Several tracks with violin solos as opposed to violin ensembles

[GeneralDisarray]

This may be a dumb idea, but I want to try a violin ensemble made of maybe 20 to 30 individual Logic tracks, each of them with a solo violin. Then I want to set all the tracks to record, or maybe just stack them into a summit track so I don't have to enable R in dozens of tracks.

But I want to humanize the performance, rather than play all of them as they come, even if the articulation is the same in each. And my goal is to make this as realistic as possible, to sound real as I play rather than go into each track after recording and use the humanize and other MIDI functions (which I still can use to fine tune later), so I was thinking that I need to make small changes in each track as well as the plugin itself, while keeping the same articulation.

For example, I won't have 10 play sustain, 10 marcatto, and 10 pizzicatto because it's not what I'm looking for. I'm looking for 30 playing long sustain legato with portamento. But, I can change the portamento and legato time depending of the instrument (I plan on mixing solo violins from different companies).

If the engine has a humanize option, I plan on using it, but if not, I was thinking of experimenting with the MIDI FX and whatever else Logic has.

What MIDI FX would you use for this, and how would you use them? Anything else I can use, Either bundled in Logic or 3rd party that is free or very inexpensive?

[Atlas007]

The first two (stock) MIDI Fx I’d experiment with would be Note Repeater (as delay source) and Randomizer (in various combinations/configurations)…

[GeneralDisarray]

Thanks, that's actually two I was thinking about, after I learn to use them, which shouldn't be too complicated. 

I was also thinking of adding some differences in EQ that would be subtle but to give each violin its own sound signature so to speak, which would be a way to get closer to real players, each with their own brand and model of violin, and even if they were all playing the same brand and model and year of manufacture, what I have learned about violins so far is that even in that case, they wouldn't all sound exactly the same.

Does that sound like something that would work?

I have another question about violins, but I'm going to start another thread because it's not totally related to this specific setup.

[Atlas007]

4 minutes ago, GeneralDisarray said:

Does that sound like something that would work?

There is nothing like trying it!

Keep us posted!

[GeneralDisarray]

20 minutes ago, Atlas007 said:

Keep us posted!

Definitely will!

[David Nahmani]

There are a few more things you could randomize to differentiate individual violins. Assuming you're using Logic's Studio Strings instrument, adjust different values for the cutoff, resonance, and attack. Add a Pitch Shifter plug-in to each violin track and detune the different violins by various amounts in a -5 to +5 cent range (experiment for the best results). 

Finally pan each violin at a different angle in the stereo field, and feed them all to a stereo reverb to place them in the same room. 

[fuzzfilth]

On 2/19/2023 at 8:03 AM, GeneralDisarray said:

This may be a dumb idea

You are probably right.

The idea is to play back the (single line) score and make it sound as being played by an entire section, right ?

I'm sure you have already found out that duplicating a string track just makes it louder, unless there is at least a minuscule amount of randomisation and/or round robins going on. In essence, you're playing the same sample twice, which does not make it sound like two violins, but still one that's precisely 6dB louder.

In an effort to get around that, you might try all kinds of randomisation tricks, like random and/or slightly modulated delays, different EQ, stereo position, etc. all of which do not address the basic problem: you're still playing the same sample twice, and if you do this, you will always either get simple additions in level (with EQ and Pan) or get really weird and  obvious phasing effects (with delays and modulation).

So you need to play as many different samples as you want voices. If you pile up all your different libraries playing the same note, you actually get the effect of many simultaneous voices, but the differences in these will be huge, as each library will have its very distinct sound to it, due to players, instruments, mics, mic positions, rooms, processings, masterings all being radically different. Unless you manage to find one library which actually has every violin player sampled separately with all articulations which would be a huge and costly project. I don't know of any. I'm not sure also if you can set up a system that can actually play all the single voices of an orchestra with a reasonable amount of resources.

And even then, you're not taking into account that every one of the thirty violin players has a slightly different approach to when to actually play the note, how to craft the attack, swell, bow pressure, micro pitch, when and how to bring in the vibrato, how to sculpt the vibrato speed in relation to its depth, all of which are dependent on the meaning of this very note within the entire phrase/line/movement/piece.

So, regardless of all the trickery you might come up with, in the end it may turn out as elegant as the drawing of your four year old kid, which may be endearing if you know its source, albeit, elegant it is not.

Edited February 20 by fuzzfilth

[Atlas007]

Despite @fuzzfilth discouraging caveats, I'd still experiment and use those as clues to workaround them...

[fuzzfilth]

Of course. And you're cordially invited to prove me wrong.

Edited February 20 by fuzzfilth

[GeneralDisarray]

13 hours ago, fuzzfilth said:

So, regardless of all the trickery you might come up with, in the end it may turn out as elegant as the drawing of your four year old kid

Well... are we talking about the drawing of my four year old as in me drawing him, or him drawing me? Because they would be drastically different.

This would be my 4 yr old drawing of me:

While this would be my drawing of my 4 year old:

As you can see, the difference is night and day.

Sorry, couldn't resist. Don't take it personally, Smartass is my second name.

Now, about your very valid points about this project, which I really appreciate:

 

13 hours ago, fuzzfilth said:

I'm sure you have already found out that duplicating a string track just makes it louder, unless there is at least a minuscule amount of randomisation and/or round robins going on. In essence, you're playing the same sample twice, which does not make it sound like two violins, but still one that's precisely 6dB louder.

Well, I have duplicated string tracks, but even in my intermediate at best knowledge of Logic and DAWs, it wouldn't occur to me to just leave them all carbon copies of each other, because what you said is exactly what would happen.

13 hours ago, fuzzfilth said:

In an effort to get around that, you might try all kinds of randomisation tricks, like random and/or slightly modulated delays, different EQ, stereo position, etc. all of which do not address the basic problem: you're still playing the same sample twice, and if you do this, you will always either get simple additions in level (with EQ and Pan) or get really weird and  obvious phasing effects (with delays and modulation).

So you need to play as many different samples as you want voices. If you pile up all your different libraries playing the same note, you actually get the effect of many simultaneous voices, but the differences in these will be huge

Since I'm not a musician, barely an aspiring one, and it takes me too long to play long note progressions, I play a couple of bars at a time, unless it's an easy one, but I don't have the best motor coordination, even at typing text on a keyboard I keep hitting the wrong keys or two keys at the same, besides having dyslexia, so half the time I'm typing text with letters swapped, and that happens when playing notes on the MIDI piano.

So when I'm done with the whole thing, or at least that section, I open the MIDI transform panel, and start doing some things like set the notes all to the same velocity, because of my awful playing when recording, I may end up with notes that are 1, some 127, and everything in between, but not the velocity I was trying to accomplish. 

Then I start editing the velocities mostly based on what I hear in the original song, sometimes just doing a crescendo or diminuendo, listening to it, probably doesn't sound right, so I do it again with a different velocity range, and so on. Then I apply a humanize and a couple of other things I can't remember right now, and then in the piano roll I may quantize and probably move some notes manually, because my internal clock is not that great, and I can't stand the metronome, it ends up confusing me even worse when I play.

But since I started going into more places in Logic and learning more features, I started loading the MIDI FX and saw there was some great stuff in there. So I started thinking, what if I could use these as one way to introduce minor changes in the playing as it happens, so that one "player" has a little extra vibrato, or delays in notes, and other things that will give each virtual player its own uniqueness, while still feeling as part of the whole.

And then there are other tools beside the MIDI FX. Unless it's a section of those $200 violins that were manufactured in an assembly line and all sound the same, I understand that each quality violin would have its own unique sound, within reason.

My idea for trying this is that all the libraries I have that come with violin ensembles, they sound great, but when I put them side by side (in a manner of speaking) with the original track I'm trying to play, the ones in my library don't sound similar at all. For example, I have Spitfire's Hans Zimmer strings, which is by far the most expensive considering that it's just four instruments, when I paid the same for the entire Eastwest Hollywood Orchestra Opus edition, which not only comes with more string instruments, but also with brass, woodwinds, string solos (which HZS doesn't have). I'm a huge fan of his, and it was something I thought that it was worth the $400 or so, even if it's just string ensembles in large numbers.

But I did some comparisons between several of his tracks from different movies, imported from the CDs in AIFF, so while it's not Hi-res lossless, it is still the same track as the CD, and I tried to match the long strings in the original songs with the virtual instrument, trying out different choices, like 60 violins, or 20 left, 20 centre and 20 right, and finally what they call Violins Galleries.

And if I listen to those just playing and moving MOD and EXP, while adjusting the plugin's own controls like reverb, release, vibrato, etc, the sound really nice. But when I recorded the same notes as a section of the original song, usually long sustain and legato, the sound is really not very similar. 

Of course given my inexperience I wasn't expecting the same exact thing, that would be absurd because his original recordings were heavily processed first by him and maybe by him again at the mastering stage, but they don't sound anywhere near. And the truth is they don't even sound that much as a section of 60 or 20 players playing at the same time. And this is not just his plugin, if I use Eastwest's Opus 18 violins for example (at least when I can because it's not very stable and it crashes unless Logic is ran in Rosetta 2 mode, which means slow), they all sound like all the players playing exactly the same, as if they were androids programmed to play exactly the same exact note, with the same exact intensity, at exactly the same time.

So I thought, if I try to replicate a real violin section and give each virtual player its own uniqueness, perhaps it would get closer to the sound from a real orchestra. Does that make sense?

That's why 

 

 

[Jgirv]

Actually, all of the opinions and advice here are good.  Definitely try any and all. 

A thought, though.    Playing orchestral libraries realistically and well is a skill that takes practice.  A lot of it!  

If you are specifically after getting a realistic string section sound, any of the top-tier libraries will absolutely do it.  You need, however, to allow yourself time to learn how to approach them on their own terms.   It takes careful performance as well as editing to get stunning results.

Since string sections are not recorded by placing a mic on each player, your idea prob wouldn’t sound like a professionally recorded ensemble.      But go ahead and see what you get, you might like it.  

Last caveat…. Back in the early days of MIDI I tried to layer up a huge ensemble using every synth I could get my hands on…. Eventually it all sounded like a giant accordion.   Not that that is a bad thing (love that instrument) but it certainly wasn’t the huge orchestra I was trying for.     

[fuzzfilth]

15 hours ago, Jgirv said:

Playing orchestral libraries realistically and well is a skill that takes practice.  A lot of it! 

Absolutely

15 hours ago, Jgirv said:

It takes careful performance as well as editing to get stunning results.

Absolutely

15 hours ago, Jgirv said:

it all sounded like a giant accordion

This very sound has actually been preserved for posterity in the 'Tutti' section of the original "Peter Siedlaczek's Orchestra"-AKAI-CD-ROM from the 90s of the last century, one of the first trying to make the orchestra playable on a sampler (650Mb of sounds for a 32Mb sampler, no less). He got it substantially better in the follow-up "Peter Siedlaczek's Advanced Orchestra", though.

[Plowman]

Quote

Back in the early days of MIDI I tried to layer up a huge ensemble using every synth I could get my hands on

We've probably all tried this. What the OP is doing is an obligatory rite of passage. It's a worthwhile, educational pursuit.

But imagine this: if it did work, okay, let's proceed to senza vib, espressivo, trills, detache, spiccato, staccato, and tremolo (and those are just with the bow). Now multiply all the needed computing power by four other sections. 

Ironically, if it works at all, a useful pizzicato sound might be found with per-instrument stacking -- particularly if slight delays in attack are used. 

Echoing the point made above about miking, VSL's Synchron Strings have up to nine microphones, and none of them are on solo instruments. The Synchron Elite Strings (chamber ensembles) have a solo mix channel, but even that is not completely dry.