Yes. It looks like it's true. I was a mastering engineer for 6 years and participated in many court cases where I did audio analysis.

In the image below, the first 2 images are of the stock 4sec pfp Bassoon sample L and R.

The second 2 green images, are timecompressed from the stock 6sec pfp Bassoon L and R samples to match the same 4sec length.

http://evanevans.dyndns.org/E3_Studio/Macintosh%20HD/Users/evanevans/Documents/Bengt%20Nilsson/Timestretch_Evidence.jpg">

Now the process of timestretching and compressing is an intelligent one so you do not always get symmetrical placement linearly of the waveform crests, however, the most fanciful and capable timestretching algorithms do an excellent job of preserving all the crest's phase orientation. That is they preserve the voltage positive and negative distances vertically.

Based on these images I could certify that both the 6sec and 4sec samples are from the same recording.

Evan Evans

But what happens to the sound quality, especially when so much care has been taken to record everything 24/96 ?

Just curious to know the truth about this stretching rumour..

A picture is worth 1000 words. This appears to be no rumor.

Evan Evans

The other important question to ask is: "Does this deteriorate the sound quality or not?"

na i don't think so, the time-streching is well-done.

I just don't like that this wasn't disclosed. I mean this is supposed to be a library of "samples" not contortions. I feel a little bit jipped by it. Well actually, I don't feel jipped, but it makes me less trusting of what is found in this library.

For me there are now two large wastes of space/time in the VSL orchestral libraries. Because the repetitions are repeated, verbatim samples and all, the libraries are exactly 33% bigger than they need to be. So teh 222GB Pro Edition needed to only be 167GB. ... And now this. More duplicate data.

I guess you could say that 150GB of Pro Edition is easily worth the huge price tag, but it still doesn't sit well that we pay for around 40% of stuff that is nothing new , duplicative, and a waste of space.

Out of curiosity, could some kind of real-time time-stretching tool be innovated to create dynamics of many lengths during playback? Sounds pretty space age and CPU intensive, but that's probably what people would have said about having 200,000 samples of the orchestra a few years ago (ok, 10 years ago at least).

Evan Evans

Funny thread,
some informations:

There is no sampler engine, which does realtime stretching in a acceptable quality at the moment.

But there are some really good stand alone programms out there.
One of our major concerns is, that high quality realtime stretching will be part of future sampler engines.

Therefore we are searching and developing in this direction. Part of this developement was an assignement with Peter Neubäcker (Melodyne) to develop a dedicated VSL Melodyne version.
If I remember right, there was more than 1 year developent time for the final algorythm and all the engine features.

So back to the recordings - the dynamics:
First, it's impossible for the musicians to perform all dynamics with 100% tight length. There are always little deviations. For woodwinds and brass samples we got so good results with Peters engine, that we were able to offer 100% length optimized samples. You can correct always in two directions, therefore for brass and woodwinds do have more than one optimised versions of one sample.

So here you've got the results of a high quality developement, which makes the library more flexible.

So please, Evan: there are very often incriminations in your postings, some time more subtle, sometime very dumpy:
Of course I'm not very pleased about such statements.
So I would like to ask you for more fairness in your postings.

Keep in mind, that this company is not "Microsoft". All people here are passionate musicians and technicians, and all togehter we are trying to develop the ultimative composing tool of the 21st century.
That's all.

thanks a lot
Herb

Mr. Evans -

Lucky you that Herb is such a patient character.

Personally I'd say your tone is _way_ off an accaptable limit. I can't find this funny anymore. Allthough I know that one shouldn't "feed the troll", I'll do you the favour and give you the attention you're obviously longing for:

First of all: Although we accept controversial topics liberally, this is still a private forum (or should I say: not _your_ private forum).

Your messages show that you are not able to accept (or not even considering) another truth than yours. You can't even wait until all facts are evident. That way, you give away the chance to be taken serious (which is a shame as sometimes you seem to have impressing in-depth knowledge). What's more - you not only show that you didn't fully understand certain aspects of the topics you write about, but you cause serious troubles for others: for us, because you publicly run down our products out of pure ignorance; for our costumers, because they could gather misinformation from your messages.

On the top of it, you got more feedback from the whole crew (barring your fellow forum-members) than anybody else. Reason enough to hold back your horses a little bit from now on, I'd say.

I'm _not_ writing in the name of the Vienna Symphonic Library; but as someone who tries to keep this forum a helpful, friendly and vivid place each day, and as someone who puts his heart, soul and lots of work into this remarkable project --- I'm fed up with your arrogance and your know-it-all attitude.

I want to make a very clear statement: As much as I appreciate the fact that you have strong opinions and don't hesitate to express them, I will not tolerate another verbal faux pas from your side. The next time I will ask our webmaster to initiate precautions to have your mails reviewed by a moderator before they are posted *).

Get a live and give us a break. And please: Do yourself and us a favour - don't post an answer.

/Dietz

*) ... oh, BTW - I am a moderator ... [8o|]

Phew, that feels better now :-]

/Dietz

Makes me feel better aswell.. thanks [:D]

@Another User said:

Therefore we are searching and developing in this direction. Part of this developement was an assignement with Peter Neubäcker (Melodyne) to develop a dedicated VSL Melodyne version.
If I remember right, there was more than 1 year developent time for the final algorythm and all the engine features.

Wow, this is the first I've heard of this, and I would say a very exciting potential development. I've thought for a year now that a marraige of VSL and Melodyne, if reasonably easy to use, would be the way to go. It'll do wonders for the dynamic samples in particular as well as allowing short notes that are just the right length. Very nice.

@Christian Marcussen said:

Makes me feel better aswell.. thanks [:D]

Same here - keep up the good work and spirit, guys.

Peter
(happy new Opus 1 owner [:)] )

[:D]

YEAH....

Do not say everything you know, because you may look like a fool...... [[;)]]


Iwan

Dietz,

It looks like my observations turned out to NOT be out of pure ignorance. As always.

And Herb,

Thank you for your response. Don't worry, like I said I still feel I got what I payed for. But it was and is a little startling when you think you have in your hands performances of 2s,4s,6s dynamics when that's not true. many times your employees and yourself seem to find any negative comment about your library as "dumping". Most other companies would take them as recommendations on future policy and public opinion. But time and again we get a kind of elistist response from you guys when someone says something sucks or something's wrong.

In all honesty your library ain't flawless. And I would hope in the future you would welcome all the reactions from all your customers.

If any customer's reaction isn't easily overturned by VSL than I'd say there's a reason to listen and not to fight. Here you have validated the customers concern. This is reason for apology not for telling the customer that they are out of line. Although seemingly outlandish, my thoughts and reactions to this library are bleeding edge. Others share my viewpoints or will sometime soon. These concerns I seemingly quibble about are goin got become all that users care about in the future.

My apologies for casting any bad news, but it's based on the merits and public opinion.

Evan Evans

I just talked to a friend of mine who is a very high up actively working film composer in Hollywood. I asked, "What do you think about that VSL uses time stretching to create their 2s and 6s dynamics from their 4s dynamics?" I was very straightforward without any emotion in my voice so as to get an unbiased response.

Well, he nearly flipped off the handle. He said that is a complete misrepresentation and other users should likely feel that's not fair. I also left out some colorful words he said.

So that's 2 out of 2 in my camp (myself and him). But I'm sure my colleagues would take this the same as I do.

I really don't need to feel vindicated here though. I trust my own opinions well enough. But so far if a poll were to be taken my guess is that this, although a small nearly inconsequential issue, reflects poorly on the representations and merits of the library.

But in VSL's defense, I LOVE the library, and we are only talking about perhaps 1% of the library here ... right? [[:|]]

Evan Evans

I don't care about that time-stretching, we only discovered that "superchérie" now and we didn't noticed it before , so ... should we care?

@evanevans said:

[...] are exactly 33% bigger than they need to be. [...] that we pay for around 40% of stuff that is nothing new , duplicative, and a waste of space. [...] and we are only talking about perhaps 1% of the library here [...] Evan Evans

Now what? 33%? 40% Or 1%? You can't distinguish between a "kind of elistist response" and a completely stunned reaction to some incoherent allegations in a company-owned forum which has proven many, many times to be very open to critics, and even praise for seemingly competing products. Although you're as powerful as 50 users (or was it 60?), you destroy the vibe _and_ the sense of this board for several hundreds (or thousands) of others. /Dietz

to add an interesting point to the discussion, i took the liberty to layer the two graphs of the left channel
http://www.vsl.co.at/upload/forum/Timestretch_Evidence.gif">
so what does that say to you [;)]
christian

I would imagine that getting the original sampled performers to play the exact same way for similar articulation samples would be near impossible. I imagine that getting each tone sample to match the rest is a feat on its own. The consistency which allows these samples to appear to us as "instruments" seems to me to be a great effort. That is what we paid for.

What draws us to this product is it potential for realism in the sound of the final audio. Personally, I don't care *how* they achieve it. If a competitor of VSL figures out how to do it better -- maybe without timestretched samples -- guess they will be able to charge more. God bless them!

I think what we have here is a matter of marketing seeming a touch ahead of technology. VSL brags about the samples, the silent stage, the "real" elements used, while saying little about the special tricks and audio blending involved. Marketing-wise, this makes sense for this kind of product. We want realism and would be jarred by seeing "timestretching" in the marketing.

But, heck, we *know* that the samplers we play these samples through use crossfades and envelopes and all manner of "synthetic" elements to make this all blend. And these forums are riddled with the digital advice of how to make things sound more realistic -- reverb settings, new instrument programming, etc. Basically: whatever works to achieve the end result.

Herb & co. : You used a synthetic method to achieve a certain result. This is not in your marketing anywhere. But -- oh wait -- these are samples, playing through samplers... on computers, no less!

And darn it, it works!

For my money, I am happy.

When you -- or someone else -- figure a way to make it better, you will get more money from me. [:)]