orchestral samples -- worth the while?

[condor]

I'd like to get into scoring orchestral music using orchestral samples on a Gigasampler system. I've read glowing reviews of some libraries, but I'd like to know from members of this forum who use orchestral samples:

 

Can I produce completely first-rate orchestral music using samples? I know how to write for orchestral instruments (15 years experience), and have a good knowlege of sequencing/programming as well as mixing in DAWs (ProTools, DP, etc.).

 

How about articulations, dynamics, phrasing, etc.? I know these questions are broad, but before I jump in, I'd like to know the limitations.

 

Thanks!

[Guest]

There was an excellent feature on keyboard mag about it, maybe on the december 00 issue? I believe there's a book called midi orchestration, which is probably worth checking out. Check out:

 

http://www.keyboardonline.com/features/jnhoward/index.shtml

http://www.keyboardonline.com/features/rupertgw/index.shtml

http://www.keyboardonline.com/features/orchsamplesbg/index.shtml

 

 

Can I produce completely first-rate orchestral music using samples?

 

I suggest check out music scores from guys who use this approach, like James Newton Howard, and Hans Zimmer. Let your ears decide if it's convincing or not.

 

Since you already know how to write for orchestral music, the thing you will need to learn is how to use continuous controller in midi to vary the sound in real time to give it the proper dynamic and articulation.

 

Rod

[soapbox]

Originally posted by condor:

Can I produce completely first-rate orchestral music using samples?

 

Ive done a fair amount of this sort of thing and I would have to say that occasionally you can. However, this is only if youre willing to let the high quality samples you have steer your writing so that you use the most believable samples and avoid the others and if youre willing to spend a great deal of time mixing samples of different articulations and dynamics and if youre good with effects and EQ. You will then occasionally be able to produce a believable orchestral sound that stands on its own using nothing but samples.

 

BTW, there are good sounds to be found in the Peter Seilaczek Advanced Orchestra collection, the Miroslav Vitous library, Quantum Leap Brass, and The Ultimate Strings Collection. Ilio also adapted some of the Synclavier collection of percussion and string samples. Theres a lot of other worthy stuff out there too. Unfortunately I dont have my entire collection at hand, but this will get you started if you decide to follow this path.

 

Rod (rsaboiasilva) provided some good links. You might also want to check out an SOS article on this subject.

 

In addition, Rod may have been talking about " The Guide to MIDI Orchestration ," by Paul Gilreath. I own a copy but have yet to find the time to read it. Maybe if I quit posting so much... http://www.musicplayer.com/ubb/wink.gif

 

This message has been edited by soapbox on 04-23-2001 at 10:38 PM

[Ethan Winer]

condor,

 

> I'd like to get into scoring orchestral music using orchestral samples on a Gigasampler system. I've read glowing reviews of some libraries, but I'd like to know from members of this forum who use orchestral samples <

 

I write and record a lot of classical music, and I use both samples and real players. I can tell you for certain that real players - if they can play in time and in tune http://www.musicplayer.com/ubb/smile.gif - are always better than samples. (And I have some pretty good samples!)

 

> Can I produce completely first-rate orchestral music using samples? <

 

So yes, you can produce first-rate music using samples, but probably not a first-rate recording.

 

> How about articulations, dynamics, phrasing, etc.? <

 

And there is exactly the problem. Suppose you want the violin section to play a legato half note followed by two staccato eight notes and then another half note, etc. You might be able to find one sample set that can handle both. But as often as not, the samples that have a sharp enough attack to handle the staccato notes will be too marcato for the sustained notes. That is, every note begins with a down bow. So now you need multiple sets of samples loaded, and constantly switch between them with MIDI program changes or use separate tracks. All the while hoping the sound quality and ambience match so the parts sound like the same section playing in the same room...

 

I just finished a new mix of my cello concerto. If you're interested it's at:

 

www.ethanwiner.com/concerto.html

 

Most of the instruments are real but some are sampled. I mention my concerto page because there's also an article I wrote for Strings magazine that explains in more depth how I made the recording using both real and sampled players. And I didn't have to pay an arm and a leg to do it.

 

--Ethan

[jnorman]

i have been chasing this particular problem for the last 6 years, and my take on it is - no, you cannot make convincing orchestral music using sampled instruments. my son is a composition major at USC, and after blowing through literally every available piece of gear and libraries of samples, he just wont try to use any of it any more. in fact, he has gotten to the point where he hardly even uses his gear to listen to drafts of his work. there are many professional applications where you can get away with using samples/synths, such as theater work to augment specific live instrumentation. but for actual film scoring work, you need to use real players if you want it to sound right, or at least a preponderance of live players with minimal synth support. it can be extremely difficult and expensive to hire and record large orchestral groups, and many new film scoring composers, such as steve gutheinz (check some his work on mp3.com), are compensating by using more minimalist arrangments. my son is using this same approach for his senior recital because of this.

[condor]

Folks, thanks for the replies.

 

I can see that some of my suspicions are valid. I haven't auditioned many samples sets yet, but I think I'll find that their usefulness is pretty limited.

 

In my experience, getting an ensemble to play the right notes is only half the battle. There is an infinite range of tempo variations, articulation, phrasing, and tonal color within a performance that helps drive the music forward, and I don't really see how I could recreate this using samples.

 

It makes recording a lot simpler, too. If the players are 'on,' and I do a good job communicating my intent for the piece (via proper notation), I can set all faders to 'zero', push 'record' on the DAW/deck, and have very little to do at mixdown. Infact, even EQ is almost unnecessary when the arranging/orchestration is attentive to the change in tonal color throughout the range of each instrument at particular dynamic levels, and how those tonals colors blend to produce a particular sound.

 

So, I guess the next question is: if I get into sample-based composition, even in conjunction with live players, what is the limit? At what point do I tell myself, "ok, this is all you can do within this time frame and with this budget?" I've always had to work within a budget, but I can see already that the few sentences I tell live players before a reading will translate into hours and hours of futzing around with velocity and automation curves, and to what end?

[gabrielk]

Originally posted by condor:

So, I guess the next question is: if I get into sample-based composition, even in conjunction with live players, what is the limit? At what point do I tell myself, "ok, this is all you can do within this time frame and with this budget?" I've always had to work within a budget, but I can see already that the few sentences I tell live players before a reading will translate into hours and hours of futzing around with velocity and automation curves, and to what end?

 

I think this largely depends on your audience, your client, and the type of music. (Big "duh's" on all three of those, right?) But honestly--if a honed classical listener will be listening to your music on its own, they'll appreciate mediocre amateur musicians better than they'll appreciate a few pretty decent musicians playing over sampled sequences (read: they may rag on it, but they won't lose all respect). Folks who listen and play this type of music have much better ears on their heads than most amateur/semi-pro engineers! =) ;P

 

Now, if it's going to be, say, in the background of a car commercial, or theme music for a boy scout troupe, your sound will be much more appreciated if it's filled out with the sampled music.

 

And I believe that "how much futzing around is too much," in regards to getting the samples believable, etc., should be largely dependent on either a) your devotion to the project (is it worth staying up 22 hours a day for two weeks to meet that deadline and get it perfect?), b) how much you're getting paid in relation to the work (are you getting the big bucks to do it *right*, are you doing it for yourself, or...?), or c) a combination of a & b?

 

Hope my post wasn't completely useless to ya,

--Gabriel

[dansouth]

Originally posted by condor:

Can I produce completely first-rate orchestral music using samples?

 

No. It will never sound like a real orchestra. Never. No matter how much you pay or what you do. Get that in your head at first, or you'll be disappointed. That said, there's a lot you can do to bring out the musicality of orchestral samples, especially if you're willing to put time into (a) selecting sounds that work well together, and (b) painstakingly editing the timing, velocity, and duration of every note.

 

For an example of what can be done with far less quality than a Gigasampler, have a listen to the trumpet concerto on my site. It was programmed in logic and played "live" by a single Roland JV-1080 with the orchestral expansion board. Most of the programs were "stock" with a couple of envelope parameters modified for more distinct attacks.

 

http://www.mp3.com/DanSouth

 

I have yet to find orchestral samples that I really like, and I've dropped a couple of grand already. If they successfully port Miroslav Vitous samples to Giga, I might try that. Virtual Orchestra and Virtuoso Strings are a waste, as far as I'm concerned. I'd love to have my 2000 bucks back. I prefer the little 8 Meg Roland board, because the samples blend together very nicely.

[dansouth]

Originally posted by rsaboiasilva@hotmail.com:

I suggest check out music scores from guys who use this approach, like James Newton Howard, and Hans Zimmer. Let your ears decide if it's convincing or not.

 

Keep in mind that Zimmer and company sampled the London Symphony themselves. They do not use off the shelf sample libraries.

[condor]

Ethan and Dan,

 

Thanks to both of you for providing links to your work. Unfortunately, I am a true luddite, in that I'm just learning the 'complexities' of basic email and web browsing, so I haven't been able to hear your work. The points in your posts here, however, have been very useful to me.

 

So, is it too early in my 'investigation' to conclude that samplers et al are best used for sounds that are _not_ supposed to sound like acoustic instruments? If so, it seems like a marriage of the two would produce more musically satisfying results. The cue music for 'The X Files,' for example, uses a lot of this (synthy pads, hits, etc. coupled with those wonderfully eerie lines played by a bass clarinet). Seems like this approach works very well, and though it would be quite a learning curve for me to become proficient making my own sounds with samplers and synths, I guess I'd be avoiding the 'cheeze' factor, wouldn't I?

[Sergievsky]

You really can't conclude anything. You need to dive into this yourself and find out yourself. Everybody works different ways and have different budgets and different skills and patience and time futzing with the sequence, which is critical. A friend of mine who sat in on a tracking session was asking me where I got my string samples. I told him it's the exact same one he has. I just messed with the the sequence and volume cc more and tracked it through a Sony R7 reverb that had a really nice program simulating a really nice orchestral hall.

 

And like any library, some samples are nice and some suck, so you almost have to own them all and pick and choose, and some that sucked actually sounded better on some mixes than some that sounded great on its own. I have 4 different samplers including Giga and the samples sound different in all of them, and though they all have Akai conversion capabilities, they all end up sounding different.

 

If you're planning on reproducing Beethoven's Fifth, don't even bother. But if you have a budget for some live musicians, well, a live french horn player playing the same passage along with the sequenced sampled French horns will derinitely sound more real.

 

Anyway, good luck if you do dive into this. Hope you swim really well. http://www.musicplayer.com/ubb/smile.gif

 

Raul

[Curve Dominant]

Hey, folks...

 

I have some experience with this, so I'm tempted to chime in. But I would like to preface by saying that I'm equally tempted to keep my mouth shut. That having been said...

 

I have to disagree that it cannot be done. One can create "convincing" orchestral recordings using purely sampled sounds (take note of the use of the word "convincing"). But it is not cheap or easy. I composed two 15 minute long "orchestral" pieces for a local modern dance troupe, and the dance reviewer for the Philadelphia Inquirer asked me "Where did you study classical music?" and "Who conducted the piece?" in an interview after one of the performances. I had to laugh, because I'm a self-taught rock guitarist, and the piece was recorded in a MIDI studio using sampled sounds. But here's what we used:

 

Roland A-90 keyboard: 88 weighted keys, touch-sensitive and after-touch capable. The parts were all played on this keyboard, violins, violas, cellos, flute, piano, horns, etc.; and effects such as marcato, staccato, legato, etc. were emulated by experimenting with the capabilities and limits of the keyboard. None of the tracks were done in one take, I can assure you. It was painstaking, time-consuming work.

Digital Performer on a Power Mac 9600 with all the upgrades available (the owner of the studio was a computer consultant).

Roland JV2080 sound module

Emu E6400 sampler with all the expansion cards the darn thing would hold

a truck load of outboard gear, MOTU interface, Roland electronic drumset, digital mixer, pictures of naked women, etc. etc.

 

Then, you have to bear in mind that the mixed recording was professionally mastered, and played LOUD in a fairly large theater. We took this into account when applying effects: for example, we only applied stage reverb to the tracks, but never any room or hall reverbs, because these would be created in the theater.

 

So, how convincing was it? Well, that depended on the individual audience member.

 

The average Jane and Joe Schmoe, who are not "in the biz," but rather were just out for a night of stimulating entertainment, didn't think for a minute that it was anything but an orchestra.

 

A more interesting question is: what would Ethan, or J Norman's son, have thought? I can assure you that they would have known within moments that it was not really an orchestra...but would they have cared? Would it have bothered them? That would have depended on how much they were enjoying the dance performance - which is really what they would have came for.

 

Here's my point: There is only so close you can come to the real thing, and then there is how well it works with what it's being presented in conjunction with. There's an apex where, if it's reached from both angles, a kind of "suspension of disbelief" takes effect, where even Ethan and J's son will say, "Screw it, I'm enjoying this." And THAT, my friends, is achievable.

 

It also helps if you are a brilliant composer. But that's why, if y'all are in Philly, ya call on me. http://www.musicplayer.com/ubb/smile.gif

 

I welcome feedback or questions regarding this.

curvedominant

[Sergievsky]

What was it that Jeff Rona said on one of his simulated orchestral scoring session articles...something like, "it sounds like a real orchestra, that is, until you hear a real orchestra" http://www.musicplayer.com/ubb/smile.gif

 

Raul

[Curve Dominant]

Raul: LMAO! That sounds like Jeff. I like his articles.

 

curvedominant

[gabrielk]

I think ya just said what I was trying to say, Curve, but a little clearer, and a little more precise. Of course, you've also had more experience with it than I have had.

 

Hey, are you really that brilliant?

 

--Gabriel

[dansouth]

Originally posted by Solfatio:

Hey, are you really that brilliant?

 

Yup, he sure is! http://www.musicplayer.com/ubb/smile.gif

[Lee Flier]

Originally posted by Sergievsky:

What was it that Jeff Rona said on one of his simulated orchestral scoring session articles...something like, "it sounds like a real orchestra, that is, until you hear a real orchestra" http://www.musicplayer.com/ubb/smile.gif

 

Raul

 

ROFL... boy he got that right! Now if only people felt the same way about drums... and keyboards... and guitar amps... and... oh nevermind... http://www.musicplayer.com/ubb/biggrin.gif

 

--Lee

[Ethan Winer]

condor,

 

> if I get into sample-based composition, even in conjunction with live players, what is the limit? At what point do I tell myself, "ok, this is all you can do within this time frame and with this budget?" <

 

You'll find out soon enough! http://www.musicplayer.com/ubb/smile.gif I'll venture that whatever you can get done in 10-20 hours (not counting the time to write the music) will show you how far you can go.

 

--Ethan

[Ethan Winer]

Gabriel,

 

> if a honed classical listener will be listening to your music on its own, they'll appreciate mediocre amateur musicians better than they'll appreciate a few pretty decent musicians playing over sampled sequences <

 

I'm not sure I agree. I have concert recordings of amateurs playing my pieces, and the biggest problems are passages out of tune, "cacked" brass and woodwind notes, and lame entrances. With samples and overdubs of decent players - amateur or pro - at least the notes will be correct!

 

--Ethan

[Ethan Winer]

condor,

 

> Unfortunately, I am a true luddite, in that I'm just learning the 'complexities' of basic email and web browsing, so I haven't been able to hear your work.<

 

Gee, all you have to do is click the blue underlined links. http://www.musicplayer.com/ubb/smile.gif When you get to my Concerto page click the button "Download Concerto MP3." If you have a current computer and a high-speed Internet connection it will play. With a slow connection it will sputter and try to play, but after the whole file has been downloaded (maybe 20 minutes) you can click play to hear it.

 

> So, is it too early in my 'investigation' to conclude that samplers et al are best used for sounds that are _not_ supposed to sound like acoustic instruments? <

 

Not necessarily. You got a lot of good advice from others here about this too. Some samples are great and some are terrible. Regardless of how much they cost. And some orchestra instruments can be sample-played more successfully than others. Strings are always a problem because so much can happen after the note has started. Same for woodwinds and brass. But timpani or piano samples, for example, can theoretically sound as good as the real thing since you're playing back a real recording of a real instrument.

 

--Ethan

[Ethan Winer]

Curve,

 

> There's an apex where, if it's reached from both angles, a kind of "suspension of disbelief" takes effect, where even Ethan and J's son will say, "Screw it, I'm enjoying this." And THAT, my friends, is achievable. <

 

Absolutely. And since Condor's initial goal is (I think) to write and hear the music, not necessarily release a CD on Sony Classical, there's no reason samples cannot be very satisfying.

 

--Ethan

[dansouth]

Originally posted by Ethan Winer:

> if I get into sample-based composition, even in conjunction with live players, what is the limit? At what point do I tell myself, "ok, this is all you can do within this time frame and with this budget?" <

 

You'll find out soon enough! http://www.musicplayer.com/ubb/smile.gif I'll venture that whatever you can get done in 10-20 hours (not counting the time to write the music) will show you how far you can go.

 

--Ethan

 

Buy the score to a popular piece of music - The Four Seasons, The Ninth Symphony, The Nutcracker, etc. These are all available on Amazon, etc. Pick a segment of one of the movements and program is note for note. Then listen to a quality recording of the piece. I've got twenty bucks that says that it will be a humbling experience.

 

On the other hand, if the audience listens to your piece without a reference, these differences will not be significant enough to detract from their experience, given that you did a REALLY good job of programming dynamics, phrasing, interpretation, etc.

[Ethan Winer]

Dan,

 

> Buy the score to a popular piece of music ... program it note for note ... I've got twenty bucks that says that it will be a humbling experience. <

 

You don't have to convince me. I've done this many times!

 

--Ethan

[dansouth]

Ethan, I was suggesting it to condor as a follow up to your message regarding the limits of sample based orchestrations. Looking forward to checking out the cello concerto. - Dan

[Ethan Winer]

Dan,

 

> I was suggesting it to condor as a follow up to your message regarding the limits of sample based orchestrations. <

 

Ah, now I see. http://www.musicplayer.com/ubb/smile.gif

 

> Looking forward to checking out the cello concerto. <

 

Let me know what you think. It's still "in progress," but I think I'm pretty close. I have an appointment with a real classical mastering engineer in two weeks (at $125 per hour!), and he's promised to give me his best advice and suggestions. This guy does CDs for some of the biggest labels, and I am very excited to see what I can learn from him.

 

BTW, I listened to your trumpet concerto. Great job! It's a real pleasure to hear "new music" that's not overly disonant or full of chords that never resolve...

 

--Ethan

[dansouth]

Ethan, I had a chance to listen to about two-thirds of your cello concerto today. I chose to listen to your version, first. I'll listen to the other gentleman's version when I have a chance.

 

Some impressions - The recording quality is very good. The instruments are well-balanced. Interestingly, it sounds like a much large ensemble than what your page mentions. You must know some pretty potent orchestration tricks. The cello comes through loud and clear. There's nothing worse than a concerto where you can't hear the soloist - unfortunately, this happened to a friend of mine on the recording of her senior recital (violin). - Your orchestra and conductor are very sensitive to dynamics, thankfully.

 

The music sounds very theatrical to me. During the faster passages, I could almost imagine Indiana Jones racing across the Sahara chased by bad guys on camelback. The slower, more lyrical passages remind me of an aria or a song from a musical, a testament to the strength of your melodies. I like the way the piece moves from aggressive to laid back, back and forth. The segues are very smooth - nice touch.

 

Good luck!!

Dan

[Ethan Winer]

Dan,

 

> I chose to listen to your version, first.

 

I gotta take that one off! A few weeks ago I made a new, and much better, mix of the version with Steven Thomas playing. That is the one to listen to!

 

> Some impressions - The recording quality is very good. The instruments are well-balanced. <

 

Thanks!

 

> Interestingly, it sounds like a much large ensemble than what your page mentions. <

 

I had eight violins, six cellos, etc., but during each session I had the players record the entire concerto twice. I did that to have a backup performance for editing, but except for a few short passages both takes were good enough to use. So it's really 16 violins, 12 cellos, etc.

 

> Your orchestra and conductor are very sensitive to dynamics, thankfully. <

 

My "conductor" was a click track. http://www.musicplayer.com/ubb/smile.gif Though I made CDRs for all of the players with my original synth version, so they had a chance to practice the dynamics and tempo changes ahead of time.

 

Thanks very much for your comments.

 

--Ethan