to follow the instructions, particularly HC members?

[Nevandal]

Hi Nevandal. Here's why I switched them. In any experiment with living subjects there is the effect you're trying to test for and there are other things that may skew the results, called 'confounds'. Confounds may lie in the instructions or in the subject pool or in a variety of other places such as--and this is why I switched them--in the order that the examples are presented. I know that when I hear a piece, I'm likely to enjoy it better the second time because I have a better idea what to expect and I can listen more closely to what's going on inside the music. Thus, in this experiment, if I left the order the same throughout, I could easily be testing whether people prefer any piece the second time around. In a well-financed 'real' music perception experiment, I might control for this by having a second group of subjects hear truly identical versions and ask them to rank them and then apply that to the findings (there are statistical methods for doing stuff like that--utterly beyond me). But I decided, instead, to control for it by simply switching the order half-way through the experiment. There may be problems with that, too; I'm not an experimental psychologist (though I do some programming and administrative work in that field and took a graduate course in music perception/cognition at one point so I'm not utterly ignorant). Anyway, this is a rough-and-ready experiment about a topic that I find interesting. We can discuss how relevant the results are when the results are complete and I tell you all what the difference really was. The comments so far have been fascinating, so I'll be very interested to hear from you then.

 

 

Hmm, yeah. That's actually a good idea. I don't know if switching the two was the best way to go about that, though. I tried to listen as subjectively as possible, and I listened again a few hours later. I do know what you mean, though, and I've done something like this before. What I did to counteract it is to put a clip which I thought was inferior as the 2nd clip, because (it seems) like most people might like the 2nd clip better. *shrug*

 

 

I definitely did listen to both clips more than a handful of times before deciding

 

I just think that switching them halfway through makes the after-discussion a bit painful and confusing! (especially since I saved the files to disk, and listened to them many hours later, who knows which is which now, haha)

 

 

Thanks for the response...I can't wait to find out what the actual difference was.

 

 

 

EDIT: Okay, listened to them again, and my first post was analyzing their original configuration. I can tell that they've been switched around.

[alcohol]

I did an experiment with myself with guitar cables. Wasn't a blind listening test. But what I noticed with myself. That if two things sounded close. What the last thing I listened to I perceived it to sound better. This mistake is often repeated in recording studios when someone says "add more EQ," then they turn the knob and everyone agrees it sounds better even though the EQ wasn't engaged.

[llamastorm]

First piece felt like too much sustain or something, but the second piece over-reverbed.

 

So #1 I guess.

[grumphh]

I figure I'll wait until there are a total of 70 votes (35 + 35) and then tell you the secret

Within a few months we shall be very anxious to know that secret

 

... i think 70 votes is quite a high number to shoot for... - especially if you don't keep bumping this to the front page.

[pogo97]

Yeah, you're probably right. I'll consider pulling the plug tomorrow. I did, though, want a balance between the two orders.

[r33k]

Most of you guys are way off the mark. The two examples demonstrate two different tunings on the piano sound. Someone earlier had it right when he mentioned the very first chord. That's all you need to hear the difference. In one example the chord beats internally, and doesn't in the other one. Go back and listen again, peeps. This is good ear training practice. Oh, and Doug, it probably would've been better to use an acoustic piano sound instead of that CP80 sound. My $0.02 only.

[Oscius]

The piano sounds a bit out of tune in current example 2. Kinda like recorded with tape recorder but not even close as bad, it is brighter and colder too, which I like.

 

Thus I like the first one better but there is way not enough treble. I can't really say I like the piano sound on either... Is that kurz? Or roland?

 

I was too lazy to listen to the backgrounds so no word on them.

 

Vote: example 1

 

...But like I said fix the EQ, maybe add some low pass filter too. the piano sounds just too cheesy on the 2nd. Even though I acualy prefer colder sounds The example 2 with more in tune "cold" piano patch would be great.

[aeon]

The first thing I noticed was the beating of the first chord in the one example, and not in the other.

 

I preferred the one with the beating. That was example #1 when I took it.

 

I liked both for the melody. I didn't care much for the soundstage and recording.

 

 

cheers,

Ian

[Tullsterx]

The strings are not as muddy on the second one. Seems like the strings are "cloudy" on the first one.

[Nevandal]

Most of you guys are way off the mark. The two examples demonstrate two different tunings on the piano sound. Someone earlier had it right when he mentioned the very first chord. That's all you need to hear the difference. In one example the chord beats internally, and doesn't in the other one. Go back and listen again, peeps. This is good ear training practice. Oh, and Doug, it probably would've been better to use an acoustic piano sound instead of that CP80 sound. My $0.02 only.

 

 

Yeah, after I noticed the difference about 2 or 3 seconds into listening to the 2nd example, all I did was switch back and forth a bunch of times to confirm my suspicions.

 

oh, and /bump

[pogo97]

As promised...

 

The two examples are identical, except that:

(current) Example 1 is in equal temperament

(current) Example 2 is in Werkmeister III temperament (google is your friend)

 

 

Once you adjust for the switch in example order after 30 responses, 31 people "liked" the equal temperament example better, 19 people liked the Werkmeister III example better, 11 didn't have a preference and 5 didn't hear a difference.

 

The 'piano' was Pianoteq's CP-80 preset with unison width and octave stretch reduced to 0, and piano length increased to 10 m. I did this in order to achieve as 'transparent' a sound as I could so that the difference in temperaments would be relatively obvious. As some pointed out, it did not result in a very realistic piano sound; such is life.

 

The synth strings were Logic ES2's JP Strings Dark preset. I reset the pitches of the oscillators to the same pitch, rather than slightly detuned. I probably should have, but didn't, turn off various modulators, including poly AT.

 

Back to work; I'll have more comments and possibly some more examples later.

[asynchro_nous]

Why don't moody and woody rhyme?

 

Interesting thread; nice demo of the modeling software as well.

[Nevandal]

Cool.

 

I'm still not sure which one I prefer in terms of sound, since the piano didn't really sound too great on both of them...but after reading about just intonation and myriads of other tuning systems, I can tell you that I greatly prefer 12 tone equal temperament and any other tuning system is just...crazy, convoluted, and too much of a hassle to work with.

 

Cool experiment, though.

[-groovatious-]

Thought so.

[pogo97]

Me, I'll keep working with unequal temperaments. More meat there.