How do you fake a trill and make it sound real?

[Shooshie]

No, I certainly don't mean a legato trill on a synth, though some patches could conceivably be programmed for that. Percussive sounds work, such as piano or marimba.

More later if I think of something.

I've been without internet for a week. A little overwhelmed right now.

Shoosh

[FMiguelez]

Glad to hear I won't have to give you my VSL Super Package then, Shoosh! Phew!!!!

So then, would you be comfortable with the following?
- My other post stands as written if we're talking about regular hardware synths
- Your post stands as written if it's something similar to Wallander (modeled) stuff

I know nothing about modeling at all, so I could not comment on any of it.

[Shooshie]

Glad to hear I won't have to give you my VSL Super Package then, Shoosh! Phew!!!!

So then, would you be comfortable with the following?
- My other post stands as written if we're talking about regular hardware synths
- Your post stands as written if it's something similar to Wallander (modeled) stuff

I know nothing about modeling at all, so I could not comment on any of it.

It's not just either/or of the above. It's my libraries, as I set them up. There are so many things that work when you set them up right, but then there are things that don't work, too.

You're way over-thinking this. Or else, way over-skeptical. Remember, I've been doing exactly this for close to 30 years, and I was a concert performer before that. (and didn't stop being a performer when I started doing MIDI)

Wallander Instruments, however, are often very good at this sort of thing. Key noise and all. Sometimes, they inexplicably don't sound as clean as other times. After a few alterations they usually come around, but it's trial & error type stuff.

The real education was trying to get old synths to do this kind of thing. I thought I'd gotten very close, back then, but now when i hear those old recordings I cringe! It's all relative.

Shooshie

[dosuna11]

Great knowledge. Thanks for sharing. Is this in the tips section?

[Shooshie]

Great knowledge. Thanks for sharing. Is this in the tips section?

Probably not. I'm trying to think if there are any tips or posts in that thread which contain this kind of info. I really should add something. I just have to figure out what exactly to add, and where to put it.

I still have a backlog of tips to add "one of these days." Takes a lot of time to get them all in the tips-sheet format, edited for clarity. Still, I absolutely do plan to do that... one of these days!

Shooshie

[kwiz]

Thanks for sharing, Shooshie!

[FMiguelez]

You're way over-thinking this. Or else, way over-skeptical. Remember, I've been doing exactly this for close to 30 years, and I was a concert performer before that. (and didn't stop being a performer when I started doing MIDI)

I am totally skeptical about being able to make a hardware synth play realistic trills (and tremolos and scales and arpeggios).

I don't mean to sound negative or unappreciative of your posts and all the knowledge that abounds in them. I've learnt so much from you throughout the years and I have an enormous amount of respect for you. But in this particular case I just know that no amount of tweaking would fool me into thinking I'm listening to a flutist or trumpet player. I know what they are supposed to sound like because I've been listening to them since I'm 2 years old.

I tried for years and it always sucked. When I thought it wasn't that bad, is because I most likely fooled myself (or lowered my standards or the tolerable threshold) into thinking: Well, at least the music is nice. But if I listen to those old productions now, I just smile and... Well. Let's just say technology has allowed us to go into a whole new ball game.

I used the techniques you've described. They are great techniques that will certainly improve things, but it won't fool anyone into thinking it's the real thing, like the OP wants.

The OP will discover this for himself, I'm sure (with the VI library he mentioned). One can spend a full night working on a tremolo or trill or scale, but it will not sound "real" because the necessary technology simply wasn't there for those synths to begin with (his library is different and newer, of course, but the point is that even VSL would suck if one tries to make a trill with a non-trill patch, such as sustained/short/medium. Ok. Maybe the performance legato patches could work, but that's because of the special programming of the releases and different samples/layers per note).

Now, I'm sure you could make a convincing trill with Wallander (or similar) instruments and your breath controller, but like I said, I can't comment on that because I don't know anything about that particular kind of VI.


I might have been a little brusque in my first post. I just wanted to let Mike know my opinion on the matter. Especially because there's not always too much time to spend editing and mangling with MIDI for these "little musical details" when there are impossible deadlines to beat. At that point the priorities (and time one can spend on detail) become totally different.

[Shooshie]

I am totally skeptical about being able to make a hardware synth play realistic trills (and tremolos and scales and arpeggios).

I don't mean to sound negative or unappreciative of your posts and all the knowledge that abounds in them. I've learnt so much from you throughout the years and I have an enormous amount of respect for you. But in this particular case I just know that no amount of tweaking would fool me into thinking I'm listening to a flutist or trumpet player. I know what they are supposed to sound like because I've been listening to them since I'm 2 years old.

[...]

I might have been a little brusque in my first post. I just wanted to let Mike know my opinion on the matter. Especially because there's not always too much time to spend editing and mangling with MIDI for these "little musical details" when there are impossible deadlines to beat. At that point the priorities (and time one can spend on detail) become totally different.

He asked for a way to make trills without playing them. He didn't ask for something that can fool a panel of highly attuned critics. You guys immediately jumped to "it's impossible." I said "here's a quick way of doing it."

Again, you're WAAAYYYYYY overthinking this.

What you're really debating is whether mock-ups in general can fool a trained ear. Not just trills, but any kind of mock-up. The answer, I've found, is two-fold:

1) if you tell them it's a mockup, they will always say they were not fooled.
2) if you tell them it's just a recording, they'll listen and keep that opinion to themselves, whether or not they are suspicious. Most of the time, I find that they think it was real.

As for trills, let's talk about that. I've practiced trills even longer than you, assuming our age differences, and I'm well aware of the differences between playing a trill in Bach, Telemann, Mozart, Beethoven, Brahms, Stravinsky, Ravel and Prokofiev. Even Phillip Glass has a trill style, based on his proper voice leadings and his mathematical divisions of rhythms, but he'd probably be the last to admit that there is one, or to ask someone to play it that way. He's a very humble guy! (and yet very exacting) Each uses trills slightly differently in their musical language, for different purposes, and often those purposes change from one trill to the next, depending on the context.

Maybe you wouldn't like my trills, but I'm perfectly capable of forming them in MIDI the way I play them, and the technique in the video is the first step toward that. (The video was created a couple of years ago for tool demonstration, not how to create a finished, perfect trill) But let's talk about trills themselves. They are very important in music, but they are used for completely different reasons by different composers and in context of what is being said or the effect being created.

I studied trills by slowing them down. Great artists play trills by ear and feel, not by formula (but within the contexts such as those I laid out above). The best of them are uneven. Ravel even WRITES them unevenly to fill space while trying to avoid a pattern that bores the ear. Example: the opening of Ondine, from Gaspard de la Nuit. It looks like a pattern, not a trill, but functionally it's an expanded trill that attempts to avoid repetition in the ear. Marta Argerich does that better than most people, because she gets that. Many others try to play it precisely, and it comes out too much like a mixed-meter pattern. When Marta plays it, she fudges the values and velocities, getting a wonderfully suspenseful sustain without the feeling of any repetition. The humanization feature actually works pretty well in this if you go back and fix certain notes by ear.

Other trills are about perfect repetition: Mozart, for example. Then there's the baroque. In a 1957 (give or take) Columbia recording of Philadelphia Orchestra "First Chairs," principle trumpet Samuel Krauss plays Jeremy Clark's Trumpet Voluntary's trills with about 8 dyads in under a second. I used to know the exact number, because I slowed my record down to 16 rpm and counted them, but that was in the 1960s, and I've forgotten. But that was my first introduction to playing trills on a wind instrument, and was a model I sought to imitate for many years until I learned that all trills are different. A trill at that speed ceases to be a series of notes, and becomes a sound, something distinctive for royal ears. Those trills are very easy to emulate in MIDI, by the way.

Every trill is different. Learning how to express them is very similar to learning how to enunciate Shakespeare or Tennessee Williams. Nobody can "teach" you that. You have to read the context and BE the voice, saying it with emotion (or lack thereof, if required). Trills, like every other aspect of music, are like that, too.

To emulate a trill in MIDI, you have to start out with a trill. It has to be non-mechanical, and your instrumental voice has to respond to quick notes. Violin trills are more difficult, and I prefer using a sampled trill if one is available, but I've got nothing against making my own if I don't have time to find one, or if one doesn't exist in that library. Sampled trills have a sweeter sound, in part because violinists play trills differently than wind players or keyboard players. They can actually vary the pitch of the upper note during the trill to compress and expand it for greater emotion. Listen to the big three romantic violin concertos for examples: Beethoven, Mendelssohn, and Brahms. But we can actually do some of that ourselves by selecting the upper note, then altering the pitch in the Pitch Bend section, using proportional drag to change them across the board, or using % reshape and a spline tool. It's a little more work, but doesn't take long. (haven't done that in a while, so I may have forgotten some details)

Anyway, the big question remains "can MIDI mockups actually sound like the real thing?" Each year the answer becomes closer and closer to "Yes!" But the smaller question, about trills, is simple: within the confines of MIDI Mockups and their ability to fool the listener's ear, the trill itself requires some skill, but is just as convincing as any other part of the genre.

There is no question about this. Again, if I have sampled trills, then it's all the better, but when you don't, and when you can't play them in for one reason or another, it's very possible to cut up some long notes and create a trill that is very convincing. It's part of the MIDI craft.

Shooshie

[MIDI Life Crisis]

Totally agree Shoosh. Frankly it's kind of a no brainer in reality. Lots depends on the responsiveness of the VI or MIDI instrument but the bottom line is is basically two notes (for a simple, single trill) played very fast. I don't really see what the problem is - aside from bring less experienced and/or creative in the use of MIDI. No offense intended to anyone, but it's really not that big a deal unless you decide to talk it to death instead of just experimenting and mastering it as best you can.

[FMiguelez]

My dear friend Shooshie, I think we agree on more things than we disagree. Our disagreement is in the details and even in some of the "wording"...

First off, let me tell you I won't be caught red handed skimming your (or anyone else's) posts. I read every word you wrote, and I enjoyed your writing.

He asked for a way to make trills without playing them. He didn't ask for something that can fool a panel of highly attuned critics.
[...snip...]
Again, you're WAAAYYYYYY overthinking this.

I don't think I am. The title of the thread is "How do you fake a trill and make it sound real?"
That, to me, means he wants to fool an (unspecified) audience into thinking his trills (or MIDI sequence) were performed by a real player. Maybe not a panel of judges, but his intent seems quite clear to me.

You guys immediately jumped to "it's impossible." I said "here's a quick way of doing it.

Yes, but that sounds like what you propose is as effective as a sample or the real thing, and it isn't (yet).
If you had said "here's a quick way of doing it if you don't have the samples, but it takes longer and it won't sound as real", then we wouldn't be having this conversation.

I mean, like they say, the proof is in the pudding: Show me an exposed trill (or even a full chamber quintet mockup) that doesn't use samples that fools me into thinking I'm listening to real players and I will shut up and concede all my points.

Every trill is different. Learning how to express them is very similar to learning how to enunciate Shakespeare or Tennessee Williams. Nobody can "teach" you that. You have to read the context and BE the voice, saying it with emotion (or lack thereof, if required). Trills, like every other aspect of music, are like that, too.

Agreed. Trills have changed over time with style. But now you're talking about performance, which is fine, but that would be getting a little off-topic, because I insist, no matter how much one knows about trills and how to play them according to style, etc., the problem remains: they still suck when done via MIDI with non-trill samples.

There is no question about this. Again, if I have sampled trills, then it's all the better, but when you don't, and when you can't play them in for one reason or another, it's very possible to cut up some long notes and create a trill that is very convincing. It's part of the MIDI craft.

Again. We just don't agree on the details. I wouldn't say "very convincing" (they won't fool me or anyone intimately familiar with orchestral instruments). I would say "you can achieve a result that might get you by, but it will not (Ok-maybe in a few years) sound like the real thing".

That's why I insist that Mike should adjust his expectations accordingly. He will only get so far faking trills. He might do a great job, but he would only "fool" people with no music training. He will discover this for himself, I suppose.

We're generally very very close to the real thing (with samples). But that's why samples were invented... This is what drove Paul and the VSL gang to develop their product, the frustration that came with getting only "meh results".

So I suppose the bottomline is:
Yes. One can try to fake trills, tremolos, etc., with MIDI tools.
Will they sound convincing? Well, that depends on the skill level of the artist, of course. Trained musicians will spot the difference from 3 miles away. Lay people would probably be convinced.
Given the choice to use samples or doing them yourself, and assuming the sample works in the context of the piece, it would be a no brainer anyway, would it not?

[MIDI Life Crisis]

If the musician can play the trill, and the MIDI software can record it, and the VI can reproduce it accurately, I don't see what the problem is. It's either the player, the DAW, the computer, or the playback device that is creating the problem. Just figure out which it is and fix it. Done. No?

[FMiguelez]

If the musician can play the trill, and the MIDI software can record it, and the VI can reproduce it accurately, I don't see what the problem is. It's either the player, the DAW, the computer, or the playback device that is creating the problem. Just figure out which it is and fix it. Done. No?

I tried that for years with my synths and it just never worked.

I got good results, but nothing that would replace the London Symphony Orchestra.

The problem is that all those intricate interactions of fingers, keys or strings, intonation, legato, subtle portamento are not there with those synths. One can program a trill accurately in terms of note placement, velocities, CC ramps, etc. But there are many other factors that if they aren't there they can't be faked.

I mean, come on, Mike. You know this. You've been doing for longer than I have. Play an oboe trill or tremolo on your Korg and tell me what it sounds like...

[MIDI Life Crisis]

So what sounds worse? Breaking the constancy of the sound with a recorded trill or having the crappy MIDI sound? If I need a convincing performance of a difficult instrument I hire a player to run the part down. If you add up all the time you spend chasing your tail on something like that, it is more cost and time effective to hire a good player. At least that's what I think. But as far as faking, I can fake most trills and other stuff "well enough" for mockups and even in some performances for broadcast. It depends on a number of factors (instrument, style, tempo, range, et al).

No, it's not the London, but frankly what virtual setup is? Even in high end feature films with gigantic budgets and fees (but still using VIs for realization - duh!) you can hear the crap when it's played back. It's my opinion, but I maintain that electronic instruments just cannot compete in terms of expression and execution to a talented player on a acoustic instrument.

[FMiguelez]

It's my opinion, but I maintain that electronic instruments just cannot compete in terms of expression and execution to a talented player on a acoustic instrument.

Oh, I TOTALLY agree with you there! Not any time soon anyway, if ever!

But samples, being recordings of real players, are as good or as close to real as we can get at the moment.
I think the main challenge for VI companies has been to create software that makes those individually recorded notes NOT sound like individually recorded notes.
That ability is one of VSL's high points with their legato and performance patches, and they can be made to sound almost real. So much that I would have to listen veeeeeery carefully to tell them apart from a live performance (especially if the engineering aspect is great as well).

Needless to say, it all depends on the musician's ability and skill to program them effectively. But the tools are there.

[Shooshie]

I thought he just needed help with getting the trill started. Everyone knows that a performance is only as good as you make it.

Again, the video was meant to show the tools that will help do trills quickly. The finishing of a trill (not part of the video) is another matter, but not difficult to do.

I'd rather play in a trill any day. Sometimes the instruments you're using just don't permit that. They don't respond fast enough.

Speaking of Wallander Instruments, which Fernando has mentioned a few times, I happen to have a recording of literally the first time I ever blew a note on Wallander Instruments. In fact, it was the Wallander Demo, on which most of the controls were deactivated, and not even all the notes were present. But I picked up the trumpet and decided to try Jeremy Clarke's Trumpet Voluntary (often misattributed to Henry Purcell). It was just by ear; I didn't have music in front of me, so it starts out a little unsure. I expected the trills to feel delayed, so I did them slowly at first.

You can hear me getting more confident as the song went on. Remember... this is the FIRST time I ever blew on it. I just hit record first, for whatever reason. I did not go back and adjust anything (or it would have sounded a LOT better) I was playing too loud, but the controls to adjust all that were locked out on the demo, so this was before I even knew what its dynamics were. By the end, I was playing pretty much full-speed trills.

I post it because I want you to see that Wallander responded as if it WANTED to follow me, and I was playing it as if I were expecting something like MOTU Symphonic Instrument until I saw that it was actually going to keep up with me.

So... here's Wallander Instruments, 2007, first time I ever blew a note:
Trumpet Voluntary (by ear)

The next thing I tried was the horn. This time I got out a piece of music by Hindemith. Again I was playing with too much force, but this was literally the 2nd thing I played on Wallander. I'd learn to refine it later. This was just to see if I could make music with these instruments. Up to this point I had never had ANY sampled instrument play like a real instrument. I had synths that played very expressively, but they lacked a convincing sound for doing real instruments. What I heard when I began playing this piece (and the preceding one) excited me very much.

Again, remember that the controls were locked down in this demo. I could only play it with default settings, and I needed to make it less responsive, but the point comes across, I think.
Hindemith: Waldhorn Sonata, mvt. IV, Wallander Instruments horn

I immediately bought it.

Shoosh