An AI-driven violin library - might be interesting to take a look

[funky lime]

But the idea of this video is to showcase the performance if the only input is melody - no CC controls, no keyswitch articulations, etc.

But the people who use these libraries don't use them in that way, with "no CC controls, no keyswitch articulations, etc." So it comes across as a bad-faith comparison; you are showing sample libraries in a way they aren't meant to be used, to make your product look better by comparison. Kind of like pulling a pizza out of the oven before it's done baking and then saying "see, look how bad it tastes!" (furthermore, you're comparing solo violin to a "faked" ensemble patch, which muddies the waters even more)

To demonstrate confidence in your product you could compare it to a real violin recording. But if you are insistent on comparing it to your competitors, you might use something like Straight Ahead Samples Bel Canto Violin ("Smart Delay"), or Tokyo Scoring Strings ("Intelligent Lookahead" and "Easy Artic"), which have a similar concept of "plug in the melody, and let the technology handle the articulation stuff for you behind the scenes" for a bit more of an apples-to-apples comparison.

But if the idea is truly is just "to showcase the performance if the only input is melody," then you don't even need the comparison to other libraries. Just show what your thing does. Most other library developers are classy enough that they don't compare their product to competitors' products in their promotional material; they leave that to the reviewers and users. To do otherwise smacks of bad faith, even if you claim that's not your intent.

[mcbpete]

To demonstrate confidence in your product you could compare it to a real violin recording. But if you are insistent on comparing it to your competitors, you might use something like Straight Ahead Samples Bel Canto Violin ("Smart Delay"), or Tokyo Scoring Strings ("Intelligent Lookahead" and "Easy Artic"), which have a similar concept of "plug in the melody, and let the technology handle the articulation stuff for you behind the scenes" for a bit more of an apples-to-apples comparison.

Also the very fabulous 'Bohemian' series by Virharmonic - https://virharmonic.com/ (not associated with them in any way, just a very happy customer/user back since when it was a UVI soundbank)

[Saffran]

But the people who use these libraries don't use them in that way, with "no CC controls, no keyswitch articulations, etc."

.
You can't know that. You assume it.
Not a good idea to base your knowledge about others based on your own taste.

[joe_acestudio]

But the idea of this video is to showcase the performance if the only input is melody - no CC controls, no keyswitch articulations, etc.

But the people who use these libraries don't use them in that way, with "no CC controls, no keyswitch articulations, etc." So it comes across as a bad-faith comparison; you are showing sample libraries in a way they aren't meant to be used, to make your product look better by comparison. Kind of like pulling a pizza out of the oven before it's done baking and then saying "see, look how bad it tastes!" (furthermore, you're comparing solo violin to a "faked" ensemble patch, which muddies the waters even more)

To demonstrate confidence in your product you could compare it to a real violin recording. But if you are insistent on comparing it to your competitors, you might use something like Straight Ahead Samples Bel Canto Violin ("Smart Delay"), or Tokyo Scoring Strings ("Intelligent Lookahead" and "Easy Artic"), which have a similar concept of "plug in the melody, and let the technology handle the articulation stuff for you behind the scenes" for a bit more of an apples-to-apples comparison.

But if the idea is truly is just "to showcase the performance if the only input is melody," then you don't even need the comparison to other libraries. Just show what your thing does. Most other library developers are classy enough that they don't compare their product to competitors' products in their promotional material; they leave that to the reviewers and users. To do otherwise smacks of bad faith, even if you claim that's not your intent.

Thanks for the advices and sorry for not being classy enough. But we wouldn't even thought about comparing us to a sample-based library if it was another sample-based library we're promoting.

We were not trying to compare to or discredit a specific competitor. But sample-based libraries is such a wildly adopted thing, and we're rolling out a new thing. It's kinda natural to think of comparing a new thing to a wildly adopted mainstream thing, and highlight why the new thing is different.

I'm not saying a sample-based library is supposed to be used in a melody-only way, but that's indeed why we are different, right? I admit we should've organized the video in a better way, maybe we can say "Traditional sample-based violin is super powerful and convincingly vivid and realistic if it's programmed well. That's why it helps millions of music producers in decades. However, if you don't have the time or the know-how to program it well, and just want to dump your melody to it. It won't perform so realistic and expressive. Because it's not designed and shouldn't be used that way..." Then do the comparison in melody-only scenario.

[joe_acestudio]

But the people who use these libraries don't use them in that way, with "no CC controls, no keyswitch articulations, etc."

.
You can't know that. You assume it.
Not a good idea to base your knowledge about others based on your own taste.

It might be hard to imagine for KVR folks to imagine( KVR is kinda a high-end hard-core kind of technical forum for audio engineers or generally techy people in my opinion), but after more and more user interviews I've done, I surprisingly realized that there are many producers or composers who have made music for decades but barely program a sample-based library well. Especially, at least based on the pattern I recognized, the composers in classical music world or film music world. Many of our users have asked us if we can generalize our AI vocal tech into virtual instrument space, to let them just dump the midi to it and get a decent mockup. That's kinda why we made this AI-driven violin.

[jules99]

Ask around at VI-control and you'll get a pretty solid idea, that working with articulations and other expressions in sample orchestras is exactly where many film and TV composers put in most of their work.

On the other hand, I agree that it's mostly music producers, especially those who work in more commercial genres, that mainly program MIDI and don't want to deal with finicky articulation programming. These are your target audience. These will love what your AI violin offers.

For film and TV composers, you'll want to show not how much better your product sounds, but how much faster you'll get a similar result. Because programming expressions to sound natural can take ages. But presenting your product this way, like you've done in the video, seems a bit uninformed. As good as it sounds.

[vanerio]

It sounds really amazing. In my opinion, with the rest of the orchestra in the background, it would be a real challenge to identify that it is not real. Congratulations. I miss some trills in the video!
It is a pity that it is not that affordable for hobbyists.

[_al_]

It sounds really amazing. In my opinion, with the rest of the orchestra in the background, it would be a real challenge to identify that it is not real. Congratulations. I miss some trills in the video!
It is a pity that it is not that affordable for hobbyists.

I couldn't even find a price for it!
So I guess we have to do the whole ACE Studio DAW thing to get the Violin?
Yeah, that's a bit steep.

[jamcat]

Would love to hear your thoughts.

This is of great interest to me because I do a lot of string quartet work with SWAM solo strings.

This is what I would like to see for ACE Studio to be something I would definitely be using:

First, it would need solo cello and viola, at a bare minimum. Ideally double bass as well, and also string sections.

Second, I would want to send it audio, and get back a “real” performance of my source audio. I get a pretty nuanced performance out of SWAM, but being completely synthesized, it can be a little uncanny valley. It’s hard to keep the performance sounding organic at all times. In particular getting a really woody tone where you can hear the resin on the catgut. So what I would want most is to use it as a finishing touch to turn my mock-up into a 100% real performance.