extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Discussions about our next-generation scoring application, Dorico.
Rob Tuley
Grand Senior Member
Posts: 4095
Joined: Fri May 20, 2016 12:41 am

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by Rob Tuley »

Dorico must be sending something to NP that switches off the delay for note entry.

If you edit the score with note-echoing on while NP is playing it, there is a 1-second delay on the notes you edit.

Rob Tuley
Grand Senior Member
Posts: 4095
Joined: Fri May 20, 2016 12:41 am

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by Rob Tuley »

Brian Roland wrote:
Thu Jul 23, 2020 8:28 pm
Unless NP has added a lot of things like marching and jazz brass, outdoor battery and pit percussion, weird stuff like flugal horns and modded trumpets/bugles, extensions throughout entire woodwind families like all the saxes, alto and contra flutes, every measure of alto and bass clarinets, Orff and Kodaly instruments for children's music, etc....it's not what I require on my home rig at this time.
I have no idea how good it is for marching bands and jazz, but one nice feature of the modelled instruments is that there are no limitations on the range.

So if you have a scordatura violin with the bottom string tuned to F (and even Haydn did that sort of thing!), a natural horn part in Bb or C alto that goes too high for a modern orchestral French hornist, or a Wagner bassoon part with a low A extension, NP "just works" with no tweaking.

FWIW it has patches for Eb, bass and contrabass clarinets, basset-horn, saxes from sopranino to bass, recorders from Garklein to contrabass, piccolo and bass trumpet, flugelhorn, alto/tenor/baritone horns, alto and bass flutes, etc.

If the Orff instruments are basically standard percussion with different ranges, they will play back, though maybe not a realistically as you want. It doesn't have an option to specify the player's age when "randomizing" playback for example ;) )

IMO the weakest instrument in NP is the piano, which IMO sounds like a completely worn out Bosendorfer (and having played one or two Bosendorfers that weren't worn out, if somebody gave me a Bosendorfer grand for free I would sell it and buy a piano instead...)

User avatar
gdball
Member
Posts: 222
Joined: Wed Sep 18, 2019 10:51 pm
Contact:

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by gdball »

Rob Tuley wrote:
Thu Jul 23, 2020 10:29 pm

I have no idea how good it is for marching bands and jazz, but one nice feature of the modelled instruments is that there are no limitations on the range.
Definitely useful (to a point) for some Jazz Trumpet parts, but yah not exactly how it would sound coming from Mr. Ferguson or Mr. Gaylen Jeeter.
Dorico Pro 3.5+; EWQL Platinum; Windows 10 laptop 12 core SSD 16GB Ram; Studio One Pro v5; Scarlett 18i20;

robjohn9999
Member
Posts: 771
Joined: Tue Feb 28, 2017 7:58 pm
Location: greater NYC region
Contact:

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by robjohn9999 »

I was presuming this was an issue but I, too, haven't noticed anything (but had also switched to using an external piano sound with MIDI thru off just in case) - I'll have to try it again. re: Note Performer - to me it's a no-brainer sort of purchase since it can be used in combination with other things at the same time (if it's otherwise missing a particular sound) - for example, for jazz I typically use it (with a bit of tweaking to get around the fact that it's more classical-centric in orientation), and then use things like Superior Drummer 3 for drums instead, as well as Trillian for bass and my Ivory 2.5 piano, etc... But still nice (and affordable) tool in my arsenal for the reasons already sighted (most importantly it's ease of use for what you get in return)...
- D.D.
2019 MacBook Pro with 64 GB ram; Apogee Ensemble 88; Mac OS Catalina 10.15.2

Brian Roland
Senior Member
Posts: 1104
Joined: Fri Jun 19, 2015 5:24 pm
Contact:

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by Brian Roland »

I'd think the better latency test would be:

Do some instruments in the score with end points to NP.
Do some others using end points to some other plugin, such as HALion SE.

Play the score...

Are the HSSE tracks ahead of the NP ones when playing the score?

If NOT, you're good to go! Mix and match anything you like with NP :)

It seems to me that if NP is sorting out the difference between all those short bow techniques, and you aren't having to manually plug that kind of thing into the score rather specifically, then it's most likely buffering and analyzing it in some way to make such decisions.

Take the dotted (staccato marking) on a note that lives under a slur example.

Should it use sautellie, martele, spicatto, or staccato?

Is there something specific you must do to the score, or is it fully automatic? Or does it just use a universal short bow sound for all of the above?

Example:
In the Zortzico score, I specified that I want legato.staccato notes to inform HALion to pick the martele sound.

In some expression maps I did for GPO, I needed an assortment of things for the same score markings for different phrases of the piece. With Dorico 2, I had to set up some sticky nodes, and a lot of redundant entries to get it working. I.E. I made a playing direction for martele, and another for sautellie that stick until an ord/nat is sent. I had entries in the expression map like:
natural
natural.martele
natural.sautellie
natural.martele.sautellie
accent
accent.martele
accent.sautellie
accent.martele.sautellie
staccato
staccato.martele
staccato.sautellie
staccato.martele.sautellie
legato
legato.martele
legato.sautille
legato.martele.sautille
legato.staccato
legato.staccato.martele
legato.staccato.sautille
legato.staccato.martele.sautille

Since I was using the martele and sautille nodes in a sticky form, quite a few of the listings above actually sent the regular nat/arco instruction, or nat/acro plus the legato pedal. The format just made it easy for me to enter, or import scores and establish a default to use a particular short bow 'dot' translation in a given plugin until the node was dropped or changed.

So, in a fast piece, I could have all staccato notes NOT living under a slur mark played martele, and staccato notes living under slurs played sautille, just by setting both martele and sautellie as sticky 'direction' type playing techniques at the start of the score, and NOT have to go through the thing and individually tag all the the notes in the score I want to use these such conditions. In this case the two sounds triggered for short bows end up being:

staccato.martele.sautillie (martele key switch here)
legato.staccato.martele.sautille (sautille key switch here)

Those martele and sautellie nodes stick around until a nat/ord technique clears them.

In version 3, we got the ability to make our own techniques and playing nodes. I could ditch the sticky concept, and attach the relevant key-switch to a custom glyph of a dot that looks just like the regular staccato, and apply that to notes instead while entering the parts. Don't need all the redundant entries anymore...can just do one straight 'martele' and one straight 'sautellie' entry in the expression map. Still, I would 'manually' and specifically need to set up the short bow choice in the score in some way, for each note that needed it. Without something like lua script to automate the process....yuck....no fun approach for 'importing' a score. Gotta go through and tag all the notes individually for short bowing variations.

Exclusion groups might help ease things a bit more, but Dorico and I still have to tell the plugin which short bow articulation to use.

GPO had no way of distinguishing on its own which short bow articulation to use, just by sending a common event for 'the note has a staccato dot over it'.

If NP does indeed analyze things and use something like the time between note-on and note-off events, perhaps combined with note velocity, to guess at the best choice, then there's likely to be some buffering and latency in the process. It could well be noticeable enough delay that mixing different plugins playing parallel with NP in the same score would be a no go.

Rob Tuley
Grand Senior Member
Posts: 4095
Joined: Fri May 20, 2016 12:41 am

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by Rob Tuley »

Of course NP synchronizes properly with other VSTs. It wouldn't be much use if it didn't.

You get a 1-second delay between hitting "play" and playback starting, and a 1-second delay after you hit "stop" before it stops (which is actually quite like what happens with a real orchestra!)

I never use the HSSE samples for anything (unlike you I'm not going to first buy HALion6 so I can tweak them, and then spend hours doing it) but I use NP + Pianoteq all the time.
Take the dotted (staccato marking) on a note that lives under a slur example.
Should it use sautellie, martele, spicatto, or staccato?
I used to play violin (though I haven't touched one for years) and frankly, those are just words. As I said in another post, real violin playing doesn't use separate little packages of bowing technique like you find in a sample library.

I entered a bit of your Schubert quartet, and amusingly it played better if I deleted some of Schubert's articulations. NP was doing the right thing automatically, and explicit instructions in the score were making it over-do it. Of course Schubert's marks are "harmless" for human players who are smart enough to know that a staccato dot doesn't mean exactly the same thing on every note that is marked with one.

Brian Roland
Senior Member
Posts: 1104
Joined: Fri Jun 19, 2015 5:24 pm
Contact:

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by Brian Roland »

Rob Tuley wrote:
Fri Jul 24, 2020 10:54 am
Of course NP synchronizes properly with other VSTs. It wouldn't be much use if it didn't.

You get a 1-second delay between hitting "play" and playback starting, and a 1-second delay after you hit "stop" before it stops (which is actually quite like what happens with a real orchestra!)

I never use the HSSE samples for anything (unlike you I'm not going to first buy HALion6 so I can tweak them, and then spend hours doing it) but I use NP + Pianoteq all the time.
Take the dotted (staccato marking) on a note that lives under a slur example.
Should it use sautellie, martele, spicatto, or staccato?
I used to play violin (though I haven't touched one for years) and frankly, those are just words. As I said in another post, real violin playing doesn't use separate little packages of bowing technique like you find in a sample library.

I entered a bit of your Schubert quartet, and amusingly it played better if I deleted some of Schubert's articulations. NP was doing the right thing automatically, and explicit instructions in the score were making it over-do it. Of course Schubert's marks are "harmless" for human players who are smart enough to know that a staccato dot doesn't mean exactly the same thing on every note that is marked with one.
I don't blame you if you'd rather spend your time working with the higher levels of composing and arranging. That was kind of the point of the thread. How NP is great option for many users, and that is available NOW, and for a reasonable price.

No, they are more than 'just words'. They represent actual techniques, and they each produce easily distinguishable sonic characteristics. The changes in bow speed/pressure/etc. are quite real.

In tutti instruments, it really is essential to have an actual sample of several people playing the instruments together, on different instruments (or building it up to seem that way...but all ultimately mixed down to a single sample). It's going to be a very complicated waveform. Hence, why my martele-like mock-up sounds 'artificial'. Best I could do with my current scope of knowledge in HALion, was do a little dynamic shaping to entire sample.

With a solo instrument, one can mimic the various techniques more closely with changes in dynamics and pitch, and exciting or filtering a certain frequency range, and theoretically get much more convincing variations of bowing techniques out of a single sample.

It's possible to build up tutti presets that actually meld solo samples. I don't think it's practical with what comes in HSO through. It doesn't have but two sets of samples for each solo string instrument (that I know of at this time). Mix too many of the same sample together, and you get phasing issues that can lead to artifacts or frequency cancellation effects. It also drives up the voice count, and memory or D2D bandwidth demands...meaning potential users might need higher minimum system specifications to use them.

No one said that humans adhere to them in any ridged form. No one says that it can be done well with mere key-switches alone. Getting a base template for the sound instantly IS a good START however. Refinement begins with basic concepts. It's hard to reach the top of the ladder without going up all the steps in-between. Dorico, and/or the instrument he talks to needs to eventually be able to step on all the rungs, but we still approach them one by one in the process.

For me, when I'm designing a sound, I tend to overdo it at first, so I can more easily detect what effects a given set of tweaks are having on the sound. Over the long haul, of course I'd dial things down for the default configuration. Again, proof of concept...that was the goal.

When things are 'overdone', it's easier to take away than it is to add ;) Like the artillery battery that sends the first shell long, the second short, and the third on target. It was a sketch, but there is a dial there that'd eventually get logic applied for the user to easily dial it back ;) Educators often present overdone examples ON PURPOSE....so a young student can more easily detect the concept. Students often break a lot of leads when first learning to write with a pencil (even the fat stumpy ones)! If a kid is overdoing accents, it's easy to ask her back off. If she doesn't know the technique at all, then the performance must go without, or you stop and try to teach it to the kid on the spot.

I'm glad you are happy with the current state of Dorico + NP, and that it already does everything you need. I don't recall asking anyone to buy H6 and design a sound. If one doesn't care to design or deep-tweak sounds, they won't need it.

As you can see, I've merely created a 'shell' of an instrument, that can ultimately be polished into a plug and play string instrument that doesn't require much user intervention to get nice sound, and can be vastly improved or customized for a given musical piece/style/phrase with a small amount of tweaking.

I'm looking forward to better over-all score translations in the future, regardless of the library one connects to those stave endpoints; thusly, I am sharing some of my experiences in working with the way scores are translated and communicated to plugins. I've offered specific examples of things I'd like to see added to Dorico for the purpose of translating scores nicely with as little fuss as possible on the behalf of the user.

I don't know how many man hours went into building, testing, refining the string sounds for NP, and teaming them up well with expression maps and such for Dorico, but I'd be willing to bet it's WAY more than I put into the concept sketch above, and I'd also bet they were using MUCH better hardware, in a much better room to work on it.

The reason I bothered to spend some of my personal leisure time LEARNING to build a macro page for the free HSSE player, using samples that are included in any standard installment of Dorico Pro, as I've said many times before was for easy collaboration and proof of concept.

IF I were to share anything I wanted translated in a certain way, I could simply provide the instrument presets along with the score, and the person/people I share it with don't have buy thing to pick up where I left off. They double click a single vstsound file...it installs to the Steinberg Library. Load the score...and they're seeing and hearing it exactly as I last left it. Minimal fuss, and no extra expense to the target audience.

It's not necessary to jump through ALL the hoops I did in that example just to get a good basic string sound with HALion 6/Sonic/HSO. HSO Strings are far from being the only sounds included with HALion 6 and HALion Sonic 3. Quite a few of them are quite nice with nothing more than the general expression map with either velocity or cc dynamics, and nothing else.

Learning and doing the macro wasn't a requirement to make HALion work, and make good sounds, but for the sake of user controls the free HSSE player, I did it.

There are a million and seven simple things that can done to humanize the short bow selections.....
That's my real reason for joining in the thread! The discussion started as where things might go in terms of Dorico features for score translation, and communication with plugins. What improvements can be made to the system that help it make more musical, or human like choices.

How much of that is the responsibility of Dorico, and how much is up to the plugin/sample library?

With older instruments, the expression map system really needs to be MUCH smarter and easier to work with. Little things that prevent, or automate a user having to do a lot of redundant entries to a score to get the translation desired. The play tab editor is also still very important, because there is so much one can do to really refine and polish the ultimate rendering quality. With a great play editing tab, the entire expression map thing is totally optional. Having a great play tab editor provides a workaround for any missing features in the expression map and automatic translation capabilities right away.

Newer plugins like the entire HALion family, can take advantage of cutting edge feature sets of the VST protocol. Things like Note Expression. Snooping information from the transport like the current key and tempo, the actual db level of the mixer channels, and more. So, there is MASSIVE potential to make very smart instruments that don't require an extensively complicated expression map.

Note Performer is beyond version 3 now. It has been in development since 2013, by PROS. Of course it sounds better than my garbage rendering!

I brought it up because the discussion drifted into the field of modeling vs lots of samples vs a hybrid of the two. I took a few samples, and gave an 'example' of some possibilities of what sound shaping can do, using sounds that SHIP WITH DORICO.

A. It's built using samples I, and every single Dorico Pro user has the rights to install and use. A valid reference point.
B. HALion IS the default *Steinberg* foundation for nearly ALL of their VST instruments.
C. It is the flagship VSTi platform for ALL PLUGINS VST. If it's part of the VST protocol, HALion will get it first, and it'll set the standard.

Being a subsidiary of the Steinberg/Yamaha line of products, it kind of made sense for me to use this as my reference point. If the Dorico team gets it working in the HALion engine, for which they have more access to the technologies and dev kits, without having to pay as many/if any extra licensing fees...it makes sense to collaborate with HALion examples. Get it working with HALion, and others will follow. HALion IS the industry standard for VSTi protocols.

So, even though my example string presets are 'in the barrel and overdone', it doesn't negate the points I've tried to make.

I'm appealing for continual support and improvement of expression maps in Dorico, attaching more of the VST informational parameters, AND improving the play tab's DAW like features. I put some proof of concepts as to why I think BOTH are important on the table.

If someone asked me which to do FIRST, I'd say:
1. Improve the legato/slur concept and implementation.

2. Export/Import of flows in midiLoop format (endpoint info included, for easy exchange with other hosts)

3. Beef up the transport....connect the tempo parameter so plugins can have access to that information.

4. While working on the transport, let it send mtc code over a MIDI port of user choice, and serve as the Master.

5. Add mtc slaving abilities. (This could wait in my personal dream priority que if we had feature 4, or vice verse)

6. Improve the cc lanes, so it's easier to work with isolated entries. I.E. Terraced entries.

7. Scaling, and curve drawing tools for the CC lanes.

8. Logical editors, or similar conditional/scripting abilities in the play tab.

9. Groove templates.

10. Gradual improvements to the ease of working with note on/off positions on the timeline.

11. Improvements and additions to translation options with the expression map and score technique systems.

12. Option to use individual Note Expression events rather than channel CCs. I.E. Double click a note in the play tab, and draw or record expressive data on a graph there. It becomes relatively time-locked and part of the specific note, rather than floating freely as channel data. Several notes on the same stave/channel can have independent expression data.

Plenty more, but that's enough to keep a team rather busy for quite some time.
Last edited by Brian Roland on Sat Jul 25, 2020 1:54 am, edited 4 times in total.

robjohn9999
Member
Posts: 771
Joined: Tue Feb 28, 2017 7:58 pm
Location: greater NYC region
Contact:

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by robjohn9999 »

I'd still give Note Performer a try, though just to see how you find it compares (especially given the time difference in terms of the need to program, etc.)...(I think they offer a 30 day trial)...Certainly, I'd be curious your thoughts after you played with it a bit...
- D.D.
2019 MacBook Pro with 64 GB ram; Apogee Ensemble 88; Mac OS Catalina 10.15.2

Rob Tuley
Grand Senior Member
Posts: 4095
Joined: Fri May 20, 2016 12:41 am

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by Rob Tuley »

I disagree about the "just words" thing. The things the player can control are bow speed, and direction, bow pressure, rotating the bow to change the amount of hair in contact with the string, and the distance of the bow from the finger board.

Also, the bow is flexible and will "bounce" on and off the string if you control the pressure to make it do that.

That's all there is to violin bowing. There are only three basic techniques, detache (always on the string and always moving), martele (always on the string but sometimes not moving), and spiccato (not always on the string). Add in a few special "tricks" like col legno, if you like.

But if a sample library has 57 different samples of the same note, somebody has to invent 57 different names for them. You are making the mistake knows as "reification" - giving something a name doesn't mean it is a thing. Everybody was happy to think "gravity" was "the thing that makes apples fall downwards off trees" until Einstein worked out that there isn't a "thing" called "gravity" at all (but I'm not going to try to explain General Relativity here, so don't panic).

Of course string players use more words to describe different variations of the basic techniques - and they also get into flame wars with each other about exactly what those words mean, which are often arguments about a distinction without a difference!

(Note, I'm not picking on string players here. Every other group of instrumentalists probably does the same. Keyboard players certainly do!)

Brian Roland
Senior Member
Posts: 1104
Joined: Fri Jun 19, 2015 5:24 pm
Contact:

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by Brian Roland »

Rob Tuley wrote:
Sat Jul 25, 2020 12:57 am
I disagree about the "just words" thing. The things the player can control are bow speed, and direction, bow pressure, rotating the bow to change the amount of hair in contact with the string, and the distance of the bow from the finger board.

Also, the bow is flexible and will "bounce" on and off the string if you control the pressure to make it do that.

That's all there is to violin bowing. There are only three basic techniques, detache (always on the string and always moving), martele (always on the string but sometimes not moving), and spiccato (not always on the string). Add in a few special "tricks" like col legno, if you like.

But if a sample library has 57 different samples of the same note, somebody has to invent 57 different names for them. You are making the mistake knows as "reification" - giving something a name doesn't mean it is a thing. Everybody was happy to think "gravity" was "the thing that makes apples fall downwards off trees" until Einstein worked out that there isn't a "thing" called "gravity" at all (but I'm not going to try to explain General Relativity here, so don't panic).

Of course string players use more words to describe different variations of the basic techniques - and they also get into flame wars with each other about exactly what those words mean, which are often arguments about a distinction without a difference!

(Note, I'm not picking on string players here. Every other group of instrumentalists probably does the same. Keyboard players certainly do!)
And the part of the bow used....closer to the center, or towards the ends, pinch on the bow....
To a degree, also the direction of the bow.

Let me put it this way....

If I were going to put in the score for Czardes, I'd be looking for a distinct sautellie sound. I don't know how the sound is made, but I have an idea what it's supposed to sound like. Maybe it's 'spiccato'....but when I hear spiccato, it sounds more V shaped to me where the bow is in contact with the string a shorter amount of time through the stroke, the bounce is harder...more percussive echo from the bounce. Sautelle more like bowl or U...string makes contact longer through the bounce motion, less sharp of an echo, but brighter. Something like that....

Either way....they do NOT sound the same!

Not knowing much about strings, knowing I want that 'sautellie' sound, I'd probably use staccato looking dots over the notes, and things like tempo and, context would probably leave plenty of clues that it should have that 'sautille' sound...or if every single note in a passage uses the technique, I might just use the word Sautellie as a playing direction at the beginning of the passage, and then put an ord. at the end of the passage.

If I think the phrase should be using a short bow sound with a lot of pow and ping inflection, I reach for martele.

I've never had the privilege of conducting an ensemble with strings, but if I were to stand in front on one, and they had a piece I'd composed sitting in front of them. I think I'd use these 'just words' to describe what I'd like them to play back for me.

Not being a string player, I wouldn't have the first clue on how to say, "Hey second violins, will please you tilt your bow a little more and do this and that, towards the center of the bow?"

Named or not...the point still applies. A pallet of choices are available from which software that uses sets of conditions can make choices.

At the various levels of designing a sound, depending on the engine you are using....one can keep huge pallets of level/time based envelopes, and call them up on demand, or even cycle among them round robin, and so forth. They can alter pitch and dynamics...to the entire sample, or to bands of frequencies within the sample, or manipulate programmed filters, etc.

In addition, parameters can be tweaked in absolute, or relative scales in real time. Etc.

The concept can apply to actual samples, or just templates for shaping a given sample.

dodecabilly
New Member
Posts: 1
Joined: Tue Jul 28, 2020 4:39 pm
Contact:

Re: extensive sound library expression maps vs. just using Note Performer (opinions sought!)

Post by dodecabilly »

On a subject of using both DAW and notation software for different purposes - this is how I did so far, switching back and forth for notation and mockup, but here's the problem: different working environments affects my composing. If I start composing in notation software, then go over to DAW for more realistic mockup, while editing the MIDI in DAW I can come up with different solutions regarding arrangement or orchestration, or the composition itself. Because more realistic sound gives me better emotional feedback (with good live performance with real musician being the best possible feedback), which in turn can make spark another ideas and revisions. And vice versa - when i port the midi mockup into notation software, the visual feedback of notes and staves is leading me again in another direction, or simply making me realise that something I came up in DAW won't really work when real musician is supposed to play it from the score. So, when i chose Dorico, after initially being a Finale user, it was because I hoped to have single environment for composing. And composing music in my case (and I believe I am not the only one) includes both notation and realistic mockups.

Dorico seems to be the best option on the market regarding this, but I still need to use DAW for mockups, unfortunately... Hopefully the gap will get smaller in the future.

Post Reply

Return to “Dorico”

Who is online

Users browsing this forum: Dup, dyross, HeiPet, Joesson, Lillie Harris, naomi, PjotrB, samuelmbradley, Ulf and 13 guests