MIDI VSTi Edition of Synfire

I feel the same way and have expressed it here several times. The developers see it differently...

Me personaly have no problem with the current workflow.
It actually helps me in not overdoing a track, by not adding too much and keeping it minimal.
I think the problem with vsti's only is that they have to comunicate with eachother, but I could be wrong.

I would love to understand their point of view!

No problem ;-)

The answer is: It won't work. 

1. A plug-in can not do that. 
2. It would be a horrible user experience.
3. It would require device descriptions nevertheless. 
4. The setup would be twice as complicated.

Since this suggestion keeps popping up again and again, I can only assume that the fundamental concept of Synfire was still not understood. This is obviously my fault, because all communication did just not seem to be fruitful to date. Maybe it is simply not yet the time for this kind of software (sigh)

Let me explain the above points:

A plug-in can not do that: VST may support MIDI output, but AudioUnits don't. Only few DAWs allow for MIDI routing across plug-ins and tracks. Also there is no way for a VST to read or write a host's MIDI data, let alone other track's MIDI data. The inherent limitations of VST would cripple Synfire and dumb it down to a "step sequencer", "phrase generator" or "chorder" kind of thing.

It would be a horrible user experience: Plug-in windows have a fixed size, are relatively small and extremely limited. Keyboard shortcuts conflict with those of the host, keyboard input and menus are limited. They behave and feel like a virtual piece of hardware, rather than an application. This is part of their concept and can not be changed. Even a 24" monitor sometimes is to small for Synfire, so how would it fit into 1024 by 768 pixels? Also no way to copy & paste or drag & drop phrases, libraries, parameters, sounds and devices and everything between multiple projects. Most DAW also do not allow running multiple open projects at the same time anyway.

It would require device descriptions nevertheless: The purpose of devices is to let Synfire know fundamental facts about the sounds it is supposed to compose music for (sound assignment is a byproduct only). A human composer can hear how a sound performs and make ad-hoc decisions based on envelope, timbre, common sense and their personal taste. Synfire can't. It needs meta data to "know" how to play a sound. Hence, even in a VST, Synfire would need playing ranges, categories, bank selections, custom CCs, articulations, etc for every sound (all this is kept in a device).

The setup would be twice as complicated: Imagine the trouble of setting up 30+ tracks, each with one instance of a Synfire VST, making them communicate properly with each other, making one the master, the others a slave (oops, accidentally deleted the master...) etc. All sounds hard-wired manually in your DAW, with no chance for Synfire to take control. Every change arranged in Synfire would need to be mirrored in the DAW manually. Your project became a huge static patch bay.

Besides that, making a Synfire VST would require many years of development, because it would have to be written almost entirely from scratch to fit the heavily constrained VST platform. Apart from this being a risk that Cognitone can not afford, I rather want to invest that time to make the current user interface simpler and easier to use. 

Thank you Andre,  please don't say the world is not ready for SFP.   I think the issue is, you are coming from a different point of view..  You've lived with this software for many years..  Everybody here is probably coming from years of a certain DAW.  You've invented a piece of software, which has a few things in common with DAWs,  but most of it radically different.. 

Most of the time I run SFP, Logic and Band-in-a-box.. I have no problem exporting a midifile and importing into a 2nd program, for more build-up work.. Band-in-a-box has finally got their drag a midifile from one app to another Stylus RMX allows you do do that.. Music Lab Real quitar lets you export midifiles to the desktop and then you drag into your DAW

The problem to me is that as brilliant as you are, you skill in dumbing it down for us newbies, is not as great.

This really needs a third person who can absorb what you've done, and break it down into analogue equivalents were there are some, and then basic 7th grader reading skills for the code that has no counterpart. 

Often yours and Supertonic's explainations (to a lesser degree) still leave me in the dark.. Cause you're talking 4th year Latin, and we're all freshman.. Slowly through repetition, re-asking the same questions over and over are people getting it.. It is a slow process.. I'm going on two years with SFP and still have days, where I get frustrated and shut it off. Other days I toil at it, and make some headway. 

I once found an article by the main software writer for the Tyros 4 software (a masterpiece of code).  His article was not so much about particular in/outs of the operating system, but why he came up with it in the first place.. It was character study data, of how he frustrated as a musician, wrote software to realize his vision. It was tremendously informative..  Just as you tell children fairytales to begin to teach them basic concepts or right/wrong/good/bad/ etc.. You have to somehow find a way to do that.. I purchased The Tyros 4  with the concept of using it as a compositional aid with Logic.. Surprising there was no written documentation of that being done..  I made lots of calls to Yamaha, but got little satisfaction, only by doing laborious experiments, writing down the results and redoing the procedure over and over did I gradually find the right solutions..  At first most of my assumptions were wrong, It took me almost three months, to be certain I could fire the program up in Logic tomorrow and it would sound exactly as it did the day before.. Along the way I found a handful of other musicians thinking along the same lines.  Two people went as far as to write free extensive software to help (unfortunately on PC)

I have not reached that point with Synfire yet.. Most of the time on opening a previous project, things are ok.  other days the song does not sound like it did the day before (mostly to do with DSP effects(.  One day I can't get library previews to work, I have to put into the arrangement blindly to see if it works.  I still have the bizarre issue, of when I can preview an instrument from the library, it drives my Apogee Duet into some kind of feedback loop, (it's not audio generatored, cause T4 volume is off. rather it sounds like a subdivisioin of the sampling frequency, happens at 44.1K

You're telling us to go to such and such a window and click this box... helps us in the short term but not in the over all understanding, the background, the why of this software.. I would assume this would take a 3rd outside person to do this, he needs a whole different skillset than yours..  

Just like most manuals from Japanese kbds, are obtuse to say the least.. things get lost in the translation,  But please don't shrug your shoulders and sigh after most of here have poured a large sum of money and time into SFP/SFE. 

Without a doubt this is the hardest software I have come accross but I love this software (when it reaches full maturity) For the time being, I'm learning everyday.. I'm going to give it one more year..  It is frustrating cause what you are doing is pioneer work, and we are your paid beta testers..  I think it's still worth the price for promise of what we'll gain..  Indeed I have been reading books on harmonization, and can see the similarities in your software.. And I am even beginning to do some of myself in Logic, applying the SFP principles. That knowledge and the motivation I found to undertake reading dry harmonizaton techniques, etc.  is worth the price.. 

This software is effectively makeing a user jump over 4 years of intensive theory, harmony, counterpoint, modalities, etc..  You cannot expect to give a chainsaw to a 4 year old and have everything be ok.. Those of us using the software  without the musical background in these areas, are delighted we suddenly have the power to do this, But we are yielding a Starwars light sabre.  Eventually we need to come to a basic understanding of what we are doing to be truly effective.. I hate those 20 something year olds  which buy the newest box, use the demo song, throw rap lyrics and becone huge stars. 

There has been a significant amount of progress that SFP has made..Hopefully SFE will attract more users, so you have more capitol to work with ,  We are coming from DAWs and trying to relate it to that..  

I appreciate the honest answer you gave, and I'm sure it will appease many...  SFP is not the type of software that will be appreciated and mastered by everyone.  I tried to get into the MSP/Max software - no go..  I gave a serious attempt at Karma software by Steven Kay, which is brilliant too, but I could not get it.. I'd come up with 'happy accidents' once in a while.. Often I come up with 'happy accidents' in Synfire Pro.. but I want to know understand the process, so it's just not a rote pushing of buttons.

Being mostly self taught - I spent a few semesters at Berklee on line music, taking songwriting, arranging, writing in a perticular style.. Some of this I had learnt by myself, but the good thing, is now I understood why a certain procedure worked, and I know what to do to evoke a certain emotion and I don't have to relly on endless takes and trying of ideas to reach my goal.    

(I'm still nowhere close to that with  SFP.. The things I've uploaded are 'experiments'.. I am certainly a better musician, but felt I should only upload SFP projects here. The more people who buy SFP/SFE are going to be coming from the same place.. They want it to work like their MPC box.. They want it to work like Band -in-a=box. They want it to work like Cubase, Logic, Digital Performer.  You are going to have to expand your resolution on what SFP is and what it is not and make it clear.  

Indeed one logical conclusion would be to turn it into an ultimate DAW.  That was not your original intention. You have made serious productive modification to bring it into line with what we are asking here.  Prado's question was indictive of what I might have asked last year..  To be honest though,  I DO NEED it to talk to Logic (or any other DAW).  Cause you need to get into the individual midi notes, alter velocity, duration, start  time and length note by note to come up with detailed results..  Then while you're doing this, it dawns on you, you need to duplicate some sections, transpose, provide counterpoint, harmony.. Now does one go back/forth between SFP and DAW? 

Cause I cant use drones. and rewire IAC is not desirable to me.. I keep a SFP  version of song and Logic version.  I go back and forth.  If I need to add a part beyond my technical skills. I go back to the SFP version, create it and import into Logic.. After enough changes, I export the new Logic song back into SFP and work from there..  Laborious for sure, but the only way I can find right now to accomplish my needs..

We are going to have to grope our way thru the darkness until we see the light as precieved as Andre.  Unfortunately which seems to be the case, your updating of the GUI, decreases the value of videos you've already needs.. There is nothing more irrating than following some directions but the visual imagery you see, does not apply to what's on your screen..  Apple is pretty bad at it.. You go to do something with the network settings, or email, and the videos online are for previous OSX's and doesn't correspond  to your software..

The world is getting more and more hectic and complicated by the second.  You are going to have to make adjustments, because you can't expect a user to be willing to put in months and years of serious study to understand your software.  Not unless you get a Sugar Daddie company to fund your R & D.  But you've already mentioned your feelings on that.. 

After all the work I put into my last musical upload, serious dealing with SFP.. I'm going to relax and do a piece the old fashioned way, put Logic into record and edit that.. Making music uses more different parts of the brain than most other activities.. But when we take the physicality out of it, making it totally cerebral, it begins to be something else.. If I run into a technical snag now, I try to find another way to solve the idea, or drop it, rather than go one 2 hour detour to find the minute technical details, I am lacking. 

I have a lot of faith in you, SFP, and your company.. Don't be discouraged, when new people ask questions, which implies a lack of understanding of your basic principles. Seriously a lot of people here have been making posts, which gradually clears things up.. When people keep asking the same question, it's an indication, of you explaining it better, or modifying the code.

Respectfully yours, 

Firstly, I would say that SFP is without a doubt the greatest software known to man for creating music. In my two and half years of using the software almost daily it has given me more insight into music and its creation than i ever thought possible. That being said. I think a Vst version would not be dumbing down at all. Vsts like Reaktor, Max for Live, Rapid composer exist in the Vst environment and thrive there. Ultimately SFP will have to either become a DAW or become a VST to reach a larger audience. The recent progress with drones in SFP is much welcomed, all of the changes in the last few years have been amazing. I do understand that SFP is trying to reach the large scale composition film work type of guy, but even someone like Hans Zimmer uses cubase. Setting up two different DAWs kills creativity. that is the problem. Constantly tweaking two different programs and routing kills creativity. I need one DAW that i can open that has a template setup that is ready to receive ideas as they happen. 

At the present some things are easier to do in a daw and others more easy in synfire. Recently the shift is moving more and more towards synfire for many people's workflow. It's now possible to create a song entirely in synfire using the engine, most of you vsti instruments, using vst effects, mixing in synfire and recording the output to file in the engine. The only thing I'm not sure is if it can handle audio in from external sources ( feel free to correct me if I'm wrong)
The issue with the last statement is that some of the workflow is still considerably easier in a daw and this coupled to the fact most of us still have a daw way of thinking ( natural given where we have all come from).
I guess the objective for Andre is to improve the user interface and hence the workflow further so that it becomes easier for people to use synfire. However there comes a point where people will not drop their favourite daw, so it has to be designed to work with multiple daws.
It might be too much of a development effort to produce and i woukd rather see that tine spent on the current products but a cut down synfire vst probably would be more successful than synfire express or harmony navigator.

Thanks for your replies.

Logic and ProTools can't host VST and there is no way to route MIDI outputs. Limiting Synfire to the VST platform would cut into its reach. We are rather looking to also support RTAS and other formats to further extend its reach.

A Synfire VST would need to read/write a hosts's MIDI data, because that's where recorded data goes in a DAW and where users expect to see it. Moving Synfire entirely to inside the DAW would not bring advantages, if it does not integrate significantly better with the DAW workflow than the Drones currently do. 

Although some of the limitations of VST GUIs, like window resizing, could be worked around, workflow and interactivity would still be limited. Freely swapping data between projects is a core feature and major USP of Synfire. With Synfire Pro, one can drag-copy entire multi-instrument sections between projects, while all duplication of plug-ins, presets, sound assignments and routings is managed automatically ("Container Import"). Two hours worth of fiddling done in a second. Building a new song from several existing unfinished fragments is a very common workflow. Most importantly however, this is what sets Synfire apart from all other music software.

VSTi can be multi-timbral, but this does not free you from patching all MIDI connections, tracks and plug-ins manually. Setting up a song this way is static and a one-way road. The more your setup grows, the less likely you are going to make any substantial changes anymore, even if the artist in you feels a strong desire for that.

During the prototyping stages of a project, using DAW drones (or a Synfire VST for that matter) is inferior compared to the Engine anyway. I can not stress this enough. In addition to the technical hassle with ASIO drivers, input monitoring and MIDI input quirks (Windows only), workflow-wise this is pretty much like breakdancing with handcuffs.

I'm not saying it would be impossible to make a VST. My point is that the few benefits it might yield would not justify the losses, compromises and huge investment that is involved.

The reason this pops up is that the current GUI is not yet intuitive enough in the way it visualizes the setup to the user. There is still some fundamental confusion about the way this works. This will be addressed.