Tangled

[Stephen W. Beatty]

An interesting video highlighting the production methods for the film Tangled scored by the composer Menkin. With the advent of MIR and the ability to acheive a superb realism for instruments samples in a real hall, how different are these sounds from the actual orchestra?  If  you are comparing the digital recording of an actual orchestra with an equivalent live ( not a step edited ) midi performance using MIR and VI can a listener really tell the difference in a double blind test? The problem with comparing a step entered (Jay Bacal)  well known Classical compositions and a digital recording of the same compositions by an live orchestra is not the quality of the resulting sound but the gesture and expressive flow of the music. Are we on the verge of  or in an era of equivalence of sound of the recorded real orchestra and a recorded midi sampled orchestra?

Regards,

Stephen W. Beatty    

[tomhartman]

There is simply no way in the known universe that a sampled orchestra is going to sound as good as the score to "Tangled," especially songs like "I See The Light."  

I'll take the double blind test any day of the week;)

[William]

 

@Stephen W. Beatty said:

 Are we on the verge of  or in an era of equivalence of sound of the recorded real orchestra and a recorded midi sampled orchestra?

Regards,

Stephen W. Beatty    

yes, that is an interesting way of putting it and I think we are actually IN that era right now, not just on the verge.   Most samples now are performed by better musicians than the ordinary live orchestra has.  I know that from painful experience and  have noticed in programming several pieces of mine that were played live that the samples are more expressive and musical than the live performances were.  And with MIR the sound quality is now total realism.  for example, I deny that anyone could tell the difference between a trumpet placed on the MIR representation of Vienna Konzerthaus playing a series of dynamic sustains, and a recording of a live player doing the same.  And if that one example is undetectable,  then it becomes a matter of drawing a line between "real" and "sampled" based on musical complexity and expressiveness which is constantly shifting and becoming less and less discernible.

[tomhartman]

Certainly it depends on the piece. Someone plucking one note on a harp while a flute plays is going to be a lot easier to fool the ear with than trying to do "Back to the Future," which at this point in time would be a joke compared to the real thing.

Also, given no reference point (the real thing playing the same piece) allows a sampled mock up to sound convincing.

I would encourage anyone that thinks samples=real to record their best mock up, then have a real section play it, and then play the two results for any ten people. The real thing is always obvious, no matter how good the samples. Without that reference point, constantly at our side to compare, we often get sucked into believing we have reached a point which we haven't.

I'm not sure VSL or any of the others were conceived to replace the real thing, the creators of libraries like this have too much respect to put forth such an arrogant stance. They have allowed us to realize music which we could not afford to in the past, and to provide clients with close approximations of reality, for which we should all be grateful. 

[William]

@aerovons said:

I would encourage anyone that thinks samples=real to record their best mock up, then have a real section play it, and then play the two results for any ten people. The real thing is always obvious, no matter how good the samples. Without that reference point, constantly at our side to compare, we often get sucked into believing we have reached a point which we haven't.

I don't believe this for a second.  It could easily be done and is done every day on television for example.  Almost no one knows that samples are being used instead of live orchestras.  I do, but 99% of people in an audience do not.  Also, I  was never interested in merely creating a duplication of something real. 

I have had the exact same things that I programmed played live.  And the point in programming them was not to reproduce the beauty of the live performance.  It was to do a halfway decent performance of the music which was not accomplished by the live players at all.  In fact, hearing the sample performance AFTER the live performance was a revelation.  Because I realized that the piece I had written wasn't  actually as bad as the players had made it sound.  if I had tried to reproduce the "real thing" it would have been out of tune strings, clams in the brass, squawks in the clarinets, etc. etc. 

"record their best mock up, then have a real section play it"  -----   Oh sure, I'll just go and rustle up a  world-class 90 piece symphony orchestra and have them do a perfect recording session for me and then I'll post it right here for you.   I advise you also, please submit your symphonic size compositions to any professional orchestra and get back to me as soon as they respond enthusiastically that they are anxious to perform your masterpiece.  Actually what will happen is your score will go into a slush pile and sit there for about a decade, untouched, before the janitor throws it away.  

Unless you happen to have a philharmonic waiting for you in an outbuilding on your estate, ready to record at a moment's notice.  I'm eager to hear the results of your test.  Be sure and post it soon! 

[tomhartman]

You don't need a philharmonic William, you can do it with a simple piece.

Why not go by the opinions of the people who do this for a living? All the interviews with them mention doing mockups for the director, then hearing it done by the live orchestra at the real session.

In fact, why not just comment on the VSL demos sanctioned by Herb to show the capabilites of the software at it's best.

TH

[William]

 I do go by those demos as I did some of them.  Also, why do you assume I don't make a living at this?

Never mind though - you are trying to be negative and stubborn and I don't like discussing things with people like that.

I do like the original topic of this which is positive and interesting about samples, and also forward-thinking which is exciting. 

[tomhartman]

Actually, it seems like you are being stubborn. I merely am saying that samples do not equal the real thing, no matter what positive spin you put on the subject. And I do this for a living too, and started doing it before midi (I'm old) so have plenty of experience with real sessions of real people. I love what has happened to sampling, that's why I use VSL and others. But I don't confuse them with the real thing.

And I noticed you didn't answer the question about the demos.

Do you think they sound like a real orchestra playing the material?

TH

[tomhartman]

Thanks. Your reply speaks for itself.

TH

[jeremyroberts]

Boys and Girls,

What you are missing from this "discussion" is that we don't necessarily  write for real the same as we write for the samples.

When i am creating with the library, to be played by the library, I have the luxury of manipulating what I am writing at any given moment to make it sound as good (and feel as good) as I want it to be -- and sometimes that means we write things that we may not have done for the "real guys". And sometimes, it's MUCH harder to write for the samples -- since they are unforgiving.

I mean -- with a good room of musicians, you can give them just about anything, and the best players will make most anything sound as good as it's going to sound. The samples are not nearly as accommodating. If the writing stinks, the sample lib will make sure everyone knows.

Conversely, simply taking an orchestration and trying to render it with the lib is not always going to get results -- since a good orchestrator has the luxury of knowing that the musicians will make it right. But the samples only can do what they can do... and if the orchestration is not written specifically for the samples, you may or may not get the results you wanted.

Yes, I do this for a living. And my work can be heard internationally. And I can't tell you which tracks are real and which are fake (I get to conduct full orchs as well as doing the fake work) else my clients will shoot me. There's a new album coming out in about 8 weeks for a major artist in the US, for a major label, and this label is known for paying for big orchestra dates... except on this project - and the head of A&R for the label (who is a respected conductor) wrote me a note, congratulating me, and wanting to strangle me at the same time -- since he knows what I just did.

The point is that there are some guys doing some very high level work -- and it usually involves some serious audio engineering and producing skills to go with the musicianship. The samples alone only get you so far -- then you have to make them sound and feel real.

Can I fool the listener? You betcha. It may be a hybrid -- sometimes using a soloist + section samples. But i can fool the listeners. Even if the listeners are musicians and conductors. If I can do it, so can others. But remember, we're probably writing for the samples, not the real guys. It's not nearly as easy to do well as some represent. And then add decades of audio and conducting and playing experience -- yes, there are some guys pulling it off. And I'll guarantee that many of us have signed non-disclosures, so we can't really tell you which projects are fake. 

And are we trying to sound "real" or are we trying to sound organic and emotional? What's the PURPOSE of the music?

If you want to be real good at this, go conduct and play in real orchestras for about 20 years, then come back to the samples.

[William]

 jeremyroberts -

I totally agree with that and believe you could fool people.   Also, you made a perfect point about playing/performing with real orchestras and taking that experience to samples. Also about how it is real work to get a good sound out of them but with enough labor and patience and knowledge of orchestration/performance/conducting it can be done.  The amount of work and musical taste that went into the Rite of Spring performance here is incredible and made it sound extremely realistic as well as faithful to the score (not always accomplished as Stravinsky himself always pointed out) AND musically expressive.  It is an incredible thing to listen to.  I remember Jay Bacal told me that he was going to force himself to stop working on it because he could always do more, but had to stop sometime. 

So when some guy like this aerovons comes on here and says "no way in the universe" crap like that it irritates me.    I played for twenty years in symphony, opera, chamber orchestras and in many small ensembles and that was the way I learned about orchestration.  I have always applied that to my sample performances.  In fact, it is only with knowing intuitively from experience what is characteristic, what is difficult, what is easy, what is within a preferred range or dynamic or whatever  that one can really do a good sample performance.  And as pointed out by the first post here  the samples are becoming astoundingly capable of realism as well as musicality.

[Dietz]

Dear all,

Please keep the tone of this interesting topic on a friendly and polite level. Thanks!

[tomhartman]

Hi Jeremy,

Like yourself, I've been at this a long time, on a professional level (it's all I do to support my family), some with real orchestra, some with samples, so I'm very used to the difference. And my point that William decided to make fun of was that I'm old enough to have started before all the sampling stuff really got going. In the beginning down here in Florida I did nothing but live sessions with live strings, woodwinds, etc, for TV and radio. I jumped on sampling as soon as I could afford it, with Emulators, etc, all the way up to now with VSL and LASS, etc, etc.

And again, I think it's great that we have these tools. Not every client has the budget for live musicians, hardly any anymore where I am, save for Sandals Resorts whose commercials I do and use real strings almost on every spot. And I certainly agree with you about having to write "around" samples. It becomes less restrictive it seems all the time, but there are still many things you have to watch, and many mixing tricks to enhance what realism you get from the various libraries. I do it daily so I know what you mean.

 I do believe that if it were possible to play a large orchestral piece side by side with any sampled version, it would still, here in 2011, be fairly obvious which one sounded superior. 

I'm not sure "fooling the listener" should be the litmus test...most listeners couldn't tell whether we recorded a master with proper levels or quality gear either, in that they listen on mp3 players, but WE hear the difference. Jerry Goldsmith once said "Don't ever kid yourself that you are not making music for yourself" and I couldn't agree more...we have to be happy...first.

If there are many among you who, given the choice of using a roomful of excellent live players and a sample library, would choose the library, that's super and more power to you. I would not have think about which I would choose for a nanosecond.

TH

[jasensmith]

Oddly enough, a test like this was done recently in Japan (I've spent the last hour looking for the link but for the life of me I can't find it).  If I remember correctly, they took ten members of a professional orchestra, sat them down in an acoustically treated room and played two versions of the same piece.  They were asked to identify which version was a real and which was sampled.  Wouldn't you know it?  They all chose incorrectly.

 

Actually, the test proctors pulled a fast one on their subjects.  Both versions were sampled.  But they were sampled by two different skilled MIDIstrators, if you will.  They never mentioned which libraries were used only saying that they were top of line, the best money could buy from the most professional manufacturers in the business.  I think it's safe to assume that VSL libraries were used to some extent.

 

Now some flaws that I see with this test, right off the bat, are that they limited their test subjects to musicians who play in an orchestra.  I know a flautist who plays in a professional orchestra and she's never even heard of sample libraries.  So if she were to hear a sampled piece programmed by the great Guy Bacos, she probably wouldn't know the difference either.  Why didn't they use composers who work with samples for a living?  Also, why trick the subjects by using two sampled versions?  Why not use a real one?

 

@Aerovons

You’re right.  Most people who do this for a living would be able to tell the difference but that doesn’t necessarily make a live piece sound better than a sampled one does it?  (I think that’s the point William was trying to make) Maybe it’s all a matter of taste.  I’ve said this on this forum before and I’ll say it again, there are just times when sampled performances sound better than the real thing.  Ask yourself, what is it about sampled performances that give them away?  For me, it’s the fact that samples just sound so pristine and clean.  There are times when that’s what I want in my sound, so samples are a godsend.  I agree with William in that we are at the point where it’s nearly impossible for the AVERAGE JOE (not professionals) to tell the difference between real and fake and if the average Joe comes away remembering my little melodies then who cares whether the orchestra’s real or fake. 

I think we composers think within the realms of the orchestral box too much as well.  How many of you guys run your string or brass samples through various guitar amps and then put a car horn impulse verb on it?  It makes a pretty interesting color.  How many of you have over three hundred players in your Double Bass section as a standard?  There are an infinite number of things you can do with samples that would either be too cumbersome or too expensive with a real orchestra. 

  

I think of samples as another means to an end and not as an artificial means that approximates a real orchestra end.  If the real deal is what you want and you have it at your disposal and you restrict yourself to that orchestral box then more power to you.  If I had a real orchestra at my disposal I’d probably be looking for ways to make them sound like a fake sampled one.  But that’s just me.  Again, I guess it’s all just a matter of taste.        

[William]

My apologies to aerovons if I was offensive but I was in fact joking with the old geezer stuff since I myself AM an old geezer. 

@aerovons said:

If there are many among you who, given the choice of using a roomful of excellent live players and a sample library, would choose the library, that's super and more power to you. I would not have think about which I would choose for a nanosecond.

 

This is a totally artificial hypothetical situation created to support an argument and is irrelevant to what I am saying.  Almost no one has a roomful of great orchestral players ready to play the second he finishes his  music.  Not even John Williams, so how are you going to ge that?  Answer  that.   

And the same goes for smaller ensembles.  You act as if you can just mosy on down the street and find them (along with the union salaries) ready and eager to play for you.

And again, you are making the same old assumption I have heard here before - every orchestra is just like the London Symphony you hear on CDs and soundtracks, and sample libraries are pathetically trying to copy that.

That is totally false.  99% of orchestras for both concert and film have musical abilities far below that of the few great ones which are incredibly difficult to get to play your music.  That is contrasted with the Vienna players you have at your beck and call within the samples - they are by and large the equivalent musical ability, and are in fact ready to play the instant you turn on your computer.

...............given the choice of using a roomful of (the usual) live players and a sample library, I would not have think about which I would choose for a nanosecond..................    (and in fact I've been given the choice and the samples won out each time)

[tomhartman]

William, thanks for clearing that up. With age comes experience, but also, just...age;)

I'm sorry you seem to have had problems with live players. What I've found down here in Florida (Miami) is that there were jobs I did where the samples were just performed better (some difficult string passages were sloppy with the live players) and there were times I'd have to sneak the samples in the mix at those points to "help" out the live players. Fortunately, those were pretty few and far between, most of the time the biggest problem I have had with live sections is making sure the string players were in tune;)

Still, I have just always found the live session sounded more...real. Well, duh...but there was more air, more transparency, and no amount of artificial reflections with room simulators, etc could make up for it.

So I'm not saying the real players' instruments sound more real, as in most cases the samples have gotten to the point where that is no longer an issue, but there is some magic with live people that isn't there with samples ...in most cases...that I haven't heard with samples. The best mock ups I've heard still sound like great mockups to me. I do agree that I have heard things on TV where even my wife has come into the room when something pretty was playing and said "Are those real strings" and I've had to say "I don't know...". But again, that's without a reference point. If I were able to hear real vs non real of the same passage I'm pretty sure I'd be able to tell.

I recently did a pop piece where there is a section where all the pop stuff stops and an orchestral section comes in, some strings, horns, and flutes. I mocked it up and I liked the samples, they sounded fine. But I had a chance to use real guys as they were already doing something else for me so I had them do it. Maybe it would be interesting to post the results. I think it's good example of samples vs real sounding DIFFERENT, but neither one sounding particularly BETTER....

Tom

[jeremyroberts]

@William said:

99% of orchestras for both concert and film have musical abilities far below that of the few great ones

As a New Yorker, I don't understand this comment -- it's pretty hard NOT to assemble a group of pro musicians from within the NYC community that isn't superb. If a player is actively working in NYC, I can count on a great section. Unfortunately, 99% of our recording studios have closed. We only have one big room left.

And as a touring conductor, I have had the pleasure of working with pops orchestras in: Boston, Washington, Atlanta, Houston, Dallas, Cleveland, Birmingham, Salt Lake City, Miami, Denver, and about a dozen other large cities in the US with SUPERB orchestras. There's no lack of talent out there. So be careful -- 99% of the orchestras are NOT sub-par. maybe 5% -- but not anywhere near your 99% number.

[William]

I guess I should specify exactly what I mean by that.   I know there is a lot of talent, but probably what I am thinking about is how even getting a good orchestra - not the greatest but a good one - is first of all very difficult for a composer and involves a lot of buttkissing or schmoozing, or being in the right place at the right time or whatever - and also, I am comparing the average to good orchestral/symphonic band/chamber ensemble/recording session players that I have had on my own pieces with the Vienna Symphonic Library specifically.  I was blown away by the recent performance I got out of the VSL sounds on pieces that were played by a fairly good college orchestra, a ballet orchestra, a very good university symphonic band and a nationally known chamber orchestra.  They are talented as you put it, but  I am being influenced in my own attempts at perfectionism WHICH ARE ENCOURAGED BY SAMPLES FAR MORE THAN LIVE PERFORMANCES because you can tweak until you die.   Even a group of talented musicians - not the Berlin Philharmonic, no, but just talented - cannot do as well as VSL samples. And I am NOT talking about sample libraries in general also - I mean only VSL because nothing else is remotely comparable with the full orchestral sound.   The choirs out there and a couple string libraries now are getting close, but not the full orchestral sound. 

[William]

 Also, to specify it a bit more - you have your great New York musicians all assembled and eager to record, your world-class studio all powered up, and all the money for the union salaries and the engineers.  O.K.   Now, you decide to change the exact articulation of a phrase in the middle of your composition. Because, it is truly evolving and getting better.  Are you going to do that?  What about doing it five times in a row?  Ten? 

So that aspect - being able to perfect not just a composition, but a PERFORMANCE is what lies behind my championing of samples.  I have often compared serious music done with samples - not crappy cheapskate substitutes for film orchestras but serious music performance by samples (and I know there isn't much of it)  to painting.  A painter doesn't write out specifications for his colors, he paints them himself.  With a sufficiently developed sample library - and VSL is the one - one doesn't have to stop with merely specifying sounds as with a score  but can paint the actual performances himself.  That is a revolutionary development which  is unique in musical history.  It has always been attempted - for example, the piano forte and the pipe organ are both attempts at allowing one person to control a huge amount and variety of sound - but it has never been fully achieved until now.

[jeremyroberts]

Ah... this is where we will never agree.

I use samples. Extensively.

I orchestrate and conduct and work with live orchestras. Extensively.

Music to me is about emotion. NOBODY, i repeat, NOBODY gives a crap about any single articulation.

Part of the magic is that sometimes, collaborators will all inspire and influence each other. If you view your orchestra as robots trained to perform on cue exactly your vision of perfection, you will be one very disappointed composer. And a composer without musician friends who don't look forward to playing your work.

I use VSL as much as anyone -- at the highest levels, but never, EVER discount the importance of the hybrid and the collaboration with the real guys. Never.

We're only making music. It's just not that important. Sure, I get notes from fans, "your music helped me through a difficult time" or "we played your music at our wedding and it brought us joy" -- that's a fringe benefit of what we do... but it's never so important that we miss the bigger picture: "does it feel good?"

As a non-film maker, I would never say, "but the DP should have used a different lens for the wide shot... it ruined the film" -- I may leave the film unsatisfied, but the viewer never would call out any technical moment. Unless you're a pro. NOBODY will ever say, "but the viola section didn't articulate the detache's with the snap that VSL library would give it..." that's total bullshit if you think that ANYONE anywhere is actually listening to your music that way. Trust me, nobody cares. Do they like the song? Does it make you feel something? Does it score the scene and make an emotional impact? Everything else is just stuff you do for your own entertainment.

If you truly think that you can make better music with machines -- great. I use the machines as a creative choice and usually financial issue -- machines vs. $30,000 session?  Then there are the hybrid scores...  so we do whatever services the music. But maybe it's my age, but I no longer believe that anything I do is important. Yeah, people enjoy it. And they pay me. And I get pleasure doing it. And I get pleasure knowing others enjoy it. But if not me, it will be someone else. I am not the only guy who can make a song sound good. So I just do my thing... speaking of, I have 170 bars of orchestration I need to deliver by tomorrow.