Menu
What's new
By:
Filters

VSL Dimension Strings - divisi test demo and what's your mixing strategy?

S

shawnsingh

Senior Member
I have fond memories of using VSL dry libraries - how complete and consistent the articulation sets were. Because of that consistency and dry sampled approach, it was possible to use articulation "micro switching" as a way to create character in performances, not just to select an articulation for the composition. And that made it so rewarding to use. But the dark side, for me at least, was that I never succeded in getting a satisfying mix.

For me, the challenge with mixing VSL dry libraries is that the dry recordings don't have a natural "distance" sound. It always seemed simple enough to start by reducing mid-low frequencies with gentle broad strokes in EQ. But that always led to accidentally emphasizing a lot of high frequency resonances and noises. I can't tell if those noises come from the instruments themselves or if there was something about the silent stage acoustics, or some accidental nonlinearity in the capture process - but those high frequency resonances are not just a few narrow frequencies - they span large sections of the spectrum from 1 kHz to 15 kHz. It becomes whack-a-mole with trying to reduce the scratchiness, and usually I end up with a muffled over-engineered EQ. I've always had a devilish time trying to get an EQ that achieves (a) distance emulation, (b) no high frequency scratchiness and resonances and (c) still sounding clear and bright.

Now, I'm tinkering with VSL Dimension Strings, wondering if I can try again. The divisi, string control, and legendary consistency of sampling are very attractive. So I tried again to get a clear bright non-scratchy mix. Attached demos is what I came up with so far, I think this attempt is far better than any of my previous struggles with VSL in the past. Also, I made these examples to test drive the divisi, and I absolutely love it. The sound of individual players being inaccurate (from humanize features, but also from some variation I manually recorded/edited) adds an unparalleled level of authenticity - it's so natural sounding, and makes it possible to get a whole world of tones and orchestrations that virtual instruments usually can't reach.

So yes, please give feedback, share your own demos, and most of all, can you please share your mixing strategies, if you feel you got a mix that is clear, bright, medium distance, and not scratchy in high frequencies.


Violin 1
8 players, unison and 2-way and 3-way divisi in various instrument registers
0:00 - 0:50, normal legato (0:20 - 0:40 sul-G)
0:50 - 1:10, sforzatissimo
View attachment dimension-violins.mp3


Cello
6 players, unison and 2-way and 3-way divisi in various instrument registers
0:00 - 0:50, normal legato (Sul-D is removed around 0:27, hence the brightness and intensity tone change at 0:27)
0:50 - 1:10, sforzatissimo
View attachment dimension-cellos.mp3


Violin 1 and Cello together
again with extensive divisi in various instrument registers
0:00 - 0:50, normal legato
0:50 - 1:10, sforzatissimo and a playful but satisfying grand ending
View attachment dimension-violins-cellos.mp3


Same demo but with only 3 violin players and 3 cellos, with divisi they really get exposed.
View attachment dimension-chamber.mp3


And finally, a quick test of Violins ppp-pp sul tasto legato (Dimension Strings III)
View attachment dimension-sul-tasto-violins.mp3
Click to read more...
 
shawnsingh said:
I have fond memories of using VSL dry libraries - how complete and consistent the articulation sets were. Because of that consistency and dry sampled approach, it was possible to use articulation "micro switching" as a way to create character in performances, not just to select an articulation for the composition. And that made it so rewarding to use. But the dark side, for me at least, was that I never succeded in getting a satisfying mix.

For me, the challenge with mixing VSL dry libraries is that the dry recordings don't have a natural "distance" sound. It always seemed simple enough to start by reducing mid-low frequencies with gentle broad strokes in EQ. But that always led to accidentally emphasizing a lot of high frequency resonances and noises. I can't tell if those noises come from the instruments themselves or if there was something about the silent stage acoustics, or some accidental nonlinearity in the capture process - but those high frequency resonances are not just a few narrow frequencies - they span large sections of the spectrum from 1 kHz to 15 kHz. It becomes whack-a-mole with trying to reduce the scratchiness, and usually I end up with a muffled over-engineered EQ. I've always had a devilish time trying to get an EQ that achieves (a) distance emulation, (b) no high frequency scratchiness and resonances and (c) still sounding clear and bright.

Now, I'm tinkering with VSL Dimension Strings, wondering if I can try again. The divisi, string control, and legendary consistency of sampling are very attractive. So I tried again to get a clear bright non-scratchy mix. Attached demos is what I came up with so far, I think this attempt is far better than any of my previous struggles with VSL in the past. Also, I made these examples to test drive the divisi, and I absolutely love it. The sound of individual players being inaccurate (from humanize features, but also from some variation I manually recorded/edited) adds an unparalleled level of authenticity - it's so natural sounding, and makes it possible to get a whole world of tones and orchestrations that virtual instruments usually can't reach.

So yes, please give feedback, share your own demos, and most of all, can you please share your mixing strategies, if you feel you got a mix that is clear, bright, medium distance, and not scratchy in high frequencies.


Violin 1
8 players, unison and 2-way and 3-way divisi in various instrument registers
0:00 - 0:50, normal legato (0:20 - 0:40 sul-G)
0:50 - 1:10, sforzatissimo
View attachment dimension-violins.mp3


Cello
6 players, unison and 2-way and 3-way divisi in various instrument registers
0:00 - 0:50, normal legato (Sul-D is removed around 0:27, hence the brightness and intensity tone change at 0:27)
0:50 - 1:10, sforzatissimo
View attachment dimension-cellos.mp3


Violin 1 and Cello together
again with extensive divisi in various instrument registers
0:00 - 0:50, normal legato
0:50 - 1:10, sforzatissimo and a playful but satisfying grand ending
View attachment dimension-violins-cellos.mp3


Same demo but with only 3 violin players and 3 cellos, with divisi they really get exposed.
View attachment dimension-chamber.mp3


And finally, a quick test of Violins ppp-pp sul tasto legato (Dimension Strings III)
View attachment dimension-sul-tasto-violins.mp3
Click to expand...
Shawnsingh,

What a great post with musical examples and Thank you doctoremmet for the video references. Dimension strings are a LIFETIME of exploration - each nuance a new nugget and revelation that other string libraries do not have. Reverberation is everything with VSL silent series. What reverberation are you using?
 
Great videos, thanks!

For this example, I turned off the Synchronized convolution and reverb in the Synchron player, I panned manually, and used one EQ plugin for each section plus a send to Cinematic Rooms. The reflection part of Cinematic Rooms came in handy to actually configure some early reflection presence.

I actually feel happy with the tone, but it remains to be seen whether this mix can blend nicely with other libraries I want to use for woodwinds and brass...

I'm also experimenting with other techniques:
- using the default Synchronized sound is actually very nice
- I will try MIR pro too
 
OK so after obsessing over the mix like only a hobbyist can afford, I have some information and learning experience to share, hopefully it is useful for others.

I tried:
(1) mixing with panning and algorithmic reverb, that's the version you hear in the original post.
(2) using synchronized-mix and tweaking it with EQ. It's not too bad, but I didn't prefer it.
(3) using only the convolution from the synchronized setup, but turning off the reverb and using my own algo. This was very promising, but ultimately I got MIR to sound even better.
(4) Using individual instruments placed in MIR

After shameful amounts of EQ tweaking, I was able to tame the high frequencies. But the spaciousness is much better, it especially shows when doing divisi or humanized writing.

So, here are the MIR demos, and then I'll write another few posts with details of how I mixed it so that others can reproduce it if they want. I will also share some very interesting thoughts from what I've learned during the endless hours of tweaking EQ.

Violin 1:
View attachment MIR-vsl-dimension-violins-test.mp3

Cellos:
View attachment MIR-vsl-dimension-cellos-test.mp3

Violin 1 + Cellos
View attachment MIR-vsl-dimension-violins-and-cellos-test.mp3

And finally, testing some energetic bouncy strings with some tricky ornaments.
View attachment MIR-vsl-dimension-low-strings-bouncy-test.mp3
The martele run at 0:21 is where Dimension really shines because the individual string players don't play fully accurately together, and that actually exposes more detail to be heard, rather than a diffused unison. It's hard to hear unless you're looking for it, but this is exactly the nuance I hoped to get out of investing this time, and I will definitely be trying Dimension Strings extensively in future songs to see how far it can go.

Cheers!
 
The mix is simply MIR pro followed by one EQ per section and one EQ on the overall strings bus.

I think the images got rescaled when attaching them, then you can't read the numbers. So instead, I uploaded the screenshots to this shared google drive folder (click here). If you have Dimension Strings, MIR pro, and the Synchron MIR venue, you could try reproducing this sound from these screenshots.

Some notes:
  • The EQ settings are fragile - you need to try dialing in something very close. A dB off can change the tone very drastically.
  • It might sound obvious, but it's easy to forget - the wet/dry mix, reverb tail length, and the positioning of individual players in the MIR venue can all affect the timbre. So changing them can break any EQ settings you already dialed in! I got stuck in loops tweaking these, reintroducing some frequency scratchy or resonance problems, and then having to completely rework the EQ to get the desired tone.
  • Violins and Violas EQ was to remove a few remaining high frequency problems. The Cello EQ addresses frequency problems, but also adjusted the timbre of the cellos to taste.
  • The MIR pro settings are:
    • from Vienna Ensemble, MIR pro auto loads the instrument profile for each Dimension player. If you don't use Vienna Ensemble, you may also need to manually do this for all 32 instruments.
    • Synchron room venue
    • Room eq (doesn't affect dry), see screenshot. This was the biggest factor in removing scratchiness and boominess to let the reverb feel a bit more natural to my taste
    • I chose the Stereo HOA Synchron triple-8 Veinna Standard microphone, did not tweak it further
    • Tweaked reverb time scaling and wet/dry offset for a drier sound.
Feedback and discussion is welcome.

I'll write one more post later about EQ. Cheers!
 
Last edited:
Great stuff - this what I mean about dimension strings being a lifetime of exploration. MIR Pro truly shines on dry samples when using the secondary microphone ( your example has it off). The “balcony“ mics in Room Pack 1 halls adds real depth to the samples. Happy trails…..
 
shawnsingh said:
And finally, testing some energetic bouncy strings with some tricky ornaments.
View attachment MIR-vsl-dimension-low-strings-bouncy-test.mp3
The martele run at 0:21 is where Dimension really shines because the individual string players don't play fully accurately together, and that actually exposes more detail to be heard, rather than a diffused unison. It's hard to hear unless you're looking for it, but this is exactly the nuance I hoped to get out of investing this time, and I will definitely be trying Dimension Strings extensively in future songs to see how far it can go.

Cheers!
Click to expand...
I think this string example is particularly impressive and is really crossing into hard core realism territory. great work.
 
You must log in or register to reply here.
Top Bottom