So this is my riddle. 

 

LIVE performance or ARTIFICIAL?

 

Can you tell which of these sound files contain "live performances" and which ones contain "synthetic" or computer generated music?

 

My hypothesis is that people will not consistently be able to tell the difference (Since there are seven examples, it would be extremely difficult to guess all of them correctly).

 

Pieces to be found at this link:

 

https://soundcloud.com/olmnilnlolm

 

(This is not my main music site, but just a place where I put fragments, works in progress and experimental bits of pieces—You will find all seven pieces in order, at the top, marked Piece 1, Piece 2, Piece 3, etc.)

 

Here is the basic information for the experiment:  

 

1.  I have produced a collection of seven sound files.  (It would be too easy to guess if I just gave you two or three).

 

2.  Some files contain "live performances."

 

3.  Some contain music produced synthetically, using midis and computer software.

 

4.  Each sound file is representing music played by two or more pianos.  No instruments other than pianos are represented.

 

5.   Background noise is inserted into each file.  [Each file begins with a recording of applause, and an announcer saying,

 

"Well, I promise you, this is the piece everyone here tonight has come to experience." 

 

What follows is an overlapping sound file containing the background noise of an attentive audience.  Chair movements, people shifting in their seats, a phone or two beeping, various human noises and other sounds, are part of the environment of each file (so that no noise can serve as a tell tale sign to indicate any sonic difference the between "live" and artificial performance). ]

 

6.  All music, whether artificially produced or performed live, comes from known composers, names taken from this rather long list of modern and contemporary composers.   (Source:  University of Cambridge Music Library)  

 

http://www.lib.cam.ac.uk/Departments/Music/moderncomposers.html

 

Listeners may find the tonality of all the selections rather challenging.  None of the works are in a simple traditional or pre-modern diatonic mode.   There is deliberate dissonance, often going beyond mere polytonality and even standard dodecaphony.

 

7.  No other information or hints need be given to those who listen to the sound files.

 

Forum members are asked if they can tell the difference between the sound files containing "live performances," and those which contain "simulated" or "artificial" music, produced by the computer software.

 

Again, there are seven musical examples. They are short. These are selections only, lasting approximately one minute.

 

So can you tell which is which?

 

LIVE performance, or ARTIFICIAL performance?

 

Piece 1

 

Piece 2

 

Piece 3

 

Piece 4

 

Piece 5

 

Piece 6

 

Piece 7

 

 

 

 

 

 

 

You need to be a member of Composers' Forum to add comments!

Join Composers' Forum

Email me when people reply –

Replies

  • Thank you Bob, for participating.  Does anyone else want to try?  I will wait until after more people participate before I post the correct answers.

  • To me, 2, 3, 4, 6 are live.

    This is doubly difficult…as one feels quite some distance from the sound being produced, making it harder to tell with re: to percussiveness, and timbre. 

    Honestly, I wasn't certain of any of them..

    (After completing this experiment…And I am very curious to the results of this one, … perhaps  it would be interesting to try it from the other side of the spectrum:  Acoustic piano studio recordings compared to midi piano studio recordings… It may be very difficult to 'tell' there as well.. and ambient  noise would be needed to be added there as well, i'm sure.  breathing and such… Just don't use any Glenn Gould recordings.. lol).

    Thanks for conducted the experiment, O.

    gregorio

  •  

     

    Bob Porter said:

     

    "I believe that 2, 4, and 7 are live."

     

    I thanked him for his participation.

     

    Gregorio said,

     

    "To me, 2, 3, 4, 6 are live."

     

    Now I thank Gregorio for his willingness to participate "in the experiment."

     

    Again, I will give the correct answers after more people have had a chance to weigh in.

     

    Peter Brown said,

     

    "The a-tonal performance and the DE-tuned strings of the piano coupled with the studio sized room ambiance make it impossible for me note the difference. Now, if I can only figure out what the point is....   :- [  "

     

    I didn't say this would be easy.  (In fact, I am not sure if I could make the distinction I am asking others to make).

     

    The point is to see if people can determine the difference between a "live performance" and a computer generated performance.

     

    I must say, to being with, that none of the strings of any of the pianos are "DE-tuned."  Permit me to stress that point.  There are NO pieces for "prepared pianos," a la John Cage.  Allow me to add, also, I don't see how a performance in a "studio sized room," or some other sized room, should make a difference in determining whether a performance is live or simulated.  Presumably, there is something that " a human being" will add to a recorded "live performance" that a computer generated sound file alone cannot achieve.  At least, this is what some people maintain.

     

    Yesterday, I noticed that this thread had been viewed 30 times.  Now, the number exceeds 60.  Yet only two people have offered answers to the question.  (I take Peter's response to indicate that he does not believe it is possible to consistently tell the difference between a live and an artificial performance, so I suppose that's kind of answer, too).

     

    I continue to invite more people to participate in this experiment.  The more people the better, as regards conclusions that might be drawn from the results.

     

    Gregorio said,

     

    "(After completing this experiment…And I am very curious to the results of this one, … perhaps  it would be interesting to try it from the other side of the spectrum:  Acoustic piano studio recordings compared to midi piano studio recordings…")

     

    Can you say more about this, Gregorio?  I am not sure what you mean by "midi piano studio recordings"?

     

     

     

     

  • Sorry for the confusion, perhaps i should have said,

     "Acoustic piano studio recordings compared to midi piano  recordings"  

    (btw, now that you mention that there was no de-tuning of the pianos a la Cage, , i am certain  that one of my guesses are wrong, I think that was #6.  :(



  • Peter Brown said:

    This is insane. I have made many, many recordings of acoustic pianos. Many were actual recordings of 'the upright' piano I miked personally. Others were MIDI pianos of acoustic uprights, consoles, grands and other various sized acoustic pianos. I can make a poorly 'sampled' MIDI piano sound like a real acoustic piano recorded with cheap mikes in a bathroom with cracked ceramic tile.

    I can make a 1955 Austrian 6' 7" Bosendorfer melt the shoes off the prettiest girl in the city. I can make a spinet piano sound like a 9' Mason Hamlin concert grand.

    But I still don't get the point.  

    Help!  :-o

    I think the original point was to show how blurry the line has become between 'real' instruments and virtual ones, as far as the listener is concerned.  If i'm understanding your point correctly though, i think you are correct.  It's much easier to hide a virtual performance in a certain context.  

    I listened to the examples, but i'm not going to bother to guess.  I've seen this experiment done a couple times on the web and heard for myself.  I'm already a believer in how far virtual instruments have come and what they can do.  So that at least explains my adding the to the view count without adding to the guesses count.  

  • Susan said,

     

    "Ondib, I'll will give it a try tomorrow."

     

    Thank you, Susan.

     

    No hurry, of course.  Take your time.

     

    Reply by gregorio X 1 hour ago:

    'Sorry for the confusion, perhaps i should have said,

    "Acoustic piano studio recordings compared to midi piano  recordings" '

     

    That's interesting. Personally, I think few people will fail to tell the difference between a midi piano (by itself) and a live piano performance.  That's if we are talking about a pure midi. There is something "tinny" or highly artificial sounding to me in a midi by itself (even a piano midi). However, a midi, modified by more advanced computer software, in my mind, is able to come much closer to, and to become nearly indistinguishable from (or perhaps totally indistinguishable from) a "real piano."  This is just my impression, based on listening to pure midis and modified midis.   All the midi information on these sound files has been put through a relatively sophisticated set of alterations within computer composition software, so as make it sound "more real" (or "less unreal.")

     

    But still, the issue is not whether the instrument sounds "tinny," (or unlike a standard piano) but whether or not the human element contributes something to the overall sound patterns, TO THE MUSIC ITSELF, that the computerized performance cannot contribute.  If the human element is indeed noticeable, what might follow logically?  One could presumably tell the difference between a human being playing a toy piano and a computer generated performance of music for a toy piano.  So there is more to this than the simple nature of the instrument itself, and how it sounds.

     

    I played these experimental files for my wife this evening, and she said she thought it impossible to tell the difference between those which contained recordings of "live performances" and those which contained artificially constructed music created from midis.  (She probably has, in a practical sense, a much better ear than I do, a more acute sense of hearing, of aural perception than I do—and I admit from the beginning that I could not distinguish the two types of files).

     

     

    '(btw, now that you mention that there was no de-tuning of the pianos a la Cage, ... )'

     

    Whether or not there was deliberate de-tuning should not radically affect people's ability to tell if the piece is being played by a "live person," should it?

     

    How might that be relevant, I wonder.

     

    As a related question: Do you think a listener could tell the difference between a "live human performer" banging randomly on a piano, and a computer file of an artificial piano playing randomly generated notes?   

     

     

     

     

     

     

     

    gregorio X's Discussions
    gregorio X's Discussions | Meet composers from all over the world. Share ideas, thoughts, tech insights, job opportunities, and business knowledge!
  • Thanks for the entertaining challenge.

    I listened once to each, but in one case (of #5) I felt unsure and listened 2 more times before venturing a guess.

    Here are my guesses:

    H = human player

    C = computerized playing

    1 - H

    2 - H

    3 - H

    4 - C

    5 - C

    6 - C

    7 - C

    I had a glass of wine after the first three.

  • O- you ask:

      "

    "Whether or not there was deliberate de-tuning should not radically affect people's ability to tell if the piece is being played by a "live person," should it? "

    No it shouldn't, of course.. And it didn't,  to me.  (because I was unaware while listening). Now  that you mention  that there were no de- tuned acoustic pianos,  by simply deduction I realized my answer was wrong… 

    O- You ask:

    "As a related question: Do you think a listener could tell the difference between a "live human performer" banging randomly on a piano, and a computer file of an artificial piano playing randomly generated notes?"

    I think it would be more difficult to tell, with complex tone clusters… But still there should be some tell signs..But I think it is getting increasingly difficult…In fact, some of what i have heard in the way of demos, has been astonishing… Of course then it just gets back to the 'live player', verses midi programmer.

     Amazing piano samples …They are definitely out there… with special attention to the most subtle of qualities.. 

      

  • Thank you, Mariza, for your entry into what is becoming a more and more interesting experiment and discussion.

    I doubt one glass of wine would make any difference in your abilit

  • [continuation of message to Mariza]

    ... your ability to discern the difference.   But you do say you are "guessing."  

    That in itself is helpful information (perhaps ... though I am not giving anything away by saying that.)

    [At least, I don't think I am ... I hope not ].

    I have a question for Gregorio, but I'll save that for a future post.

    Yes, the "samples" are interesting, and they were a recent discovery, which I will share with everyone at the end of the "experiment," so everyone can listen to them, analyze them, and make of them what they will.  (Of course they sound different, depending upon how they are worked upon, within the composition software).

This reply was deleted.