Dark Lady PM guitar/bass vsti

[shogger]

Even if releasing the current SE version is tempting, my "do it right or don't do it" moto still stings a bit

After hearing the examples I agree whole heartedly. And the bass is way better than the guitar thingy. I like the bass and would be interested in it.

Shogger

[SciFiArtMan]

Really nice development and progress Mokafix! The guitar may not be completely there, but sounding very usable and better all the time. The bass is exceptionally nice! You're definitely on the right track and making great advances! All the best!

SciFiArtMan / Neomorphic Giant

[datapark]

Moka, please. Just release it.

Consider it a "beta".

Please. You aren't going beat any company by just throwing teasers at the public. Give us a real taste by letting us play with it.

[Mokafix]

Thanks forthe interest kept on that plugin, despite te lack of release...


A few things:
- who told that my plugin has no midi implementation? It has. there are 4 midi modes to suit different needs :

multichannels (very handy for sequencing especially in hosts that let you assign midi channels to each notes like Reaper).

Live1 = the strings are choosen by hand position on the fret board. The hand position can be set to "automatic" : if you play a note that needs to be done down the fretboard, then the hand will stay there for a little time before going back to it's default position (that can be set too).

Live2 = you can switch between "forced polyphony" : if you hold a key, the next note will be triggered on another string if possible (of course, if ou do E2 and F2, it'll "hammer on" on first string unless second string is tuned to F2); "forced monophony" where legato always make a hammer on or a slide on one string.

Live3 = a very diffeent approach: left hand is used to choose notes, right hand triggers the strings (very guitarish approach). If you press one note with left hand, you can play it on any string by pressing string keys with right hand. if you press 2 notes, the first one is assigned to string 1 , the second one to the rest of the strings. the best is to press 4 keys all the time, knowing that the lower key is assigned to the first string, etc. What's cool with that is that it makes it easy to make arpeggios and let you choose the strings very naturally, and very precisely, without letting the choice to any automatd stuff. It requires some practice though.

What you will never be able to do with it : make a chord with more notes than the number of strings (4 for the bass version), because it's just not possible in reality, and I have no way to make it possible here either without adding extra strings.


The truth about "round robin":
Round robin fakes something natural in a non scientiic way. Why? because there's absolutely no randomness in a guitar. Everything happens logically. It is a very straight instrument.
If you play a note with realstrat or whatever sampler, release it (until it stops sounding), then play it again, it should sound exactly the same, otherwise it's wrong. A real guitar will always sound exactly the same unless you change your picking position from at least one centimeter OR you keep picking constantly without muting the string in between (which happens often). My plugin have a "random picking pos" knob, and the fact the sound changes if the string is still moving when I pick again is taken into account. It needs to be adjusted though for some tricky reasons that would make you sleep if I tryed to explain. I thought about making it even more real by making the string mute on key press and ring on key release, but it's a bit painful to cope with the simple muting of one string then (maybe some dead keys could mute each 4 strings individually then, like you would do with a real bass by muting with the right hand à la Jaco).


My plugin is really like a guitar inside a computer, so it is much more painful to make a midi implementation for it than for a sampler, because it "reacts", it's alive, and doesn't just play a sound on demand.
But at the same time, it offers stuff you will never ever find in a sampler : try to make the "loser" slide riff with any sampler, and it won't sound better than mine, I'm quite sure about it. It will never ever ever be possible to make realistic slides with a sampler, period. it's just technically impossible. You can probably make ok slides, but that's all, and there's no way you'll ever do bottlneck, because the sustain wouldn't be long enough. In dark lady, sliding feeds the string with some noise, so you could slide forever, it would keep ringing, though in the end, the tone would not contain anything from the first impulse, only bottleneck contact noise.


This plugin doesn't fake an instrument, it is an istrument in itself, so it will be a bit more demanding for the player than a sampler, but will give probably more lively results.

Also note :
All demos I did are sequenced, not live played, so despite the fact I tryed to randomise the playing a bit, it's not like playing it for real.

[koolkeys]

Impressive Mokafix, though I wouldn't advertise it as "guitar VSTi", for this it just sounds too crappy IMO and RealStrat is still a killer in this section (also in terms of "playing" - as user of both RealGuitar and RealStrat it was like a holy grail to finally have access to "easy" keyswitches, unlike LPC or Chris Hein), though RealStrat lacks the "pickup change" feature.

The Pickup Change feature is actually something that made me think "dude, when do I get my claws on it? And how much will it cost?" - among all bass VSTi's and samplepacks, this could be a killer. Though I kinda miss round-robin in terms of sound. The fast play sounded totally static to me. Then again I know it's still in Beta Phase.


I'm looking forward to this VSTi - not for Guitar, but for Bass. And I really hope Musiclab adapts the idea of "changing pickups" - and maybe you also think about adapting plektri rather than finger pickings alone.

The difference is though, that Musiclab can't possibly change pickups any more than the samples they have recorded. They can't adjust all the variations in between, which you can do with PM.

Samples are great, and in static situations they are the best sounding. But in order for them to be able to sound as realistic as possible, you need to record each articulation from each string. With PM, you don't need that.

Samples and PM both have a place of course. But in this case, samples just won't do for pickup changes the way that PM can. And that is a good thing that there are/will be options.

Brent

[griels]

Aggh! Too many cool bass/guitar emus!

This looks to be a frontrunner whenever it comes out though - those slides are gorgeous

[Mokafix]

The difference is though, that Musiclab can't possibly change pickups any more than the samples they have recorded. They can't adjust all the variations in between, which you can do with PM.

In fact they could... (Line6 did it with Variax and roland with VG88)
But they didn't. the question is why?

- they are too lazy.
- they want to make it an update you'll need to pay for. (or maybe they want to make "real-LP", "real-dobro", etc, as completely different instrument rather than leting one do everything with a single plugin)
- they don't know how to do it (if they really don't know I'm so sorry for them as it is really easy and well documented on the web).

Honestly, I once wanted to prove that my technology is better than the variax, because variax and VG88 need to "remic" an already miced signal, while I miced a string directly in my plugin. So I did a set of samples (bridge, mid, neck), then did another set using variax technique (remicing a bride pickup sound), and the sounds were quite similar. So this technology used in VG88 and variax for pickup virtual positionning as a post effect is really trusty, even if fake theorically speaking.

hmmm... I wish I had a big company. I have ideas for a very good guitar sampler too. Not a thousand Go stupid beast, but a very smart one that uses maths to reduce samples number and increase versatility. It would have fewer samples, would let you change pickup positions, pick position, tuning. Then add to that a good body emu and a smart midi implementation, and you got the finest guitar sampler ever, an hybrid of sampling and PM = less cpu than PM, and less ram than sampling only. Still no good slides, but at least something miles beyond realstrat and the others.

[MaliceX]

Well if this does get released, if it's going to be a 'buy' app, how much are we talking?

[Highvoltage]

where are the audio demos mp3? reports a 404 error

[Compyfox]

In fact they could... (Line6 did it with Variax and roland with VG88)
But they didn't. the question is why?

- they are too lazy.
- they want to make it an update you'll need to pay for. (or maybe they want to make "real-LP", "real-dobro", etc, as completely different instrument rather than leting one do everything with a single plugin)
- they don't know how to do it (if they really don't know I'm so sorry for them as it is really easy and well documented on the web).

Okay, stop it now will ya? Your recent posts showed enough how arrogant you are compared to other companies. You insult them that you made a way more "superior" instrument than others and statet that over and over "what" you can do, and how better you can do it... with SE (unless you switched to C++ now).

It's enough! I'm a betatester of RealStrat and I know what this thing can do and what it can't do. And compared to YOUR demos, it sounds way more alive and versatile.

"A guitar string sounds exactly the same if struck again - round robin is just plain wrong", "other coders can't pull off what I can cause they're lazy".

Do you actually know that this is the case with other firms? Do you know how much work was put into the plugins BEFORE we could get our claws on them? Have you even PLAYED RealGuitar before and compared it in terms of playability and versatility to what's out there (where it's really a winner)? Do you even KNOW the fact that RealGuitar was once released with just a handful of Guitars only and later upgraded with 2 more "for free" to registered customers? Who said that this won't be the same with RealStrat?


Stop acting so big, Mokafix. This is what I disliked about you while you were still working on your Redrum clone (if you could call it working on v4 - it was rather giving it away cause you didn't like to optimize it anymore) and acted towards me like I was just a stupid fag because I couldn't "recreate" in SE what you did and thought "it's so easy, it's obvious". You didn't even release your "guitar/bass" thing yet, and so far it was only good sounding on bass, if you can call all those audio snippets and YouTube videos of high quality.

In case of RealStrat... you're still milestones away to call your VSTi superior, or call other plugins "bulky", "too CPU/RAM intensive" or "crappy concepted". RealStrat is less than 1GB with tons of keyswitches, articulations and seperated recorded singlenotes on each fret, RealGuitar2 less than 600MB with 6 guitars included - and compared to other packs like Steinberg, Chris Hein, EastWest, Prominy and how they're all called, this thing is still the best in terms of size/CPU usage/usability compared to their competitors.


You're not on the level of TruePianos, nor PianoTec yet. And unless you can't (or won't) proove that, and I couldn't confirm it myself that your "thing" is indeed superior, you should better be a tad more quiet towards your colluegues, how bad they are compared to you. Such additude is just pissing me off!


Just my two cents.

[xoxos]

if your analysis didn't have $150 for a single instrument (?!) you'd probably be pretty happy with a simple dynamic solution it's all perspective, friend.

sorry for saying stuff to you.

[Mokafix]

Never said my plugin was superiror; it's just that knowing what I know (and only learned by myself and from some websites), I just find it too bad there's nobody making a variax like sampler, that's all. I mean the technology (at least pickups positionning - body emulation is another thing) is really simple, so I'm just a bit frustrated.

But you're right, my plugin IS bulky, and cpu intensive , which is why I take so much time releasing it. And in fact the tech behind is so simple that I don't really feel like it is anything to be proud of.
That said, I demoed revitar and heard slayer (the only two vstis I can really compare with, as they use the same appraoch), and I think mine is better; so let's say I'm "glad" to obtain the results I obtain using this approach.

About round robin, the guy who sent me a "various attacks" sample had to do another take without muting between each attacks because it sounded all the same and he wanted to show me how different it can sound in another situation (no muting between strokes). Not only was it sounding all the same, but it looked the same at cycle length scale too (even the very small harmonics).
I just thought this info was interesting, because it goes against a common belief that real instruments are very random.
In fact I implemented round robin too, then tried to make the randomness happen "naturally". Now I might come back to a kinda round robin anyway, because it's easier to do, doesn't require much midiwise and does the trick quite well.

Realguitar seems very nice, but I'm in my PM research, so sampling is miles away from my mind. And I know the limits of each technique. There are some things a sampler just can't offer. I'm trying to fill that gap using a technology that has other drawbacks (cpu, difficult midi implementation). It is a complementary thing, not a replacement of all the guitar plugins.

So I'm sorry if I sound arrogant sometimes... I know I am, like we all are when the one we really try to prove something to is ourself, aren't we? I have absolutely no "real" support for this, so the support I get from this thread is the only thing that keeps me going on. That's why I sometmes need to post something new, have some hugs, smiles, thumbs up, and keep tweaking. That's quite primal behaviour, but I know I'm not the only one on this forum . If I was so sure about my skills, I would never make a thread and show off.



to MaliceX :

I don't think it will be expensive. Of course, if someone helps me doing a c++ version, the price might rise a bit, but I don't want to make it expensive. So something between $30 and $50 probably (as long as the price decision depends only on me)


to Highvoltage :

I had to remove some demos, because my web host is clumsy and it seems I reached some space limit.
This mp3 is still there though.

[Highvoltage]

so can you post a demo of a fast up/down stroke on one string? That is the thing that hasn't been emulated so far, and i think its one of the most distinct character of a real guitar sound.
At a faster picking speed it gets more and more chaotic and random, each little picking noise is slightly different, that makes its cool organic sound. At least i think that.
An interesting aproach could be to sample only the picking noise on every string (if its possible at all?) and mix it with the PM string decay. Those are so short samples, one could sample thousands of them. Would be a rather big job, but it could work.
Cause all honesty i think if you apply distorton or amp sim to your guitar, the synthetic picking noise gets exaggeratet, and you wont get that hard metallic sound of a real attack. dunno.

great work btw.

[Mokafix]

In fact, the impulse isn't synthesised. It is a very very short sample. Then it's sent into the string model.

Here are two files :
Guitar_pickingtest.mp3
Bass_pickingtest.mp3

The guitar is a picked strat with bridge pickup, the bass is a fingered Jazz bass with bridge pickup too. Both are picked right on the pickup.

I sequenced the midi in reaper, and randomised it by hand (notes velocity, positions, lengths). You can hear the way the sound vary for each stroke without any change in the impulse.

For the third part in the guitar file I just copied the same pattern for each string, and offset it a bit to simulate chord striking, so it's all donward chords.

[MaliceX]

Just a question, is it possible to do the characteristics of "Slap Bass" on your PM guitar thing?