Samplemodeling Violins vs. Audiomodeling Violins
- Thread starter windplayer
- Start date
widekeys
Member
Yes, i have experimented. I have one cubase project where I am testing to mix the ensemble sound from time to time. It is great fun to play in this project - there are 18 Audio Modeling Violin instances and I have it setup to be able to easily record tutti or {A,B} or {A,B,C,D} Divisi (apart from being able to play individual solo violins).
To demonstrate my current state of mixing the ensemble violins, I recorded a short doodle for you showing off some common use cases (18 as Tutti; 8 as Solo + A + B; 18 as 4 Part Divisi Tremolo, 18 as 4 Part Divisi Spiccato)
I have a lot of pro and contra opinions about the project. The biggest pro is the fun and freedom factor. The biggest contra is actually not the sound anymore (it was in the beginning stages of the project) but the sheer computing-power required for 18 divisi AM Violins. I can not play this live with all 18 together, I do 18 as A,B,C,D divisi. Otherwise I get ASIO dropouts at 1024 Samples. My system is 2011 mid range (i7-3770K). Another contra is the amount of fine-tuning needed. Every slight change in one of the many parameters yields different results. Testing all of them requires to open and close all 18 violins and dial in the desired values. I'm looking forward to Audio Modellings ensemble instruments which are already confirmed to release in the near future. Hopefully they come up with a good idea to control the sound and to be less heavy on my system.
muziksculp
Senior Member
@widekeys
Hi, and Thanks for the feedback, and audio demo showing AM violin made into an ensemble.
They sound a bit too tight, or focused sound wise compared to how an 18 violin ensemble would sound, this is what my ears tell me right away. but the performance sounds very convincing.
Yes it would be great if AM delivers their ensemble strings versions in the near future. Although, I'm not sure if they will be released this year.
Cheers,
Muziksculp
Hi, and Thanks for the feedback, and audio demo showing AM violin made into an ensemble.
They sound a bit too tight, or focused sound wise compared to how an 18 violin ensemble would sound, this is what my ears tell me right away. but the performance sounds very convincing.
Yes it would be great if AM delivers their ensemble strings versions in the near future. Although, I'm not sure if they will be released this year.
Cheers,
Muziksculp
widekeys
Member
Tight in terms of performance? This may be due to only 4 recording takes for all 18 violins. I am not aware of an easy randomize function in cubase, that "smoothly" randomizes CC, velocity and note on/duration. As a workaround I just delayed each midi track a few milliseconds.
muziksculp
Senior Member
No, not in terms of performance, but in terms of the sound characteristic of an 18 violin ensemble, your demo sounds like a smaller violin ensemble is performing, rather than 18 players. imho. an 18 violin ensemble will have a lusher, more diffused sound, rather than a more narrow-focused sound of a smaller ensemble.
widekeys
Member
Got it. Well, an easy fix could be spreading the stereo field a little (attached file) and cutting off some high end to simulate dampening (second attached). I'll listen more carefully to live orchestra recordings, thanks for your opinion!
Last edited:
muziksculp
Senior Member
This version sounds fuller than the previous version. but not exactly what I would expect to hear if 18 violinists are playing. It's the overall richness, and character of the strings timbre, that I feel is a bit lacking. But given that these are Physically Modeled Virtual Instruments, they are quite impressive at what they are able to deliver.
Thanks
robgb
Genius. Idiot. You Decide.
Well, their approaches are likely top secret, but the way I understand it, Sample Modeling uses recorded samples as the basis of the sound and manipulates them via sound modeling techniques.
Audio Modeling, on the other hand, creates the sounds from scratch, much like Pianoteq pianos, using their SWAM technology. Some of their instruments may use samples as a basis, but I'm not sure which. I'm sure someone else will chime in to correct anything I've gotten wrong.
muziksculp
Senior Member
Maybe a bit OT, but... What puzzles me about SM technology, is that they are using Kontakt, and using a quite small set of samples to create quite expressive instruments using physical modeling, or other types of techniques, phase alignment, ...etc.
So, how do they incorporate, or integrate this into Kontakt is kind of a mystery to me. Anyone here know what's really happening under the hood ? how can they do this in Kontakt ? I would have thought they would need their own specialized, proprietary PM engine to do it.
I think their Solo and Ensemble Strings sample size is only 5 GB. which is quite small compared to traditional Strings libraries, for both Solo and Ensemble.
So, how do they incorporate, or integrate this into Kontakt is kind of a mystery to me. Anyone here know what's really happening under the hood ? how can they do this in Kontakt ? I would have thought they would need their own specialized, proprietary PM engine to do it.
I think their Solo and Ensemble Strings sample size is only 5 GB. which is quite small compared to traditional Strings libraries, for both Solo and Ensemble.
robgb
Genius. Idiot. You Decide.
It's a mystery to me, other than they must be programming geniuses.
Vardaro
Active Member
To be clear, AM uses a small number of samples in their woodwinds, but the strings use only "waveguide synthesis" (hence the "plasticky" buzzy tone?..)
Apparently their future brass will be built without samples, as well.
BTW, in the back to back impro, I found SM fast runs more convicing (as a string player!)
Apparently their future brass will be built without samples, as well.
BTW, in the back to back impro, I found SM fast runs more convicing (as a string player!)
Last edited:
muziksculp
Senior Member
According to Audio Modeling's website, they have three versions of their SWAM Engine, SWAM-B (for Brass), SWAM-W (for Woodwinds), SWAM-S (for Strings). It doesn't mention that the SWAM-W use some samples, but if it does, I wonder why ? and why the other models don't, or why the SWAM-W is the only one that uses samples ?
Anyways.. They have a new version of SWAM-B, I think ver 1.5, for their solo brass instruments, I was wondering how their Solo Brass compares to Sample Modeling Solo Brass, i.e. The Trumpets ?
Audio Modeling's Trumpets cost $250. Sample Modeling's Trumpets cost 119 Euros. which is around $138. I don't want to go off topic here, so I might post a new topic about this. unless it exists.
I'm very happy that I purchased Sample Modeling's Solo & Ens. Strings, still discovering this deep library.
Vardaro
Active Member
I too should like to hear their brass.
However as SM shows, real strings, lips, reeds and air columns do not behave as neatly as high-school physics manuals would have us believe. Real instruments produce a mess of random distortions, pitch-bends, and inharmonicity, particularly at note onset.
However as SM shows, real strings, lips, reeds and air columns do not behave as neatly as high-school physics manuals would have us believe. Real instruments produce a mess of random distortions, pitch-bends, and inharmonicity, particularly at note onset.
Bollen
Vintage Member
I needed a day off and ended up getting myself drunk... In any case, I love this type of posts, they are my favourite! So I thought I would contribute to this comparison with a bit of VSL solo, 1st chair Dimensions and Chris Hein's...
Oh and by the way, I am a totally converted SM (or even AM) devotee...! (although I don't have either..
)
Oh and by the way, I am a totally converted SM (or even AM) devotee...! (although I don't have either..
)
widekeys
Member
Thank you for contributing to the demos! The VSL Solo was overall the most convincing of the three libraries, at least to my ears. Still, to me the AM Violin has the cleanest blend between the different articulations and dynamics (which seems obvious since there really is just one playable "articulation" that is continuously morphed using CCs).I needed a day off and ended up getting myself drunk... In any case, I love this type of posts, they are my favourite! So I thought I would contribute to this comparison with a bit of VSL solo, 1st chair Dimensions and Chris Hein's...
Oh and by the way, I am a totally converted SM (or even AM) devotee...! (although I don't have either..
Bollen
Vintage Member
Yeah I agree, however... Living with a violinist has tempered my judgement and find AM to be too "keyboardy" or as you say "clean blend of articulations and dynamics", so much so that it just doesn't sound 100% real, more like 85-90. The real violin has tons of artefacts and imperfections that I think SM gets closer...
On another point, the OP said these were improvs, it took me all bloody night to program these demos... I'm a total convert of either, I don't care, I just don't want to spend the rest of my life programming things...
@Bollen you made this thread resurface and like you the subject interests me.
I would like to add a challenger to these comparisons, but come back to the physical modeling side with Reason Friktion.
I've talked a lot about this instrument lately, and I often compare it to Audio Modeling strings, I'd like to see if Friktion can rise to the ranks of the other 2 tenors.
Can someone give me the midi file to rework it for this plugin?
For my part I own the Sample Modeling S&E Strings, and I admit being a little disappointed.
I was hoping to be able to replace my other two libraries (Embertone Intimate strings, Chris eh Ensemble Strings), but I couldn't beat these two in terms of sound with SM (especially in an ensemble).
After seeing the demos that some do, it's still a powerful plugin, but as far as I'm concerned, I don't know how to tame it.
In comparison, With Friktion the sound immediately explodes in my ears, I have the impression of being a volonist with the possibility of playing with all the aspects that make up my instrument.
It is more complex to master because it is a synthesizer whose base are the strings, and can easily leave the register of traditional strings.
Besides, we can create other equally realistic instruments.
In addition this instrument works under Reason Rack, so you can use the modular rack interface to create an ensemble for example.
I combine several strings in a rack, with different effects to humanize everything, I would miss more than a midi modude for divisi inside the rack and that would be perfect, but unfortunately this module does not exist.
Here is an example of my cooking on a midi file without any CC, just a placement in space and a reverb with Eareverb:
I would like to add a challenger to these comparisons, but come back to the physical modeling side with Reason Friktion.
I've talked a lot about this instrument lately, and I often compare it to Audio Modeling strings, I'd like to see if Friktion can rise to the ranks of the other 2 tenors.
Can someone give me the midi file to rework it for this plugin?
For my part I own the Sample Modeling S&E Strings, and I admit being a little disappointed.
I was hoping to be able to replace my other two libraries (Embertone Intimate strings, Chris eh Ensemble Strings), but I couldn't beat these two in terms of sound with SM (especially in an ensemble).
After seeing the demos that some do, it's still a powerful plugin, but as far as I'm concerned, I don't know how to tame it.
In comparison, With Friktion the sound immediately explodes in my ears, I have the impression of being a volonist with the possibility of playing with all the aspects that make up my instrument.
It is more complex to master because it is a synthesizer whose base are the strings, and can easily leave the register of traditional strings.
Besides, we can create other equally realistic instruments.
In addition this instrument works under Reason Rack, so you can use the modular rack interface to create an ensemble for example.
I combine several strings in a rack, with different effects to humanize everything, I would miss more than a midi modude for divisi inside the rack and that would be perfect, but unfortunately this module does not exist.
Here is an example of my cooking on a midi file without any CC, just a placement in space and a reverb with Eareverb:
Bollen
Vintage Member
Hi Lychee! Yeah, there's nothing quite like listening to the same musical extracts played with different libraries... Really shows the virtues and flaws!
I don't have a MIDI file, I just transcribed it and after I finished I deleted the whole project because I'm really pressed for space in my HD...
lettucehat
Senior Member
Wow, Sample Modeling could really take a page from this library in terms of its sound... the scratchiness is really convincing. I imagine that if SM could improve the base tone of the ensembles in this area they would immediately make the SM ensembles much more usable for shorts.
