Originally Posted By: Tangmo
Therefor, if it doesn't exist already, software could be used to more closely emulate the characteristics of a selected patch/instrument and be tweakable on-the-fly without a user having to get into the file itself to manually with complicated commands and language change the line to emulate realistic play. Hardware controllers can do this. Why can't software? Audio FX software often has the "look" of analogue. Why can't midi editing software "feel" more natural?


This question has come up on various forums for years by people who do not have a good understanding of midi. To start, it all begins with the original samples. You've heard of multisampled piano's? Some software/hardware will not talk about it meaning it's only one layer, others with say 3, 5, 6 or even 8 layers. What's the difference? Velocity and tone based on how hard the pianist hits the keys. If it's 8 layers they have to hit each individual key using 8 different levels of force. Manually with their finger. You try that sometime. Each layer is a separate recording and then those layers are put together and mapped. Ever read about a high end piano patch being 2 or 3 GIGS in size? That's why. More layers the bigger the file size is.

That means the more powerful computer you need with a ton of RAM to load it. It's the same with all instruments. Better sample recordings with more layers, the more control you can have. Take horns for example. They're not just straight tones. Oh no, they have falls, doits, slides, honks, lipping effects, tons of stuff that makes a sax sound like a sax or a trumpet sound like a trumpet. Each one of those has to be in the original sampled recording or else how could you possibly get it out of some piece of software? Is the software simply going to add a growl to a trumpet line? From where? It's not been recorded in the sample so where would that come from?

Similarly, if you have a piece of well written software to make a piano sound more realistic but the synth being used only has say, two layers of piano then no matter what that software does, it's not going to give you the different tonal characteristics you get from hitting a piano very softly then harder up to banging it with 20 pounds of force. This is why an 8 layer piano sample library can cost $1,000 by itself forget all the other instruments.

Mrgeeze mentioned steel guitars. One of the hardest instruments to emulate. It's not just pitch bend, it's how the player operates the levers and pedals while he's playing. Same as my piano example, how was the steel samples recorded? There's a big difference between how single notes are bent vs a chord and where on the neck it's being played.

No software is going to make up for how the original samples were recorded, it's called garbage in/garbage out. Many of these more difficult instruments have their own specialized controllers to get a correct emulated sound. AND, most importantly don't forget the skill of the musician doing it.
It's true the reason for all this is to allow otherwise very good players to be able to create tracks of instruments they don't play. BUT, they have to be good enough overall musicians to have a very good understanding of what goes into a good horn sound for example and use their own skill and understanding to emulate it believably.

You're talking about us schleps with little training in all this trying to come up with something realistic in our living rooms with maybe a few thousand bucks worth of equipment. And, watching a few YT vids showing how to do it given by a pro who's got a college degree in music production and who's been doing it in studios for years.

Ain't gonna happen easily, at least not yet. I fully appreciate how fast tech is improving so who knows what could be possible in a few years but right now, in a home PC that is affordable to us at vs a super computer running one off custom software in a university lab, there's nothing yet and may never be. Why? Because in spite of our passion for it as musicians this stuff is a tiny percentage of the overall computer market. Our passion is specialized and expensive and it's amazing to me that for the amount of money Biab costs we can get the results we get.

As you can see this stuff is not easy to explain in a few words. To reply to the comment that PGM should make the RT's available as single note samples, reread what I just wrote above. A recording session to play an instrument like we all think of, recording phrases, licks and solos is one thing. Setting up a recording session for sample recording for a software synth is a whole other thing.

Now you asking the artist to record one note at a time up and down the instrument. That's just for the basic note. To add all the other stuff as individual layers takes hours and hours of studio time which is why companies such as Native Instruments, IKMultimedia, East/West and all the others specialize in that. I'm not saying if PGM should or shouldn't do that, it's up to them but it doesn't make a lot of sense to me. To say they already have the RT recordings is irrelevant. They were not recorded with individual samples in mind and if you want to go into the wav file and simply cut out individual notes then reread what I wrote above again. Won't work, it will sound like a chopped up mess.

Bob

Biab/RB latest build, Win 11 Pro, Ryzen 5 5600 G, 512 Gig SSD, 16 Gigs Ram, Steinberg UR22 MkII, Roland Sonic Cell, Kurzweil PC3, Hammond SK1, Korg PA3XPro, Garritan JABB, Hypercanvas, Sampletank 3, more.