Big Dorico news coming! Dorico Live Stream

[Dewdman42]

To be clear, I'm not talking about generic movement of surrounding CCs. I'm talking specifically about the CCs and/or keyswtiches associated with and generated by the note articulation. If they don't come before the note on, they don't work. In Dorico currently (3.5) if you move the note ahead of the beat, the controls stay where they were and don't work.

have you submitted a bug report about that? That seems Like a simple bug that shouldn’t be that hard to fix. You are saying that when the piano roll is used to alter the performance of a note to an early attack while leaving the notation as is, the expressionmap keyswitches are being triggered by the notation rather then by the note itself, which is definitely an oversight that could be fixed pretty easily I would think. I don’t see that as major design flaw.

i started looking into Notion last night and indeed their rule set approach looks to be much more complex and capable then expressionmaps, though also complex enough That i think most of the target audience, ie composers, would be turned off by it at the best and many probably would not even be able to fully utilize it due to its depth and complexity. It’s kind of lurking there for years and you don’t see really any relevant example rulesets for modern sample libraries. I can’t get any of my Vsl products to work with notion due to elicenser problems so there is that too, but anyway presonus may be on to something there if they put a little more effort into it. They should add their rule set capability to S1 also which has zero articulation management right now.

i get why you are trying to get the dorico team to pay attention to notion for ideas though as it seems the notion folks have thought through the playback issues better then anyone, which isn’t surprising; notion was really the first notational program to offer any kind of reasonable sample playback quite a few years ago andI remember having conversations with the original founder of the product where he discussed that primary focus. So they really were the pioneers. But it never got off the ground really, then sold to presonus who hasn’t really done much with it since. So notion is not without its own faults I’m afraid and somehow whatever lessons they learned about notional playback have flown under the radar with hardly anyone noticing. Meanwhile steinberg is reinventing it all over again, but painfully slowly it would seem.

 

[Gene Pool]

I honestly do not believe there is, nor need be, some tradeoff between playback and the rest of Dorico.

Molto agreement, bro.

 

[mducharme]

It is just very awkward to do the kind of CC shaping and humanization in Dorico that you can do with a DAW, and the mixer cannot compete with the Cubase mixer. These are mostly UI-related deficiencies at this point. I was pleased to see in 3.5 that now when you record from a MIDI keyboard it actually retains the offset "errors" from the performance instead of quantizing everything and requiring manual de-quantizing with the mouse to prevent a robotic performance. This solution means that when you record from a MIDI keyboard you keep the imperfections of the performance like in a DAW, while the notation is correct - kind of the exact opposite approach of Cubase's "display quantize".

I think Dorico would have to adopt most aspects of the Cubase piano roll UI in the Play tab in order to really be a viable place to do mockups, plus the mixer UI. I'm not sure they are going to want to do this since it might cause it to eat into Cubase's market share for certain kinds of composers.

 

[richhickey]

have you submitted a bug report about that? That seems Like a simple bug that shouldn’t be that hard to fix. You are saying that when the piano roll is used to alter the performance of a note to an early attack while leaving the notation as is, the expressionmap keyswitches are being triggered by the notation rather then by the note itself, which is definitely an oversight that could be fixed pretty easily I would think. I don’t see that as major design flaw.

I did report it. It is currently a 'missing feature'.

i started looking into Notion last night and indeed their rule set approach looks to be much more complex and capable then expressionmaps, though also complex enough That i think most of the target audience, ie composers, would be turned off by it at the best and many probably would not even be able to fully utilize it due to its depth and complexity.

I think the simple fact is that making good, complete expression maps (esp. from scratch) is not really in the wheelhouse of many musicians or composers. That said, there are plenty of people in the community for whom it is interesting and doable, and it only takes one or a few per library to provide basic support everyone else can use. There is a definite tension between easy and powerful tooling, for sure.

 

[Dewdman42]

I mean I even see that most expression map usage is very simplistic not even using the full capabilities of cubase expression maps as limited as they are and many people just avoiding them because they can’t figure them out. There is kind of a user wall there

 

[richhickey]

I mean I even see that most expression map usage is very simplistic not even using the full capabilities of cubase expression maps as limited as they are and many people just avoiding them because they can’t figure them out. There is kind of a user wall there

Well, I think you'd agree that we'd be better off if more things were scriptable, e.g. via JavaScript. Most of the rest of the world can improve their own lot via scripting while music software still relies on bespoke interfaces and tools. Plenty of musicians can wrangle JS, and then you have the full expressivity of a programming language. I.e. Dorico just added 'conditions' to their expression map dialog box, with 'Any of' (OR) and 'All of' (AND) radio buttons. But no NOT, no parens/grouping etc. You quickly come up with things you want to say but can't. Notion went much further with their flags system but it's still not as good as proper logic.

 

[mducharme]

I use Cubase expression maps extensively. My own orchestral template has a 1:1 mapping between tracks and staves on an orchestral score. I made my own to put Spitfire Chamber Strings and Spitfire Solo Strings into my template with each section on one track only, rather than a track per articulation. One of the biggest problems I have with expression maps in Cubase is how many lanes I end up with. I have to make the articulations change lane super tall just to see the names of all of the articulations and to use my mouse to switch. I hope that whatever enhancements Dorico comes up with in this regard are integrated into Cubase as well.

 

[Dewdman42]

of course the nerd in me gets a little giddy at any mention of "scripting" built into any kind of product, including music software. I doubt Dorico will add that in this decade. Cubase still doesn't have it either, and maybe never will. Best we might hope for from both of them is a dedicated plugin slot in front of each instrument slot, in order to host VST midi plugins (which might be in the form of a third party scripter plugin)

Steinberg has a tendency to want to abstract things...which is part of why VST3 become broken, they designed this highly abstracted API....doing away with the "archaic" midi protocol... Unfortunately ten years later the rest of the industry is still loving and using the archaic midi stream....so...concepts like VST3 note expressions, for example, are not used anywhere except in Steinberg plugins and Cubase. But that is how Steinberg thinks we should all be articulating our performances...not using midi stream keyswithces and archaic crap like that. They might be way ahead of the everyone, but unfortunately they are also totally out of sync with everyone...which has caused a lot of tension in how things operate...VST2 vs VST3 is just an example, the timing problems of CC switches, etc.. And then on top of all that they are trying to abstract Expression Maps...and perhaps that could be a great way to go, but it needs more work...and I suspect that they are not motivated to make Expression Maps work with archaic midi protocol, if you catch my drift, they would be much more interested in developing something that works with their religion....which is VST3, note expressions and perhaps even other newer abstractions that we don't even know about that from deep inside their VST temple.

 

[ptram]

Is that background music new or an existing piece? Sounds very much like John Adams.

The first time I listened to Akinola's Prelude, I was immediately thinking to Missy Mazzoli's River Rouge Transfiguration. Similar chord sequences, similar sudden ending. I know Mazzoli is deeply influenced by John Adams. Is there a piece, in particular, you are thinking about?

Paolo

 

[shawnsingh]

I took Spitfire, Berlin, Cinesamples, Scoring Synths, and Vintage Keys to the doctor's office last week.

I cant believe they all got sick at the same time...

(sorry, I'm a new dad, I just had to...)

 

[Bollen]

Why is dorico a PITA? You described very eloquently about why AI is a long ways off from being able to handle all the various aspects, but dorico is not trying to be AI.

I think you are probably meaning to talk about rule based articulation management, ie expression maps; which has nothing to do with AI. But it is true thst using rules to automatically handle articulations is what dorico is attempting to do, I think better then cubase at this point. But perhaps you are saying that the expression maps are currently not flexible enough to control things the way you would like to control?

doing things manually without expression maps or articulation sets or any automated articulation management is something you can do easily, though laboriously, with most any daw. There are draw backs to that too.

Can you give us some concrete examples and how and why dorico is difficult to bend to your will in terms of manually handling things or specific cases where the existing expression map facilities are inadequate for you to control things or how specifically they are a pita to setup?

Sorry, you're absolutely right. The above was a necessary preamble to establish why I find expression maps useless and the wrong approach (on a side note, I do consider expression maps and AI very related. In fact I would say the former is an early stage of AI, since it's based on conditions that trigger certain behaviours according to defined parameters. So not science fiction AI, but rather in the way it used in the gaming industry).

In order to answer your question as clearly as possible I believe by using a practical example I can illustrate the issues better. Say I begin a composition for about 6 instruments, after I do all the convoluted setting up of the score, VSTs, mixer and load up the individual samplers (usually hosted on VE Pro) I write a chord played forte and then begin a fugue in pianissimo. Within just those two events I start to run into problems: the forte is unbalanced, some instruments are too loud and some are just playing the wrong articulations. The notes simply have an accent, so Dorico is just playing a sustain at a high velocity. But to get the instruments to perform an accent on a first beat, some will require an fp patch, some sfz patch, etc. In Sibelius I used to trigger them with texts that I could hide or an invisible staff with the keyswitches. In Dorico I can create a staff below and then remove it, but wait..! Ah damn! I can't do that on ensemble instruments... Now we move to the first notes of the fugue, too many instruments are too loud (VSL for example doesn't have pianissimo, so you have to emulate it by using a combination of low velocity/p and another controller, either a filter or expression) so now I have to move to the Play window, find CC2 or CC11 from 128 different controllers, set it to 0, now find the filter CC... Hmm... Now it's too quite, go back to CC2, OK now it's fine. Press play, hmm... It's kicking in too late, I can hear the attacks of the first notes in the fugue as if they were forte. Back to the Play window, the view is reset! Damn, I can't grab that corner to expand the piano roll... OK, I did it, zoom in... Argh! The mouse wheel doesn't zoom in, I keep forgetting even after three years of use!!! OK use Z, oh that's too much, Y to zoom out... Hmmmm... That's too far out, so it's somewhere in the middle but I can't do that, nevermind, just put on your glasses. Oh, I can't put the controller change where I need it because it has to snap to the grid which is already set to minimum, but because of the humanisation some notes start before the grid... But I can't put it in the previous grid because that's still playing the forte....!!! OK, I'll do a little trick with CC7, oh it suddenly scrolled away from where I was working, now I'm lost, damn! Sometimes it scrolls up and down and sometimes sideways!!! I wish I could open two controller lanes at the same time to see what's going on....

Anyway, I could go on and on, but I think that explains the general issues. On the other point of DAWs, I've never been able to write complex harmony and counterpoint on it and I've never met anyone that could so they could show me how to, so that's not a solution either. I think notation, I need to see the music, vertically, horizontally and ideally all instruments at the same time.

 

[Dewdman42]

AI doesn’t work based on rules. The whole idea behind AI is that you program a computer how to learn, and then it learns through experience how to make decisions.

rule based stuff isn’t AI it it’s what you might call automation i guess since you the user can decide the rules you want enforced and you only have to specify it once and then keyswitches will be dutifully inserted in all the “right” places according to the rules you established.

AI is capable of incredibly complex decision making which would be utterly impossible to figure out in rule based system, but we give our control away to the computer at some point.

rule based system still give us complete control but the downside is that often they are not complex enough to address all the possible permutations and situations, thus leaving us frustrated and considering going back to full manual mode again, which is why they usually need to always have some kind of manual override mode.

carry on...

 

[ptram]

I would call expression maps simply patching or routing.

Paolo

 

[CatOrchestra]

I am wondering if I write a piece, can Dorico afterwards write out the figured bass for me???

 

[Bollen]

AI doesn’t work based on rules. The whole idea behind AI is that you program a computer how to learn, and then it learns through experience how to make decisions.

rule based stuff isn’t AI it it’s what you might call automation i guess since you the user can decide the rules you want enforced and you only have to specify it once and then keyswitches will be dutifully inserted in all the “right” places according to the rules you established.

AI is capable of incredibly complex decision making which would be utterly impossible to figure out in rule based system, but we give our control away to the computer at some point.

rule based system still give us complete control but the downside is that often they are not complex enough to address all the possible permutations and situations, thus leaving us frustrated and considering going back to full manual mode again, which is why they usually need to always have some kind of manual override mode.

carry on...

Naturally I agree with you since I have a basic understanding of the philosophy of AI. However, when I have worked with game developers in the past and they've talked about working on the AI, it is when they create a set of instructions for the computer generated characters (or NPCs as they called them). These instructions trigger actions according to conditions i.e. if player is in the line of sight of NPC then do this, if it's not then do this, if player is near do this, etc. Sounds awfully similar to expression maps to me... ! If no articulation play sustain, if accent play this, if accent with sordino play that, etc.

In any case, despite being an interesting topic, it's irrelevant to this thread and we also agree it's not there yet.

 

[Dewdman42]

If you are just programming some simple rules that is not AI, that is just programming. True AI is where you program the neural network infrastructure and then the rest of the time is spent teaching that AI by feeding it information. The AI neural network then gets bigger and bigger and more complicated through experience, not through programming it. In this way incredibly complex decision trees can be built up by the computer itself without much or any planning by developers.

we would have to know more information about what your game developers are doing to say for sure whether they are using some form of actual AI, the term gets thrown around loosely. But in general if you have a system of predetermined rules with predetermined actions then I don’t see that as being AI, unless those are the building blocks that enables the software to become smarter and smarter through experience.

AI actually WOULD be the correct solution for this problem we are discussing because of the complexity but that is many years away before we will see a true and useful AI solution for this kind of thing that doesn’t also get in the way of our creativity.

 

[Sean J]

I cant believe they all got sick at the same time...

(sorry, I'm a new dad, I just had to...)

I meant to reply to this... Congrats!

As a food storage junkie, I thought I'd stock up on diapers ahead of time. I had 1400 diapers stockpiled, based on average sizes and growth per others via Google who had the same idea. I should have gone higher on size and doubled the stash. Welcome to the other side of life.

...useless and the wrong approach...

The workflow of editing a note... then a technique... then a CC value... That requires more steps with your hands than our minds think it should. It's all fine until you move a note and realize other stuff has to move with it. Even with directions and attributes, there's really no musical intelligence in it. Most notation programs escape that paradigm as you'd be editing some form of engraving anyway. A DAW runs the risk of feeling half-featured unless the UI is optimized for the writing/editing process. Few DAWs pay enough attention to that IMHO.

 

[jamwerks]

I don't use the upper numbers, but the numeric keypad instead... done so for many years, much easier for me. Anyway, hitting a number and then pressing the spacebar is two moves, while the way Finale and Sibelius manage it, i.e. simply pressing the duration number is much faster. This leaves me the impression that Dorico isn't aimed at professional copysts

I know what you mean, and I thought that also for a while. There are some thing that in Dorico that take more actions (clicks, keys, etc.) than in Finale. I've gotten creative to try to lessen that as much as possible. But globally (data entry + refining + page layout) Dorico comes out being now almost twice as fast as Finale for the same job. And where separate parts are involved, that's about 3 times as fast imo.

 

[Dewdman42]

I think The Dorico team is doing remarkable things with the notion of FLOWS and LAYOUTS. Also their rule based approach to notation is in my opinion going to eliminate what normally would be a lot of manual work nudging things around to look the way you want. Sibelius was always famous for trying to do that better then Finale, but I think in Dorico they are taking it to the next level.

 

[Rob]

I know what you mean, and I thought that also for a while. There are some thing that in Dorico that take more actions (clicks, keys, etc.) than in Finale. I've gotten creative to try to lessen that as much as possible. But globally (data entry + refining + page layout) Dorico comes out being now almost twice as fast as Finale for the same job. And where separate parts are involved, that's about 3 times as fast imo.

then I probably need to put in the hours of apprenticeship. But I remain doubtful, as now for me, Dorico feels much slower, not only for notes inputting but for the responsiveness as a whole. Finale feels snappier... I have a hard time, having to honor deadlines, investing precious time learning new ways to do things that I already can do much faster. I'll keep getting updates though, I like to encourage development of new notation software.