What's new

Sample Modeling has been busy....

It sounds like the bit-rate for the video has been compressed and affected the audio? Or maybe they exported the video at the wrong settings. Can't quite put my finger on it. Have to wait and see..
 
I'm Curious how you'd compare the tone here with your experience of CH solo strings?

It's difficult to compare from just that one demo. I love SM's horns but I think instruments regularly played with vibrato are trickier to model (for lack of better term, I know samples are in use here) and maintain expression. I'd love to hear more examples particularly in a solo context. This has me very interested overall.

Guys, just to be clear, it's actually NOT physical modeling, and has nothing to do with SWAM or AudioModeling at all (they separated careers long time ago). It's actually the Samplemodeling last generation of Kontakt Virtual Instruments: if you are curious about technology, just read the information on the producer web page.

P.S. for detractors, of course they are always welcome to show if they can do it better with their tools and their skills. We all just aim to improvement, and learning. :)

Distinction noted, and I think that's why I prefer the tone here so much. These sound good based on the limited amount I've heard.
 
  • Like
Reactions: ism
It sounds good on it's own. However, if compared to the original live video, it's worlds apart in basically everything. I thought it was a bad move on Samplemodeling to use this live performance. They would have been better off using the library with an original performance. This actually works against them I think.
 
It sounds good on it's own. However, if compared to the original live video, it's worlds apart in basically everything. I thought it was a bad move on Samplemodeling to use this live performance. They would have been better off using the library with an original performance. This actually works against them I think.
I have to agree. I was quite liking the emulation until I listened to the original. The actual performance they are emulating is a real barn burner, and very nicely recorded. We're a long way from capturing that with VIs. Overall, there is a brightness and blistering attack in the original performance that is lacking in the emulation. You can also hear much more discernible body tones and bow rosin in the original. Truly great string performance emulations, particularly solo and chamber, are a tough nut to crack, but good on them for trying. Tone is everything though.
 
"A paradigm shift for virtual instruments where realism and playability share equal value."

Sorry I don't share your excitement, this is still too synthy... modeled strings just don't work for me.

The original performance

Wow, thanks for posting that. Didn't want it to end. The Four Seasons tends be treated like wall paper these days, but it really is remarkable music, especially when performed like that.
 
Despite falling pretty far short of the outstanding original performance, I’m still interested to see and hear more.
 
It sounds good on it's own. However, if compared to the original live video, it's worlds apart in basically everything. I thought it was a bad move on Samplemodeling to use this live performance. They would have been better off using the library with an original performance. This actually works against them I think.

Well I think there's a misconception here. And the fault is for sure of the wrong statement into the Web Site page: this was NOT intended to be a simulation of the original at all. Personally I think the original is often too dark and noisy, due to the style (very staccato) and the sound of the original strings. The reason to show the Video was to connect gestures to sound, so you may realize what fingering, bowing and position is supposed to produce the flawless variation of articulations, and not making an acoustic copy of the original at all.
 
I have to agree. I was quite liking the emulation until I listened to the original. The actual performance they are emulating is a real barn burner, and very nicely recorded. We're a long way from capturing that with VIs. Overall, there is a brightness and blistering attack in the original performance that is lacking in the emulation. You can also hear much more discernible body tones and bow rosin in the original. Truly great string performance emulations, particularly solo and chamber, are a tough nut to crack, but good on them for trying. Tone is everything though.

I was thinking about this, and it could also very well be whoever did the mockup. I hear things in the mock-up dynamically they could’ve been far better executed but maybe it was the time constraints to get the product out?
 
I can hear some artifacts and overall the dynamic range is clamped, but the sound is very promising, as an owner of the original stradivari and godfriler I want to believe they will make a hit with this installment. I own the now audiomodeling strings but I am not satisfied with their tone.
We will have to wait and see.
 
"A paradigm shift for virtual instruments where realism and playability share equal value."

Sorry I don't share your excitement, this is still too synthy... modeled strings just don't work for me.

The original performance


Well hearing this and comparing, seems like they are worlds apart, seems like the SM muckup is missing all of its velocity as a result your missing the different tones the bow is making fingered against the fret board,

Well the library is coming soon but we will need to see if this character can be developed.

the library sounded ok but not sure if the world want just another string library, so i hope SM kick ass with this one
 
i think the only remedy to the uncanny valley for these libraries is read-ahead function on all the libraries. no doubt this is system heavy, but what the recreation lacks is anticipation of the next attack. Theres a forward motion that creates a sense of energy and anticipation in the original piece. a library generally has individual notes with no direct context and connection to the notes before and after it. If more libraries employed something like what audiobro's choir vst does, perhaps it could better find the right sample to simulate natural motion.
 
i think the only remedy to the uncanny valley for these libraries is read-ahead function on all the libraries. no doubt this is system heavy, but what the recreation lacks is anticipation of the next attack. Theres a forward motion that creates a sense of energy and anticipation in the original piece. a library generally has individual notes with no direct context and connection to the notes before and after it. If more libraries employed something like what audiobro's choir vst does, perhaps it could better find the right sample to simulate natural motion.

Not familiar w the Audiobro choir. How would you organize this? Something like legato transitions of different intervals have been around since VSL legato or maybe before. But that just accounts for two notes really.
 
First off, they will never get any string library to sound exactly like the real thing. So let's quit nitpicking as if they can.

Second, sample modeling, as I understand it, uses recorded samples and works their magic with them, as they've done with their brass. There is some modeling involved, but utilizing recorded strings. If that's the case here, then it makes little sense to complain about modeling.

Third, instead of comparing these to real strings, why don't we compare them to other string libraries. And in that case, these strings sound superb
 
i think the only remedy to the uncanny valley for these libraries is read-ahead function on all the libraries. no doubt this is system heavy, but what the recreation lacks is anticipation of the next attack. Theres a forward motion that creates a sense of energy and anticipation in the original piece. a library generally has individual notes with no direct context and connection to the notes before and after it. If more libraries employed something like what audiobro's choir vst does, perhaps it could better find the right sample to simulate natural motion.

Well, a part for predicting the future, that is impossible for real time systems (it is for tools pre-reading a score for instance, like Finale Human playback or Wallander Note Performer), taking into account what was coming before, is already done by almost all the modern virtual instruments and sample players.

Actually Samplemodeling has one of the most accurate and sophisticated scripting for real-time sound shaping combining your input (via MIDI data/controller) with the musical context extrapolated by relation with previous note (staccato, legato, speed, bow change etc.). It's not really about individual notes, or what you listen in the demo should not be possible, even with massive editing.
 
Third, instead of comparing these to real strings, why don't we compare them to other string libraries. And in that case, these strings sound superb

I feel they sound ok, i am hoping there will be a bit of magic like there brass and audio modeling, the more velocity pressure you apply to the string you can here the tension change and the string rubbing against the fret board you can find this in audio modeling, i have also heard this in OT B Strings, it gives the impression that someone is playing a violin and not using a keyboard to make the sound of a violin and just going through the notes.

I think S M did amazing with their Brass I think its not so much complaining its more knowing what S M are capable of especially in realism and authenticity, if the one doing the muckup was skill full you found it extremely difficult to tell if a real brass was being used or not.

So the sound was ok but realism is their game
 
I didn't know about this. This evening I will listen at the demo.
 
Do we have a rough idea/feel for how the ensemble side of things will work? e.g. how large the ensemble will look to be?

I'm assuming that it'll be a case of SM layering single instruments to 'create' an ensemble, rather than recording an ensemble?
 
Samplemodeling don't do release dates. They only announce stuff when it's ready (or close). It saves people whining and nagging that they have missed deadlines and allows them to get on with the job.

Although people still tend to whine and accuse them of being too quiet and going out of business. *facepalm*

EDIT: Actually it seems they did hint at a timeframe to placate their userbase.
"Much remains to be done, to perfect these instruments. However, we expect to enter the beta phase in about one month."

https://forum.samplemodeling.com/viewtopic.php?t=39609
 
Last edited:
Top Bottom