Well - I think the future is somehow mixing the two technologies if thats at all possible. Using real recordings and then have some sound engine manipulate the sound for vibrato for instance. Sustain notes could also be altered just a fraction so that no note sounds completely alike. I'm also thinking this could be a huge benefit for srepetition programs... a staccato note which is then altered slightly and tied together with the next note - all based on real instruments and sounds.

Again I dont know if its possible, but I definatly think its the future, and think that if anyone can do it its VSL. And I hope its a route they are thinking about.

>The fact is, I want samples, not synthesis, BECAUSE they are separate recordings of notes, articulations, etc. played by real instruments. The art of sampling is using recordings, not fakes, of an instrument.

Well, i'll go for a drastic comparison now.

Imagine watercolors the only, albeit advanced, means to create pictures, and people have gone at great lengths to provide you with the means to pull it off as authentic and realistic as possible. So you have stacks of trays with gazillions of water colors in all shades and you have learnt to make good use of them.

Now someone comes along and tells you about this new technology which is called photography, which can easily capture in a mere second what would have took you days to paint.

Now, you kick and scream about it not being watercolors, so how could that be taken seriously and whatnot.

Still, folks will very likely begin to flood the new guy's studio to have their picture taken...

Point in case is, what do you head for to have chiseled on your tomb stone ? This one ?

"Here lies good old Bill, who was a master in making independent recordings of tones sound as if they were in fact being played by an orchestra."

or would you rather prefer that one :

"Here lies good old Bill, who was a master in making orchestral music."

Regardless of the current development stage and realism (or lack thereof) of <the unspeakable>, sooner or later we have to acknowledge that sampling by itself is a static procedure, a snapshot, and while it provides you with instant realism and expressiveness *for this very recording of that note*, it will fight you forever when you try to bend it into something else. That's why they pile up layers and layers of velocity switches because what you seek is just not contained in one single 14" tom tom hit.

It's a bit like trying to assemble a novel with a construction kit of words. You'd need a pretty decent stock of premade words to find exactly the ones you're looking for and a good database structure to actually find them if they're there. Now, if you had opted for the *letters*-toolkit, you'd just have to make sure you have a good amount of all two dozen different letters and you're set.

I have the impression, that <soandso> might evolve into a decent letters-toolkit.

Christian


[URL=http://www.artofthegroove.com/logic/mp3/Christian_Obermaier_demo.mp3]show reel [/URL] [URL=http://uk.geocities.com/christianobermaier/home.htm]home page [/URL] [URL=http://uk.geocities.com/christianobermaier/Studio.htm]studio pics [/URL] [URL=http://uk.geocities.com/christianobermaier/Gearlist.htm]gear list [/URL]

William,

I received that notice as well, and was intrigued enough to check it out. Actually, it's not entirely accurate to call it synthesis -- it's a form of sample resynthesis, or synthesis by analysis. It does use a store of samples as "source material" and, like VSL, that source material consists of sampled performances. The only difference is that the sample set is translated into resynthesis parameters, rather than being used directly. Now, because of this translation, it's possible to move through the "sampled" material in a non-linear fashion. So, the midi analysis part of the program basically looks for the best "match" in the sample-set to recreate what's being input via midi.

I actually freaked out a bit when I saw the details because it's eerily close to a proposal I "pitched" to a few companies a couple of years ago -- right down to using additive + noise for the resynthesis, and using comlete recordings of concert works for the sample-set (as opposed to individual notes). Pretty wild. But then, the developer has been working on it for eight years, so I only managed to be six years behind the times! [;)]

Back to your question. As Christian (1) pointed out, things are certainly moving in the direction of some sort of hybrid. It's just a question of what that hybrid will be, and how it will manage to bridge the gap between the sample and its resynthesis. Really, the only form of "ground up" synthesis that stands a chance is physical modeling, but until someone comes up with a feasible control structure, it will not likely find a place in the "orchestral" market. A sort of middle-ground can be found in the program Melodyne, by Celemony software. This program can perform some real miracles, in that it can separate out pitch, time, and formant -- it can even flatten out vibrato! If you mess around with Melodyne for a bit, you'll instantly recognize the potential for composers like ourselves. The key bit of "magic" lies in the ability to provide subtle variation to a given sample's pitch, vibrato, duration, and amplitude dynamically. This is what will really bring a sample library to life (okay, there's plenty of "life" in VSL, but you know what I mean!) There is little doubt in my mind that this is the direction in which VSL is moving. Remember this little statement?:

"Revolutionary software developments in the areas of music production and education will significantly affect the work of music creators. In the near future, quick access to samples will no longer be a question of RAM or hard-disk streaming, but will rely on a completely intuitive connection between the composer's sonic vision and the sounds required to achieve it."

Notice particularly the part about RAM and hard-disk streaming... These factors can only truly be transcended using some sort of sample resynthsis (e.g., the software that inspired this thread packs an entire orchestra into 40 MB). Again, it's just a matter of which technique is chosen, and how it's implemented.

J.

christianobermaier

Yes, I think I'd prefer the second epitaph, thanks.

I agree with your point that going to the very basis of sound, instead of large scale "units" such as samples of instruments, will ultimately be very desirable in general. And I certainly don't reject new media.

However I am refering specifically to orchestral music, not all music. If we were discussing pure music synthesis, without reference to orchestral instruments, then you would be absolutely right. But we are talking about doing recordings of an orchestra playing music in its characteristic way. If it is done via samples, it is more natural than with a digitally created tone that has been made to sound like an instrument. In the case of samples, only the time element and the degree of specificity of expression is different than actually having an orchestra playing live in a recording. But in the case of the synthesizer, NOTHING is the same. It is entirely a fake. That was my point - in specific reference to orchestral recording or duplication of existing acoustic instruments.

Also, I am somewhat set off by the tone of this company - they are doing an outright aesthetic attack on sampling, saying it is artificial. But they are more aritificial, by definition. So my response is to their aesthetic critique which I find bogus.

By the way sir, your post reveals a fundamental bias against sampling - on the VSL website. Are you feeling a little bit conflicted these days? [6]

That's funny - I was wondering what JBM thought and he was writing it as I wrote. O.K., you are right on it not being entirely a fake.

For some reason, if this technology develops and is successful, I feel I may not want to do orchestral recordings with it. Not because of being crotchety (though of course I am) but because the technological process is infiltrating the very essence of the tone. In other words, one could make a case that I prefer the very lack of perfection inherent in sampling since it comes directly - even only in moments - from a living, breathing human being. What do the recording engineers do when recording these samples - they go to the greatest lengths possible to capture every nuance. All of this is distorted severely if run through a waveform deconstructing digital process. So this seems to be a fundamentally different approach that could be considered an artistic divide.

we can't be too precious about the technology we use ... sampling is as much on dodgy moral / artistric grounds as this new synthesis technique is.

... and whatever you feel about the artistic purity of of synthesised orchestral sounds, the phrase performances of this new tool-that-shall-not-be-named is awesomely good, even if the tone and timbre is not quite there yet.

Oops! Just got your reply, William.

I agree about the rhetoric thrown around in the advertising. It's a little bit much. And unfortunately, the instruments themselves don't sound *that* great. But it is fun to play around with (there's a 15-day demo), and does possess an impressive flexibility, which I sometimes find lacking in standard samples. I think it would really take some time to learn how to use it properly, though...

____________________

Aaack! Just got the next one!

I agree, in part. The thing is that it's only a matter of breaking down the sample into its component parts. All the elements of the sound (at least insomuch as we can perceive them as temporally distinct) are still there, they're just being used differently. I think what would make VSL's use of a similar technology different is that they would share your interest in preserving as much of the original performed quality as possible, only altering it when absolutely necessary (i.e., for making those gliss-to notes longer in the perf-leg_gl instruments). This is really only a matter of the implementation, and the philosophy behind its application. VSL would work to keep their samples intact, whereas this developer basically atomizes his. Again, Melodyne is the best model for comparison. There's a demo available -- give it a try with a few melodies exported from your VSL scores. It will probably blow your mind! The only problem, for me, with Melodyne has been the fact that it comes in a form/interface that I don't find useful for what I do.

I think VSL is doing something huge. My guess is a resynthesis by analysis-based composition program, with all instruments bundled in their analysed form, and all pulled together acoustically by the mysterious MIR... but that's some HEAVY CONJECTURE. It's not too far-fetched, though. It's *entirely* possible to do. And would the fact of it being almost sci-fi advanced ever stop Herb?

J.

Dear forum-members,

no need to beat around the bush - the name of the company is Synful, as everybody knows by now. It's founder, Mr. Eric Lindemann, seems to have joined our forum, too (as we all know by now, likewise [[;)]] ), so maybe he is willing to discuss these topics openly with us. - In the worst case we could agree that we don't agree. [:)]

There's no doubt that the technology employed at Synful is interesting, otherwise Yamaha wouldn't have made the effort to get their Vocaloid-principle patented (which seems to be very similar to the Synful-approach). - But it's beyond question that the Vienna Symphonic Library has chosen a different solution for virtual orchestration very consciously.

Greetings from Vienna,

/Dietz - Vienna Symphonic Library

hmmmm... Deitz, no hints?

I mean, really — the quote I posted above certainly does imply something bigger than the creation of .GIG files going on behind the scenes. And it's been on the site for as long as I've been a member. There's got to be something tasty you can tell us! After all, MIR has been awfully quiet lately, and there's no indication that it's being designed as a VST or a standalone... and the idea itself seems inherently multi-channeled, so it makes sense that it would be a "built-in" feature... but what will it be "built-in" to?

Come on, just a little hint! [:D]

J.

>Yes, I think I'd prefer the second epitaph, thanks.

I thought so [:D]

>Also, I am somewhat set off by the tone of this company

You're right there, a little more taste in the marketroid blurb would have gone a long way.

>By the way sir, your post reveals a fundamental bias against sampling

Which tells me my example was too drastic, sorry. I'm into sampling for quite a while now, and if i wasn't, why would i frequent this forum anyway.

>on the VSL website.

So what ? I think VSL is strong enough a product and a company that it can deal with an open minded discussion. There's no hidden agenda and no plans to take over the world from me (at least as far as i would tell you).

VSL is not a religion for me. It's a (highly sophisticated) tool.

>Are you feeling a little bit conflicted these days?

Not at all. I'm curious, that's all.

Christian


[URL=http://www.artofthegroove.com/logic/mp3/Christian_Obermaier_demo.mp3]show reel [/URL] [URL=http://uk.geocities.com/christianobermaier/home.htm]home page [/URL] [URL=http://uk.geocities.com/christianobermaier/Studio.htm]studio pics [/URL] [URL=http://uk.geocities.com/christianobermaier/Gearlist.htm]gear list [/URL]

O.K., I admit I find this company's stuff interesting, even though it is synful.

Two other observations: the VSL legato is head and shoulders over those synthesized legatos. Hardly "disconnected notes." Right now I am working on an orchestral set up which is nearly all legato instruments. So almost no notes are "disconnected."

The other is a simple fact that is nevertheless an aesthetic principle - just hearing a straight, pure tone coming right from a great musician or group of musicians is what I am after. As soon as somebody tampers with it, even in a slight way, it is no longer what I am looking for. (Unless of course I am doing music which doesn't involve those acoustic instruments.)

In conclusion, we need to exorcise these synners.

"The power of VSL compels you!... The power of VSL compels you!

Like many here, I have listened to the guy's demos and read his material. It's worth pointing out that what he's doing is really something like a fancier version of the VSL perf. tool.

I don't really think one needs to take sides; if it sounds good, I'll use it!

His examples do have nice phrasing, but his sounds are clearly not as good as VSL's to start with. What he's doing does require some sample input and analysis, wherever he gets it from.

He is not alone, btw: There are other serious developers working along similar lines.

If past experience is any guide, I expect VSL to continue developping the Perf. tool AND adding to the sample library. And let's not forget that the the more comprehensive the library, the better the data available for analysis and resynthesis.

I think the next 2-3 years are going to see very interesting developments for those of us working in the field of orchestral simulation...

Alan

I too had an e-mail and was,er, COMPELLED to hear the demos.My take is this:the solo strings are really quite astonishing,the woodwind,slightly less so.The brass was surprisingly poor:the Copland demo,with tutti brass,sounded like a synthesiser (in the worst sense) to my ears.The are no ensemble strings (they'll have a hard time cracking that nut,I think) and no percussion at all!.There's a l-o-n-g way to go with this technology,but it's very promising.

The claim to have rendered sample libraries redundant (as of now) is wild hype,but I believe,in time,it will mature into a serious composition tool.The interface reminded me of the VSL's somewhat (flattery,etc...)

@Another User said:

things are certainly moving in the direction of some sort of hybrid

Gary Garritan has said that it's coming, but other than that, how is it obvious? I'm not asking rhetorically, I'm curious to know whether I've missed anything!

My uneducated guess is that most musicians will be using a hybrid of hybrids, at least for the next few years. Syn obviously does some things very well (violin), although the demos make me wonder whether some subtle nuances aren't missing - nuances that aren't missing from VSL (especially its winds, which to me are staggeringly good). I use a Yamaha VL1, which is a modelling synthesizer (no samples), for some things. VSL has carried "the new sampling" (meaning Giga and later) farther than anyone thought it could go. If I may be excused [;)], I also feel the multiple mic position idea has something to offer as well.

As a matter of fact, there are a couple - literally a couple [:)] - of DX-7 patches that I still use.

And so on.

I just found out about Eric a few days ago and would like to offer the idea that he, along with VSL, is just raising the bar which is exciting and respectable. Of course he will apply some degree of salesmanship (as have VSL) if he is to succeed. All this is perfectly ethical and rather obvious. What isn't obvious is Eric's background and credentials. Whatever we may think of a particular marketing approach, this man has done his homework and can only contribute to the art IMHO. Is he is listening in, I hope we can make him feel welcome even if some of us disagree.

Technically, isn't this what the VSL Performance tool does? It's a piece of software designed to "synthesize" legato performance created from stringing hard coded samples together. I've been doing this in some form or another for 15 years, I don't care how user friendly or real the sounds, samplers or synthesizers get, you WILL have to take time to get the mix to sound perfect in your mind..... Unless someone creates a Holo-deck like on the enterprise....Thats different.

-Ben

Nick,

Maybe "obvious" is a bit of an overstatement. But if you've been following the progress of acoustic instrument "synthesis" (including sampling), then you will have seen a vast number of technologies being investigated over the last few years. The majority of them are closing in on some form of synthesis by analysis, which is already a hybrid. Some do the resythesis through granulation, others through additive synthesis, formant synthesis, phase vocoding, and so on... But all of them are invariably finding ways of resynthesizing the original sample, so that it can be "unlocked" with regard to frequency/time. That's the nut. Samples are locked, and the only way to break that is through some form of resynthesis (at least as far as I know!).

J.

[ps - by "following the progress" I'm not thinking of commercially available programs, but rather of various competing theories. There are .pdfs all over the web following this area of research if you look around. Pretty fascinating stuff, even if the math leaves me in the dark much of the time... I concentrate my attention on the "abstract" and the "conclusion"! [;)]]

I don't know if Synful is in it's complete stage yet, but I listened to some of the samples and they sound Synthesized. The Brass sounds pretty "synthy" but interesting, the solo violin sounds quite nice. It may be worth it just to buy it for that!

-Ben

@Another User said:

The majority of them are closing in on some form of synthesis by analysis, which is already a hybrid.

Do you have any juicy links, jbm?

hmmm... I don't really...

I found a couple of .pdfs I downloaded, at some point -- try these:

http://rubato-music.com/Media/Serra_SMS_97.pdf

http://rubato-music.com/Media/cardle_CMPC.pdf

I don't know if I can really leave these up, since I don't have any rights over them. But they were freely available when I downloaded them, so it shouldn't be a problem. It's funny... you'll notice that the first one is from way back in 97, and actually outlines the general type of synthesis used in Synful.

Anyway, there's much more out there... just Google!

J.

ps - check out this little bit of madness, while you're at it:

http://www.rubato-music.com/Media/leven.mp3

no... there's no Synful-ness in there, just pure VSL!