Hosts and Religious Sensuality

[RTaylor]

See , now I had to type all that crap which should have been obvious to you right from the start. But apparently it wasn't. Do you still think it's possible?

I know it's possible. You may have to change "bass guitar" to "track 3" but it's not anything that you can't handle. If you look through the links on that page... there are all sorts of programs that will do it.

eXT simply can't be controlled externally in that way - and I don't know any sequencer which can - neither by midi nor by key-commands... a few commands are obviously possible (start, stop, record - commands like that) - but that's not the real deal I was speakling about...

The idea of building an interface to control your sequencer means just that. How you go about that is beside the point. There are all sorts of macro processors that could make your life easier... you might even be able to implement it in cal on Sonar. {Or Python on anything}

We're talking computers here and nothing that's really all that complex. If you can do simple commands you can do complex ones.

[RTaylor]

Hosts.....

That's purely a Catholic thing, not?

Victor -- blithely ignoring the discussion

No. ...Though I think they get the good ones.

[jens]

The idea of building an interface to control your sequencer means just that. How you go about that is beside the point. There are all sorts of macro processors that could make your life easier... you might even be able to implement it in cal on Sonar. {Or Python on anything}

*sigh*

one last time: as long as the sequencer is not able to receive external commands to directly select certain tracks, parts, regions etc. it is not possible.


you seem unwilling to get my point that the sequencer itself needs to have the logic implemented which is needed for this kind of control. And my example was relatively simple.

[Sascha Franck]

Jens, while it might not be possible in Energy XT (which is entirely eXTs fault), it might very well be possible in others.
For instance, in Logic you can assign like *any* operation to a keycommand or a MIDI controller. Ok, "Track 3" might not work (there's just "next" or "previous" track I think) yet, but as far as the rest goes, things should be possible.
All it'd take would be some voice recognition program to translate things into keyboard strokes. Not all that wicked, when you think about it.

[Cordelia]

You're just talking about tools. I think we need a revolution in thought... in art. Tools are just tools.

Now, see, there ya go.

[jens]

Ok, "Track 3" might not work (there's just "next" or "previous" track I think) yet, but as far as the rest goes, things should be possible.
All it'd take would be some voice recognition program to translate things into keyboard strokes. Not all that wicked, when you think about it.

*sigh* and another *sigh* - I have already answered this in detail - 'next track' and 'previous track' is possible in eXT as well - but where's the use in simply translating voice into keystrokes? Not all that easy if you think about it.


Why don't you lot get that for this to make any real sense and to bring any real advantage to the user another level of interactivity would be needed first?

The sequencer would need to know e.g. on which track the bass-guitar is, where the first bridge is, the second chorus, or maybe the coda, the outro, whatever.

Right now all sequencers are just remembering what is where without any actual comprehension. You could say the sequencer is to stupid for what I'm proposing. I'm quite sure one day we will have this kind of interactivity where the sequencer really startts to think and learns from the user (adjusts to his way of wroking like a real assistant would do) and voice-recognition built into the core of the sequencer itself is just one part of it.

If what I was would already be possible why isn't everyone already using these voice-recognition programmes then?

I'll tell you why: because as it is now it's not worth the hassle.

However the way I mean it will bring a terrific evolution to homestudio workflow.

We'll dig out this thread in a few years when what I mean is a quasi-standard as is e.g. freezing nowadays...

[Sascha Franck]

Jens, no offense meant, but you seem to underestimate the power of keycommands.

The sequencer would need to know e.g. on which track the bass-guitar is, where the first bridge is, the second chorus, or maybe the coda, the outro, whatever.

I am only agreeing on the track part. This is impossible indeed.
But, saying something like "first verse" and having whatever speech recognition software translate it to a keycommand which would then just trigger a "go to marker 1, loop selection and play/record" shouldn't be much of a problem. And the same might even go for "select track number *" (not sure whether such a command exists in Logic).
Of course this would have to be readjusted for each and every song, but that'd be the case with an integrated solution as well, simply because songs are varying too much.

Oh well, sure, it'd be *way, way* easier if such things were implemented straight in a host. But that doesn't mean it's completely impossible yet.

I'm however not sure whether there's a speech recognition which would actually do the translation job. I wouldn't want to say "Control M 1" (for "go to marker #1") either.
But that should be quite a minor deal for a programmer, assuming the actual recognition process would be working fine.

[Stupid American Pig]

the next revolution will be-









reel to reel tape.

[Al Magnifico]

I think, the clue is, that the speech recognition program has not only to translate commmands like "go to marker 1, activate record on track 3, loop 4 bars and start!", it also has to understand propositions. Imagine, I say "Track 3 is bass", "intro goes from Bar 0 to 5" or "new aux bus with bionic delay", then I could say now: record bass intro, add some delay. A kind of variable assignment.
Could be an interesting project... on some far day, when the hosts finally had learned to handle all the hardware control surfaces

[jens]

Of course this would have to be readjusted for each and every song, but that'd be the case with an integrated solution as well, simply because songs are varying too much.

no, I don't think so... - when an assistant can learn and understand what a 'chorus' is then why not the sequencer? We would need to be able to set user-definable atributes for sections of our music. Let's call it 'object orientated sequencing'.

In other words: sequencer with some kind of artificial intelligence could be the next big thing. It would rather be a revolution than a mere evolution though. Voice recognition would be a part of it. I'm convinced if someone smart from Cakewalk or so would read this post he could make big bucks from it. But when I posted a similar post about two years ago or so in the Tracktion forum (without the voice recongition part though) the reactions were similar to this thread.

Common sense wasn't far enough yet and apparently it still isn't.



edit: typo

[jens]

I think, the clue is, that the speech recognition program has to not only to translate commmands like "go to marker 1, activate record on track 3, loop 4 bars and start!", it also has to understand propositions. Imagine, I say "Track 3 is bass", "intro goes from Bar 0 to 5" or "new aux bus with bionic delay", then I could say now: record bass intro, add some delay. A kind of variable assignment.
Could be an interesting project...

Bravo - you seem to be the first one in this thread who really got what I mean.

[jens]

Jens, no offense meant, but you seem to underestimate the power of keycommands.

nope - I use them daily - hence I know their boundaries...

[Al Magnifico]

you can call me Urgestein

[jens]

o.k. then Urgestein...

[RTaylor]

Jens, while it might not be possible in Energy XT (which is entirely eXTs fault), it might very well be possible in others.
For instance, in Logic you can assign like *any* operation to a keycommand or a MIDI controller. Ok, "Track 3" might not work (there's just "next" or "previous" track I think) yet, but as far as the rest goes, things should be possible.
All it'd take would be some voice recognition program to translate things into keyboard strokes. Not all that wicked, when you think about it.

The logic doesn't have to be built into the sequencer. You're building an interface to the thing.