What's new

Real Talk On Sample Modeling Brass

aaronventure

Senior Member
One cool trick I do for SM is add a random LFO modulator to CC1 with a very small range (~2 units +/-) to simulate imperfect flow of breath. Unnoticeable on fast and loud marcato lines, pretty impactful on slower lines. It's not something you can just pull of on the modwheel. So while the modwheel input are clean curves, what's actually heading into Kontakt is a bit more jaggy and rugged.
 

Geocranium

Active Member
Best use of SM brass I ever heard is Daniel Beijbom's "Boss Battle"

https://m.soundcloud.com/danielbeijbom/boss-battle
I've heard this track before! Very impressive. Honestly the strings are the most impressive thing about it. Very expressive. I must say, this is the only posted example in the thread so far where I wouldn't have immediately said that I was hearing SM brass. It's the only one that comes close to the SampleControl videos.

I think SM brass is an incredible instrument capable of incredible sounds. Their technology is second to none, and I'm excited for any future releases they have. My only contention is that I would give a newbie looking for a brass library a fair warning about how difficult it is it achieve those trailer bwams and epic FFF passages. Takes a lot of finesse.
 

Saxer

Senior Member
One cool trick I do for SM is add a random LFO modulator to CC1 with a very small range (~2 units +/-) to simulate imperfect flow of breath. Unnoticeable on fast and loud marcato lines, pretty impactful on slower lines. It's not something you can just pull of on the modwheel. So while the modwheel input are clean curves, what's actually heading into Kontakt is a bit more jaggy and rugged.
How do you set the LFO in the SM instruments?
 

Casiquire

Senior Member
I think SM brass is an incredible instrument capable of incredible sounds. Their technology is second to none, and I'm excited for any future releases they have. My only contention is that I would give a newbie looking for a brass library a fair warning about how difficult it is it achieve those trailer bwams and epic FFF passages. Takes a lot of finesse.
While I do agree with your point in general, I think it's worth noting that the same warning should be given to anybody who wishes to do the same using real instruments. In my opinion, samples shouldn't be used because they're easier than real instruments, but simply because they're more accessible.
 

aaronventure

Senior Member
How do you set the LFO in the SM instruments?
I work in Reaper so I don't know how other DAWs do this (I know Ableton can't, or couldn't in 9.7). ReaControlMIDI, add a slider for ModWheel, Param>Parameter Modulation, click on LFO, set it up, and click on MIDI link below. Now, I actually use a CC mapper to transform my CC1 into CCX (40 or something unused by SM) before ReaControlMIDI, and then assign that to the source control in the Parameter Modulation. To sum it up, my modwheel transforms into CC40, CC40 controls the baseline for the LFO, LFO does its work on the "modwheel" parameter and the result exits ReaControlMIDI as CC1.

I also do a similar thing for Spitfire strings since they lack inflections, only I create an envelope based on sidechain input which is a hihat playing every time I play a note. That way I don't have to do those miniscule CC1 edits every time I move a note - it's done automatically. I simply use it like in CSS, for swells and overall dynamic changes, and if I don't want the inflection or want it to be slightly different, I edit that small part. Hihat volume has a small random LFO on its volume so that the inflection depth is slightly different every time. I also have the sustain pedal mapped to HiHat volume (actually it's remapped 64>65 because 64 will sustain notes) so holding the sustain pedal simply bypasses the whole effect.

All done using standard (ReaControlMIDI) and custom (CC Mapper X) JS effects in Reaper, and the performance hit is literally 0.

This saves a shit-ton of my time and I'm able to work much faster.
 

germancomponist

Senior Member
Basically, SM's instruments require prior knowledge of how sound is generated, how instruments / sound waves are absorbed by rooms, and how to operate a mixer to make it sound good. Every sounding engineer used to have that knowledge, and if you do not have it, there are many offers on the Internet to learn it. If you master that, then you will fall in love with the instruments of SM!
 

dgburns

summer of pickles and IPA beer
Sample Modelling Brass is crap.

Do not use it on it’s own, it will sound realistic. Do not add reverb, for it will only make it sound better. Do not work the parameters, for it will only make it sound almost as good as a real trumpet, in the right hands. Do not, for Pete’s sake, use the mutes, they just have no equivalent.

And whatever you do, do not take your time and divisi the playing across a full Sample Modelling brass section, cause it might just sound so close to the real thing, as to make you think you don’t need the real thing.

Like I said, they suck ;)

And whatever you do, do not blend with any other library, it might actually sound really good. You’ve been warned !!
 
It's interesting to see since the company split which products are Audio Modeling and which are Sample Modeling - to me the brass (which is Sample Modeling) sounds head and shoulders more realistic than any of the other libaries, which leads me to think the modeled-from-samples approach is what works best. Just hope the fact that they've split up doesn't mean the expertise to make something great is divided between two companies.
 

Raphioli

Active Member
Best use of SM brass I ever heard is Daniel Beijbom's "Boss Battle"

https://m.soundcloud.com/danielbeijbom/boss-battle
First time I noticed this thread.

And wow! Amazing piece and use of sample modeling.

I knew sample modeling sounded great in good hands, especially the trumpets.
But to me, it just looks too hard to use. (difficulty seems its at another level compared to regular samples)
Not just programming wise, but trying to place the instrument, like pushing it back in the hall.
So I haven't bought any sample modeling stuff (or completely dry libraries in that case for orchestral stuff).

Anyways, hats off to Daniel. (and thx for the bump. I probably wouldn't have noticed it without it)
 

Henu

Senior Member
I love everything else on the SM instruments, but I just can't get the horns to sound anything like that. :(
 

Henu

Senior Member
Salt, meet wounds. :cautious: In all seriousness, I think one of my problems is going too extreme with the mod wheel. While it sounds good with the trumpets, it sounds horrible when cranked on horns. That, and the dreaded dead space. The ER's are a complete bitch to get sounding right.
 

PerryD

Active Member
Salt, meet wounds. :cautious: In all seriousness, I think one of my problems is going too extreme with the mod wheel. While it sounds good with the trumpets, it sounds horrible when cranked on horns. That, and the dreaded dead space. The ER's are a complete bitch to get sounding right.
As a "real" brass player, I love using a breath controller with SM brass. Dynamics feel natural to me.
 

Henu

Senior Member
That's something on my "to do"- list. I've been owning a breath controller for two years and just a couple of weeks ago I bit the bullet and practiced using it. Would you care to tell me what sort of curve are you using? I have a bit troubles to find something that feels natural- it's either too soft or too loud.
 

Rilla

New Member
I've been studying the trumpet lately.

The trumpet has 15 Controllers (PB + CC's) and 25 keyswitches.

Most of these 40 parameters aren't needed in everyday usage, but it's better to have these options and not need them than to need them and not have them.

Samplemodeling has filled in the void of parameters I always felt were missing on sample-based, non-modelled instruments. So while there is a hefty amount of tools to use, that only helps get the exact results you want. This thing is really thorough.

Also, I don't get the complaint about the anechoic sound. I consider it a blessing. Sound happens in 3 stages, 1) direct sound 2) early reflections and 3) predelay/late reflections. Samplemodeling provides the direct sound, a reverb provides the early reflections and late reflections (Early=less than 30ms, Late=more than 30ms). Recording a trumpet in an ambient room and then auxing it to Altiverb = 2 sets of ER's; the ER captured in the trumpet mic and the ER captured in the impulse response. The most accurate results for realism would be to record anechoically and then add convolution reverb (ultrarealism=multi-convolution). And when I do that it sounds fantastic to me. Maybe people are used to working with samples recorded with ER's and then further adding reverb to that. But imo, recording anechoically is the most logical means to getting the most realistic results. You can't remove an ER once it's been recorded. But you can always add ER's to an anechoic recording.
 
Last edited:
Top Bottom