Sampling vs Synthesis - do i need synths?!

[René]

I see, thanks. If they have said so it's for sure that way. In any case, I'd like to mention this, which has been the object of discussion of many devs:

Out of the several synthesis methods available for oscillators, few synthesizers really calculate *everything* realtime. Even the supposedly-fully-realtime BLIT method is sometimes implemented by storing intermediate subfactors in tables.

Now, what's the difference between using a table of trigonometric functions which will be later be used to generate the sound, or just to store the sound already predigested? None, of course, as long as the output of the calculations results in the same damn number.

I don't think that in this 'world of realtime' all manufacturers will reveal to what extent their 'realtime' becomes 'almost realtime' or 'ok, a little bit offline but most realtime', but I wouldn't just deduct that all is realtime looking at the ads.

I'm not saying that anyone lies, because the truth is, noone is asking. And actually, I believe that the result is what matters. I just want to say that the language is often thought as pure and absolute, while it's not, it's just a method for communication.


-René

[rounser]

I think it is pointless to say the sampler is superior because it can sample a synth and thus be both. Something is lost. Maybe you do not value that something, but someone else does. Same with sampling a violin. It does not equal a violin. Something is lost.

You've got it backwards....when you resample and stack that synth patch with three others, and destructively add compression, reduction of DC offset bias, and normalisation, something is most definitely gained.

This depends on the work methods of the samplist, though, and whether or not you're a preset-a-holic who simply must have everthing exactly the same.

[pdxindy]

I think it is pointless to say the sampler is superior because it can sample a synth and thus be both. Something is lost. Maybe you do not value that something, but someone else does. Same with sampling a violin. It does not equal a violin. Something is lost.

You've got it backwards....when you resample and stack that synth patch with three others, and destructively add compression, reduction of DC offset bias, and normalisation, something is most definitely gained.

This depends on the work methods of the samplist, though, and whether or not you're a preset-a-holic who simply must have everthing exactly the same.

Something gained, something lost. That is subjective. I like String Studio and want to play the presets I make with it in realtime.

What you are saying is the same as saying the sampler is superior to a violin because you can sample the violin and add more sounds to it. Some people want the beautiful expressive sound of a live violin just as it is played.

The point being that what you value is not what everyone values and so there are different softwares for those different ears and interests.

Of course, the host is a sampler and you sample your performance and then adjust it with EQ, compression etc.

[aciddose]

listen to that track in "the fifth element" you know, with the samples of the opera singer.. that is what a sampler is for. you can make violin sounds with a synthesizer.. they dont sound nearly like a real violin, yet can a violin change its timbre dynamically?

you can sample a violin or a synthesized violin and it wont sound as dynamic anymore.. but can either a real or synthesizered violin be sequenced with such percision?

like i said.. i think a sampler is like salt, you have to use it for the advantages it gives you.. but you wouldnt want a full course meal of it.

[rounser]

A violin is a straw man example - who has a violinist on hand to appear in their next acid track? It's far easier for a sampler to emulate a synth, just as it's far easier for a synth to emulate a synth. Good luck getting a realistic violin out of a synth; you're holding synths up to a different set of criterion than samplers. This is also known as double standards.

That said, first stop if you don't have an orchestra available to you is (surprise, surprise) a sampler library. Using a synth for such purposes isn't on the cards unless you're making (wait for it) electronic dance music, where perfect reproductions are optional, and "synth strings" can be just what the doctor ordered. If you want authenticity, samplers regularly fool people into them thinking they're the real deal, whereas a synth generally has little chance of that.

but you wouldnt want a full course meal of it.

Maybe you wouldn't. There are entire dance music genres where the sampler is the primary instrument, and the synth entirely optional (dnb and house spring to mind). It's not mere "salt" - that's the synth. The result of adding too much "synth salt" is stuff that sounds like electro or trance, and deep down, no-one really wants that*. Think of the children.

*: Note that I do like trance and electro rather a lot, I just don't want everything sounding like it...

[Seus Hawkins]

if you have a inferiority complex towards instruments etc., perhaps computer music is not the way to go?

the way i see it, synthesis appears to be more of a celebration of the computers capabilities rather than sitting down at a computer and thinking, ok, let's make music, i'll use this sampled guitar, this bass, this orchestra etc. (all sampled instruments where information is lost as said etc.)

of course, there are many mangling possibilites which makes sampling interesting too and way more synthetic/synthesizer-like. and the argument that people don't have orchestras at hand. still, if it's not used in combination with synthetic sounds but at mere orchestral work composed without the intent of proving some point, i think it appears to be some kind of conservativist inferiority complex.

regards

note that i use a lot of sampled instruments attempting to sound real, i still have a lot to learn and would like to use 100% synthesis when it appears possible. sample mangling is of course also interesting.

[griels]

I think what's been forgotten here by the 'synths are redundant' lobby is the sounds *inbetween* the notes. You can sample a synth with a million different filter/osc settings, but you'll never capture the entire range of expression that is capable with it; synthesis is a continuous phenomenon, not a discrete one representable by distinct quanta. And no, using the sampler's inbuilt cross-modulation/filters is not the same. If you can't hear that, well, good luck to you.

On the other hand...

of course, there are many mangling possibilites which makes sampling interesting too and way more synthetic/synthesizer-like. and the argument that people don't have orchestras at hand. still, if it's not used in combination with synthetic sounds but at mere orchestral work composed without the intent of proving some point, i think it appears to be some kind of conservativist inferiority complex.

Not necessarily - I'd say there is some need for perspective here (as with the sampled synth argument) - the question is, whether the 'lost information' is significant to the listener. Sometimes the accuracy of, say, legato reproduction is not entirely relevant to the track's aims. I doubt many people listen to commercial radio and think 'ooh, that's a nasty formant shift', for example. By the way, this is applicable to sampling synths as well as sampling violins.

[rounser]

the way i see it, synthesis appears to be more of a celebration of the computers capabilities rather than sitting down at a computer and thinking, ok, let's make music, i'll use this sampled guitar, this bass, this orchestra etc. (all sampled instruments where information is lost as said etc.)

But sampling's not about loading up a bass guitar patch in Kontakt. I understand the where the views on this thread are coming from if your idea of what sampling's about are so limited...

I'm most interested in the rave branch of the dance music tree, which is all about sampling just random stuff, and has nothing to do with orchestral libraries or presets of most any kind for that matter. The otherworldly vibe that rave music has owes everything to the sampler. Those looking for the ultimate "ravesynth" are often surprised to learn it's identity. It's not the TB303, nor the Alpha Juno, nor the JP80x0s...from breakbeats to mentasms to reeses to "belgian" stabs, and countless other signature sounds, it's the sampler.

As far as standard bread and butter sounds go, like a pad, I think those espousing views like yours would be quite surprised how much resampling goes on in the tracks you enjoy. I've noticed a very strong correlation between heavy samplists (as in, it's their main instrument, or invariably the starting point for their tracks) and the most popular dance music acts around, and I don't think that's accidental. Daft Punk, Fatboy Slim, The Prodigy, Future Sound of London, Basement Jaxx, Mylo and other "pop dance" acts spring to mind, along with a whole host of genre "names" who are huge in their respective ponds.

And no, using the sampler's inbuilt cross-modulation/filters is not the same. If you can't hear that, well, good luck to you.

It's close enough to the same not to matter, except to the anal. If you are that anal, good luck to you, too.

[griels]

It's close enough to the same not to matter, except to the anal. If you are that anal, good luck to you, too.

But this whole thread is anal

Seriously, I can often hear the difference - and I have listened to commercial tracks and thought "ooh, that's a nasty formant shift".

[defjamm]

I think what's been forgotten here by the 'synths are redundant' lobby is the sounds *inbetween* the notes. You can sample a synth with a million different filter/osc settings, but you'll never capture the entire range of expression that is capable with it; synthesis is a continuous phenomenon, not a discrete one representable by distinct quanta. And no, using the sampler's inbuilt cross-modulation/filters is not the same. If you can't hear that, well, good luck to you.

nobody i know listens to something like that in music. i think it depends on the view and what you expect from your 'music hearing experience' and the genre. most people on earth imho don't care about the above when they listen to music. your view is from a synth-lover.

personally i first listen to the vocals, then the drums, then the melodies...there will be so much going on at this time(information for the brain), that i won't hear and look for things like you describe above.

but sure, if you listen just to a sampler and a synth alone und compare then the differences should be obvious.

imho if you want to make some music, get a sampler or a rompler or 2, get 3 synths and don't come ever back to kvr

[griels]

but sure, if you listen just to a sampler and a synth alone und compare then the differences should be obvious.

Yeah, see my disclaimer above...

I think the point is it depends on what the focus of the music is. If you're making some kind of minimalist synth music then it's quite possible that your audience will be unfortunate enough to appreciate such distinctions

Or, if you are playing something richly expressive, on whatever instrument, I think even Joe Public can pick the difference between that and a sampled imitation, even when masked by a complex backing..

[rounser]

If you're making some kind of minimalist synth music then it's quite possible that your audience will be unfortunate enough to appreciate such distinctions

In which case you'll make sure you have high quality samples in the first place, and use them in a way that plays to their strengths, or you'll sound like a demo. The same thing goes if synth patches are poorly chosen in a minimal techno track.

Or, if you are playing something richly expressive, on whatever instrument, I think even Joe Public can pick the difference between that and a sampled imitation, even when masked by a complex backing..

It depends. If you're going all general midi with a rompler, quite possibly, but if you've sampled little chunks of sax or bass with all the artifacts of the player still there (as is often the case on house tracks) and either played it back or built a new riff out of it, then probably not.

[griels]

It depends. If you're going all general midi with a rompler, quite possibly, but if you've sampled little chunks of sax or bass with all the artifacts of the player still there (as is often the case on house tracks) and built a new riff out of it, probably not.

Agreed

[shamann]

I think what's been forgotten here by the 'synths are redundant' lobby is the sounds *inbetween* the notes. You can sample a synth with a million different filter/osc settings, but you'll never capture the entire range of expression that is capable with it; synthesis is a continuous phenomenon, not a discrete one representable by distinct quanta. And no, using the sampler's inbuilt cross-modulation/filters is not the same. If you can't hear that, well, good luck to you.

I think quincy's point at the start though was not about Atmosphere-style synth patch samples, but whether or not a sampler, if loaded with single cycle wavetables, could act as an efficient subtractive synthesizer. I can't imagine why anyone would think that he couldn't, some samplers have a whole slew of excellent subtractive features.

There are differences of course if he wants to use something like Atmosphere-style patches, the biggest being that they can become difficult to work with. With a synthesizer, it's easy to shape the sound according to your music, but if you are just working with a recording of that sound you often have to work your music around the sound. But even that's getting easier with some of the tools available today.

[fac]

I think quincy's point at the start though was not about Atmosphere-style synth patch samples, but whether or not a sampler, if loaded with single cycle wavetables, could act as an efficient subtractive synthesizer. I can't imagine why anyone would think that he couldn't, some samplers have a whole slew of excellent subtractive features.

With today's samplers and a good set of basic waveforms, I would say you can obtain a good range of substractive sounds. Typical basses, pads, strings, EP's, leads, etc. should be no problem (I still do it with an old Akai S2800). But depending on the sampler, some things like PWM, FM, AM, RM, Filter FM, etc. may be hard or impossible to obtain.

However, let me repeat that today's samplers feature very deep synthesis engines (e.g. Shortcircuit) with multiple modulation options (modulation is the keystone to synthesis). These are "samplers" in the sense that they use sampled waveforms as oscillators, but they are also fully-capable synthesizers.

Finally, there's the issue of "character". No one can deny that certain synths have a very particular character, and a sampler can only capture some of it, but not all. You can load a set of minimoog waveforms into your sampler, but the sampler's filter won't respond as a moog filter. The same applies to the envelopes, amplifiers, and even the way the synth responds to the keyboard (legato, etc.).

There's also things that samplers can do that a synth can't. They're just complementary tools.