synth or sample?

[BillWelcome Home Studios]

A while back in the middle of a thread I was getting a lesson in synth/sample from an experienced user. I had hoped that the dialog would continue but it seems more appropriate to start a separate thread and get more input from a wider user base.

 

When I see 'synth' in a discussion about computer audio I automatically think 'sample loading and playback' because all the threads I see on this subject have to do with problems in loading samples into RAM and getting everything done. In the course of the previous discussion it was pointed out to me that synths don't need so many resources (which I knew...) and that many/most guys use the synths. So the logical conclusion that i got was that the reason that I always thought of the typical usage in the way that I did was that the guys using synths with the synth sounds had no problems, so they didn't write about it, I didn't read about it, and so I didn't know about it. I had thought that the guys with resource problems were typical, wherein it seems that this is not typical.

 

So where do most of you stand? Do you use the synth in your DAW more often, or do you rely upon sample libraries triggered from the synth?

 

I'm trying to dig more into this because as I move into retirement I'm going to be wanting to make more of the music myself, so I'm trying to get a practical overview. I use Samplitude/Sequoia, which has a couple of synths included in the package, and I've collected a couple of sample library downloads over the last couple of years looking forward to this time. Reading about how others use or handle the process helps me to extrapolate a solution for my own needs.

[Synthoid]

So where do most of you stand? Do you use the synth in your DAW more often, or do you rely upon sample libraries triggered from the synth?

 

I started with real analog synths in the mid-80's and prefer that type of sound to the "real thing." I don't own any sample libraries, other than the ones already installed in my Korg M3, Triton, Motif XS, etc.

 

Most of the music I play is rock and electronic based, but I'll play a realistic B3, Rhodes, Wurli, or acoustic piano sound once in awhile, but it doesn't have to be perfect. The audience doesn't care anyway. I prefer string pads, synth leads, filter sweeps, and an occasional Fairlight ARR sample. Heaven forbid, I'll even throw in the DX-7 Rhodes sound as well.

 

 

[ADino]

Very rarely do I use anything outboard, when I do its samples from an Ensoniq asr10, and only because Ive accumulated so many thru many years of using their products.

 

All of my samples are sw based, so over the past 2 or something years of using LogicPro, Ive accumulated quite a few. Add into that the process of moving everything to an intelmac from a G5, plus adding a few new ones, and things get hectic and expensive.

 

Luckily, there are quite a few really good deals floating around from the sw companies, so if youre going to get any new ones, nows the time. But my studio is not a studio for hire. I do everything myself for my clients, so I need a wide array and variety of sw based instruments. Whenever I finish a project, Im always looking for new sounds, or ways to mod what I have.

[J. Dan]

I think I remember the thread you're talking about. If I recall, the discussion was around how much memory/cpu VIRTUAL synths require vs SAMPLEs, not hardware synths, correct?

 

I'd say it depends on what you're after. The thing about virtual analog for instance, is that you can control the waveform and filter to change the sound realtime. With a sample, you're stuck with what's sampled. Sometimes that's ok, other times it's not - depends on how much the sound needs to move and modulate.

 

The waveform of a virtual SYNTH is generated mathematically and ouput to the sound card, so very cpu intensive, as opposed to just streaming from a file, as it is with samples.

[timwat]

It also seems these are wide generalizations that are becoming less helpful as new products are developed and released. There is/was the assumption that additive synthesis would require big horsepower but recent products seem to defy that. The burden of sample streaming would depend upon the actual size of the samples, which varies from product to product. Complex FX usage would also eat resources, no matter what type of virtual instrument is used. Omnisphere comes to mind as an essentially sample-based product that does so much DSP and manipulation that it has the reputation of being a total resource hog. Finally, poor or inelegant coding wastes resources no matter what...a good coder is still worth his weight in Twinkies and Mountain Dew.

[JeffLearman]

It entirely depends on the sounds we want.

 

I play mostly classic keyboards (piano, hammond, rhodes, wurlie, clav, and minimoog, roughly in that order).

 

For pianos and epianos, it's samples. For hammond, it'd dedicated software (NIB4 or NE2). For minimoog, samples don't cut it; oscillators with subtractive synthesys is required -- regardless of whether it's analog or digital. (I go digital because it's what I have, and I'm not deep into it enough to crave tanalog.)

 

Players who use pads need more synthesys -- you just don't get the same effects from samples. For example, Nick Rhodes of The Fixx -- no samples there. (OK, that was before samples were very big, but even now you'd a synthesized voice, not a sample.)

 

Keep in mind that there's a 3rd option, where the waveform itself is a "sample" (often just one cycle), replacing the traditional oscillator, but using the same subtractive synthesys techniques to shape the tone.

 

Also, well-adjusted patches for sample sets also often use subtractive synthesis techniques to make the sample set more responsive/dynamic with fewer velocity layers. The more memory we get, the less we need to do this, and we can use lots of layers layers instead.

[J. Dan]

For example, Nick Rhodes of The Fixx -- no samples there.

 

I'm pretty sure the keyboard player for The Fixx was Rupert Greenall, Nick Rhodes is the keyboard player for Duran Duran, but I think you're right on both accounts.

[ADino]

timwat, I agree def on the coders....there's some sw thats a pleasure to use, n some not so...I just did the atmo to omni upgrade, havent loaded it yet..

[BillWelcome Home Studios]

absolutely we are talking broad generalizations, and I'm okay with that. Inelegant code not only sucks, but looks terrible on anyone after 6.

[Sven Golly]

Players who use pads need more synthesys -- you just don't get the same effects from samples. For example, Nick Rhodes of The Fixx -- no samples there. (OK, that was before samples were very big, but even now you'd a synthesized voice, not a sample.)

 

Not to hijack this, but (as Tim mentioned) Nick Rhodes is the keyboardist for Duran Duran, and he most definitely used samples; he was one of the earliest 'big names' to use the Fairlight CMI on-stage, with a lot of use of their choir samples mixed into his pad sound.

 

Okay, back to the topic at hand.

[Synthoid]

Nick Rhodes is the keyboard player for Duran Duran,

 

Correct.

[richwhite9]

Ahhh..,,the Nick Rhodes cover of Keyboard that heralded the end of Western civilization if I remember the ensung firestorm of 'letters to the editors' correctly....

 

Acoustic samples are simply better for replicating acoustic and natural sounds and electro-mechanical sounds. Modeling will get there.

 

There's so much music being created today on computers and softsynths that is 'electronic' and isn't sample based or organic.

 

A general rule of thumb is that good sample sets have articulations to recreate many of the nuances of a organic instruments and sounds. The playing technique is much different (and less forgiving) than a softsynth controller. Trying to create those articulations on a softsynth and then map them to a controller? What a nightmare. Wrong tool for the job IMHO.

 

That may be the best guideline. Samples of synth sounds? Naah.... not unless you are a rompler type of guy that won't ever program. Better to buy Mini-Monstah and as many preset sets as you can get than to buy a pure Moog sample set IMHO.

 

Most synths are by nature computers. Acoustic and organic sounds by nature are not computers. Synths and samples each have their place.