Can people tell the difference between a recording of a "live performance" and an "artificially produced" performance?

 

As far as I am aware, no serious research has been done on this question.

 

I have heard people assert that recordings of live performances are qualitatively different from artificially produced performances (while implying that they can tell the difference between the two, presumably by paying careful attention to their aural characteristics).

 

But I have doubts. 

 

Can people honestly say that, under experimentally controlled conditions, they can make this type of distinction?  Can they consistently tell the difference between recordings of "live performances" and "artificially produced" performances in a situation such as this:

 

1.  Imagine if I produce a collection of six or seven sound files.

 

2.  Some sound files contain "live performances."

 

3.  Some sound files contain music produced from midis.

 

A question for anyone who reads this:

 

If I post all of these on this site, do you think people could consistently tell the difference between the two?  

 

[I am preparing such a set of items now, each about one minute long].

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

Join Composers' Forum

Email me when people reply –

Replies

  • Hi O.

    Interesting subject for experiment.

    You say: "If I post all of these on this site, do you think people could consistently tell the difference between the two?"

    Well for myself, I have been fooled many times… There is one site, (which I can't remember) that uses midi to produce orchestral compositions- especially ones that have been overlooked - and perhaps would continue to be  overlooked, - (without the cost needed to hire an orchestra)..  How wonderful… And to me, it totally 'worked'.

    I think of this idea as having various dimensions to it.  One, is the composition itself.. Another - is how it is realized.

    I can appreciate these qualities as separately in the same piece, and sometimes the lines are indeed blurred, sometimes to the point where the message and the means are the same thing --depending on the composition.

    To me, if it is a midi rendering of an orchestra, and Meant to sound like an orchestra, then I am Looking to be 'fooled'.. I want the spell to work! And many times it does!  

    But even of it doesn't 'come off' convincingly as a performance, I  still can look at the compositional aspect alone, as i believe we composers  can do.

    Curious to me is the idea (that seems to have little weight)  are the possibilities  Not of imitation, but the incredible dimensions available in the electronic media to compose something utterly unique, which an orchestra would be at odds to try to reproduce… In other words, composition written specifically For the electronic medium As it's ideal… allowing for infinite continuum of pitch (and new harmonic content) , rhythmic description, and an endless supply of new timbres.   

    M. Subotnik being just one of those composers in the above mentioned description.

    And Yet, (to turn it around again),  sometimes my piece actually Calls for 'cheesy string sound'..  :)  (where the 'real thing'  just wouldn't  create what i was going for)

    gregorio

  •  

     

    Below are my responses to replies by Peter (Peter Brown ), Wladislaw (Wladislaw Leszczewski) , Bob (Bob Porter) , and Gregorio (gregorio X).

     

     

    Reply by Peter Brown 1 hour ago: "People? ... "If I were drunk, maybe not. When I've a little too much libation, everything seems like 'an artificially produced performance' weather or not it's been recorded."

     

    I suppose a person who is drunk enough can't tell the difference between the voice of a stranger coming out of the radio and the voice of a familiar person  sitting a foot or two away.  [I believe there is scientific research to substantiate that point, probably a plethora of empirical data, not to mention anecdotal evidence].

     

    You say, Peter, 'I also think it might have to do with someone's animosity toward electronic music in general. i.e., "that's not real music. It's electronic." Die hard analog gurus have a penchant  for non-electronically manipulated audio. They're sometimes hostile toward music that is produced via samples and the software that supports it no matter if they can tell the difference or not.'

     

    If they cannot tell the difference, then we would know the "animosity" is simply the result of prejudice, and not based upon any reasoned out set of aesthetic criteria.

     

    'Just the idea that the music was triggered from soft samples on a DAW makes them defiant toward the artist as well as the industry that created and developed the technology.... Through the years, technology in the digital music industry has not only matched the quality of analog recording, (digital sample, vsts etc...) but has come to offer the artist more and better digital solutions for the artist's musical pallet."

     

    I think the hypothesis I want to test is a bit subtler, however, than the simple question of digital versus analogue.  People maintain that the recording of a "live performance" (whether digitally recorded or on an old vinyl record) is qualitatively different from a sound file, for various reasons, some of those being reasons you have mentioned.  I want to see if people can really tell the difference.

     

    Reply by Wladislaw Leszczewski 39 minutes ago: "What if you imitate a computer or randomize some aspects of the computer-generated music?"

     

    Good question.  I am not intending to give people a choice between "a live performer imitating a computer" and a piece of "computer generated" music, which attempts to imitate humans.  I can say more about the types of pieces I intend to put up a bit later.  Suffice it to say, all music used, whether artificially produced or played live, would come from known composers, on one or more of these lists:

     

    http://people.wku.edu/charles.smith/music/images/stats6.pdf

     

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

     

    You say, "The sound matters more than the thing that was told to make the sound."

     

    I believe that is one of the statements being tested by the experiment.  If you are correct, then a person should not always be able to distinguish whether the sound was produced by "live performers" or by a computer generated musical work, coming from a midi.

     

    You conclude, "In the end, we're all computers, but we're made of neurons instead of silicon."

     

    I think that is an ontological assertion, which can be debated.  I asked someone very recently, "What percentage of us is flesh, nerves and bone, and what percentage is mind, soul, or spirit?"  She said, "I think 3,000 % is mind, soul, and spirit."  It's an interesting statistic, and hardly less verifiable than the statement that we are "computers."  I think the number 3,000 % reflects the belief that human mind, soul and spirit are qualitatively different from any mere "quantity" of, and arrangement of atoms, molecules and biological components. That would be borne out by philosophers such as Hegel, Schelling, Bergson and Teilhard de Chardin.    

     

    Reply by Bob Porter 30 minutes ago:  "Peter, ... Some things may never change. Even if they don't, each of us needs to just keep working in what ever medium we feel we must. My wife is a colored pencil artist. In the art world, oil is king. Acrylic is accepted. Water color is getting there. Colored pencil is so far down the list that it might as well not be there. Oh well. It is possible for both sides to be snobs.  ...  It will be interesting to hear the same piece, side by side, live recording and virtual."

    I am hoping it will be interesting, and even dispositive of some basic facts about human perception in the age of computer generated music. 

     

    Reply by gregorio X 14 minutes ago:

     

    "Hi O."

     

    Hello, Gregorio.

     

    "Interesting subject for experiment."

     

    I hope so.

     

    You say, "If I post all of these on this site, do you think people could consistently tell the difference between the two?"

     

    I did say that, yes.

     

    Well for myself, I have been fooled many times… There is one site, (which I can't remember) that uses midi to produce orchestral compositions- especially ones that have been overlooked - and perhaps would continue to be  overlooked, - (without the cost needed to hire an orchestra)..  How wonderful… And to me, it totally 'worked'.

     

    I thought the Japanese site was an interesting one [link below], though I think most people can easily tell the difference between what we find there, and "music performed live."

     

    Here, try the I-dur version of Telemann's Concerto in D for 2 Horns 1.Spiritoso ma non Allegro Adagio

     

    http://music.i-dur.com/index.php?am=523&bm=1&em=rc

     

    Not bad, in my view, but clearly distinguishable from a live performance. 

     

    "I think of this idea as having various dimensions to it.  One, is the composition itself.. Another - is how it is realized."

     

    Yes. Go on.  I'm listening.

     

    "I can appreciate these qualities as separately in the same piece, and sometimes the lines are indeed blurred, sometimes to the point where the message and the means are the same thing --depending on the composition."

     

    I see.

     

    "To me, if it is a midi rendering of an orchestra, and Meant to sound like an orchestra, then I am Looking to be 'fooled'.. I want the spell to work! And many times it does!"

     

    I understand.  Perhaps you could give a specific example, from a web site, for instance.  But I know what you mean. I just listened to this, and for several moments, I did find myself "able to fall into the spell," so to speak.

     

    Rimsky-Korsakov:Scheherazade 1 op.35

    「海とシンドバットの船」シェラザードが第290夜から第315夜に渡っ語った「船乗りシンドバッドの物語」による音楽

     

    http://music.i-dur.com/index.php?am=336&bm=1&em=rc

     

    I intend to make my experiment in such a way that it will be very difficult for people to distinguish the two, live performance and virtual performance.  For starters, I will only use music written for piano.  Each "piece of music" will contain two or more pianos.  Most people will agree that virtual pianos are more difficult to distinguish from "real live pianos" than say, virtual and real string instruments, such as solo violins.

      

    "But even if it doesn't 'come off' convincingly as a performance, I  still can look at the compositional aspect alone, as i believe we composers  can do."

     

    That's understandable.  What is at issue in this experiment is not simply whether one can tell the difference between performances of the same piece of music, note for note, one version "played live," and the second version "played virtually" via computer.  It's about the general sound, and the assertion that recordings of "live performances" are easily distinguishable generally from so-called non-live, or artificially produced computer performances.

     

    "Curious to me is the idea (that seems to have little weight)  are the possibilities.  Not of imitation, but the incredible dimensions available in the electronic media to compose something utterly unique, which an orchestra would be at odds to try to reproduce… In other words, composition written specifically For the electronic medium As it's ideal… allowing for infinite continuum of pitch (and new harmonic content) , rhythmic description, and an endless supply of new timbres."

     

    That is a different issue, of course, and I am fully with you on this.

       

    "M. Subotnik being just one of those composers in the above mentioned description."

     

    I am a great admirer of what he achieved in "The Wild Bull," and "Silver Apples of the Moon," the only two works by Subotnik I know. So yes, I agree.

     

    You mentioned "your piece."  I don't know which piece you were referring to.  A recent one?

     

     

    Peter Brown's Discussions
    Peter Brown's Discussions | Meet composers from all over the world. Share ideas, thoughts, tech insights, job opportunities, and business knowledge!
  • Oh, a Mellotron. I had one of those. It was like playing a tractor. When you hit to many notes - that is more than three - the motor pulling the tape over the tape heads started to go slower and the pitch dropped! There is the same snobbish attitude about old analog synths and moderna plugins too. I definitly prefare a sampled instrument to trying to play a tractor!

    In my own music I often make arrangements with both pure synth sounds and orchestral samples, just because it's no hiding that it's not a real orchestra and I don't want it to sound fake. I guess I don't have a good enough orchestral sample library.

    I did The first movement of Mozarts symphony no 40 with synths only, as a experiment. It was very educational to hear that music in acompleatly new perspective. It taught me loads of new things about the music. You can hear it here: https://soundcloud.com/lennartostman/mozart-symfoni-40-1-synth-version




    Peter Brown said:


    Greg,

    When the mellotron first appeared on the market, The Beatles and countless Pop artists and bands wanted one to enhance and flesh-out their recordings. They almost sounded like strings and flutes. (Even though they were actual recording of them.)  And if mixed well and appropriately, could be very convincing, But not quite as a stand alone instrument. Remember The Beatles's Strawberry Fields? Yep. The flutes were 'triggered' mellotron tapes.  

    Until they started experimenting with sampling instruments. [Digitally recorded instruments processed  through Ai and synthesized electronic manipulation of waves and frequencies.]

    Composers do this every day as part of their tool box full of instruments. Think what Beethoven and Mozart might have done with Ai synthesis if it were around then...

    I believe also, that how these artificially produced sounds are executed, a bigger and better end performance can be had. If it's authenticity you're looking for in your project/production, anyone can be fooled with the technology available to us today. Even those who engineer these artificially produced sounds. It's a science. 

     

    DSC00026.jpg


    gregorio X said:

    Hi O.

    Interesting subject for experiment.

    You say: "If I post all of these on this site, do you think people could consistently tell the difference between the two?"

    Well for myself, I have been fooled many times… There is one site, (which I can't remember) that uses midi to produce orchestral compositions- especially ones that have been overlooked - and perhaps would continue to be  overlooked, - (without the cost needed to hire an orchestra)..  How wonderful… And to me, it totally 'worked'.

    I think of this idea as having various dimensions to it.  One, is the composition itself.. Another - is how it is realized.

    I can appreciate these qualities as separately in the same piece, and sometimes the lines are indeed blurred, sometimes to the point where the message and the means are the same thing --depending on the composition.

    To me, if it is a midi rendering of an orchestra, and Meant to sound like an orchestra, then I am Looking to be 'fooled'.. I want the spell to work! And many times it does!  

    But even of it doesn't 'come off' convincingly as a performance, I  still can look at the compositional aspect alone, as i believe we composers  can do.

    Curious to me is the idea (that seems to have little weight)  are the possibilities  Not of imitation, but the incredible dimensions available in the electronic media to compose something utterly unique, which an orchestra would be at odds to try to reproduce… In other words, composition written specifically For the electronic medium As it's ideal… allowing for infinite continuum of pitch (and new harmonic content) , rhythmic description, and an endless supply of new timbres.   

    M. Subotnik being just one of those composers in the above mentioned description.

    And Yet, (to turn it around again),  sometimes my piece actually Calls for 'cheesy string sound'..  :)  (where the 'real thing'  just wouldn't  create what i was going for)

    gregorio

  • "If it's authenticity you're looking for, go to a live orchestral concert. There's a performance hall in just about every state capitol."

    Did anyone here know that "Pierre" is the state capital of South Dakota.  They do have an orchestra.  But how good is it?

    Might it not be better to listen to a synthesized orchestra, than listen to some orchestras in some state capitals?

    Can't a computer generated music file be "more authentic" than a so-called live performance, if the latter is bad?

  • O- you mention:

    "I am a great admirer of what he achieved in "The Wild Bull," and "Silver Apples of the Moon," the only two works by Subotnik I know. So yes, I agree.

     

    You mentioned "your piece."  I don't know which piece you were referring to.  A recent one?"

      

    "Wild Bull" and "Silver Apples" are my favorite by Subotnik.  Sidewinder is interesting as well.

    As Subotnik continues  from that early stuff, it seems he has reduced his harmonic content, and rhythmic 'freedom' - though not necessarily a bad (lesser) thing - , and perhaps also introduces a 'folk motif' melodic element, and somewhat influenced by minimalism as well… And although I think he does a very good job here, the material to me is Not as exciting… (btw the way, there is one recording that is markedly different from the description that I mentioned, which is "Return"… It is kind of a time capsule - from gregorian chant to baroque (scarlatti ) …and kind of walks thru history.. tracing each 'return of Haley's comet -- which iirc was premiered at Griffith Observatory with accompanying visuals.

     

  • continuing:

    O- you articulate:

      "You mentioned "your piece."  I don't know which piece you were referring to.  A recent one?"

    I just mean that I have availed myself to the glories of cheesy strings, when the composition called for it… 

    I have not (yet) posted that one.

    I'd rather post one where Acoustic instruments Imitate Electronically Generated sounds..  Kind of to topsy turvy the equation here…  (only if requested)

    gregorio

  • Peter, you bring up:

      "When the mellotron first appeared on the market, The Beatles and countless Pop artists and bands wanted one to enhance and flesh-out their recordings. They almost sounded like strings and flutes. (Even though they were actual recording of them.)  And if mixed well and appropriately, could be very convincing, But not quite as a stand alone instrument. Remember The Beatles's Strawberry Fields? Yep. The flutes were 'triggered' mellotron tapes.  "

    Yes, I love the mellotron (and the clavinet, the chamberlain, etc.. love the vintage sound!)

    And, I might add that I have used the mellotron in one of my pieces, and loved the effect!

    gregorio

  • "There are frequently (or occasionally) artifacts of live performance that are missing in non-live performance. Specifically, with a little effort you may be able to hear the keys opening and closing on the oboe in a live performance (depending on the composition and size of the hall, mic placement etc.)  and the squeak on the string as a guitarist changes from one position to another. These artifacts, at least in my experience, are usually cleaned up in a computerized performance."

    Good point.  But these squeaks and so on are "defects."

    And often you may not hear them, as you say, depending upon microphone placement or angle.

    My question is, if you have a set of recordings of live piano performances and a set of sound files of artificial performances (and say you add INTO ALL OF THESE a track of background noise and defects, and people moving around, coughing, or shifting in their seats), can anyone tell which sound files are of the live performances and which are not?

  • Ondib,

    In the original post you wrote:

    1.  Imagine if I produce a collection of six or seven sound files.

    Do we have to imagine this?  Or are you actually going to produce a collection of six or seven sound files?

    And run an experiment on us.

    Perhaps it is necessary to create a great deal of suspense first?  :-)

    Just teasing you for teasing us...

    Mariza

  •   O - you mention:

     

    "Here, try the I-dur version of Telemann's Concerto in D for 2 Horns 1.Spiritoso ma non Allegro Adagio

     

    http://music.i-dur.com/index.php?am=523&bm=1&em=rc

     

    Not bad, in my view, but clearly distinguishable from a live performance. "

    Yes i listened to  the Telemann.. I can tell it is not 'live' , but thought it was pretty good.. 

    O said:

       "I understand.  Perhaps you could give a specific example, from a web site, for instance. "

    Here is the site I was having trouble remembering.  There are many recordings - orchestral and otherwise here.  The link is to Debussy's Trois Nocturnes.. The orchestral rendering of 'Fetes' is the (imo) better of the 3.  

    http://klassik-resampled.de/index.php/en/home/ensembles/orchestra/s...

    I will be curious to hear these comparative recordings you speak of.   In the direction of how things are going, i should think it may be increasingly difficult to tell the difference… But that may depend on the piece… Take for example the idea of 'rubato'  across a whole orchestral mov't… The conductor (i am guessing)  learns to feel this rubato in real time.. And if it is  the long 'line', I would think it would be Very difficult  to feel it - if one is building it  Not in real time.. LIke Furtwangler's versions of Beethoven's syms.  Or perhaps  any orchestral piece with a subtle rubato, or not even subtle… It is this sense of timing that would seem most difficult to make 'real'..

    gregorio

      

This reply was deleted.