Vst's PM Or FM or both?

[Kel]

Hi. Thank to all the reponses in my previous post about Phys modeling synths.

Been lurking about VSts lately, What is in your opinion more versatile for doing Film oriented patch design?

You know you can use PM and get more or less realistic results from it, but as you may know there are FM synths out there which can arrive at very similar results, for example sytrus got 64x oversampling+ harmonics editing +karplus algo+ damping in the operators so that make it as a hybrid. I mention sytrus because is the only one I have used over the years and have got experience with it. (appart from other subtractive synths, which are not the question here).

My question is, is FM more limited than Phys modeling when we speak of Recent FM synths not old DX7?

There are very few phys modelling out there which are not too expensive, and there are a lot of FM synths which are more versatile than the typical FM synth.

I hope my question is clear, sorry spanish speaker here I try my best

[foosnark]

I wouldn't call it "more limited" but it's like a hammer and a screwdriver. Sometimes you can use one to do the other's job, but it's going to be awkward and not work very well.

[cornutt]

I would say that FM is less limited than physical modeling, but it's also a more difficult design space to work in. A lot of the FM design space leads to sheer noise, and it's more work to get useful and non-cliched sounds from FM.

[Lotuzia]

Try NI FM8 : Depth in programming, enough operators and connections to keep you busy for a while. Smart envelopes.

In theory FM/PM with a lot of operators should make you able to approach any sound ( and would come close to FFT actually ) In real world, like with additive synthesis, tweaking all harmonics and give them the right property/EV/PitchDev etc etc would take a lifetime ...... per sound. So like all synthesis methods, these are shorcuts.

However FM8 is a captivating synthesizer and offers a lot

[Kel]

Try NI FM8 : Depth in programming, enough operators and connections to keep you busy for a while. Smart envelopes.

In theory FM/PM with a lot of operators should make you able to approach any sound ( and would come close to FFT actually ) In real world, like with additive synthesis, tweaking all harmonics and give them the right property/EV/PitchDev etc etc would take a lifetime ...... per sound. So like all synthesis methods, these are shorcuts.

However FM8 is a captivating synthesizer and offers a lot

Thanks.

Well the only way to get a faster approach is analising the spectrum and try to mimic it as near as possible, then you tweak the dynamics of said instrument. Even with this, it will always sound different.

Then there are others like Zebra or Blue.

I may purchase Sakura (or any other pm synth) in the future which makes the work way easier and its got a very simple GUI with all needed to model. Using FM to model instruments is hard and time consuming, I mean, only the dynamic parts, because finding the harmonics is easy.

Lets say you want to do a banjo:

-get a sample/youtube sound whatever of a real banjo for reference.

-press C5 on your keyboard

-Tweak till the end of time

And then the dynamic part, that is when My brain explodes. That is where difficulty is. I do it for fun as I do music for hobby.

The only good (of those free) PM modelling synth which works like a charm and is easier than FM is Ugo's M-theory and as it was made with synthedit it has problems with multicores so its unusable for me, even renaming dlls crashes a lot, a shame.

I may put my sytrus patches, synth1 and oatmeal. Sometimes there are things that cant be done with a few osc, and you need more osc, even if you have FM available, like in those in substractive synthesis. As I see there is no difference from operators from an FM synth to a simple substractive synth because at the end of the day you just put together a bunch of harmonics and make that sound... the more Oscs/operators, the better.

[V0RT3X]

Zebra is good for creating PM style sounds and it does FM.

[Kel]

I would say that FM is less limited than physical modeling, but it's also a more difficult design space to work in. A lot of the FM design space leads to sheer noise, and it's more work to get useful and non-cliched sounds from FM.

do you think is less limited?

I always thought the opposite. Because FM is only a fast approach to make the trick into the ear that a bunch of harmonics sounds good enough to represent an instrument. PM is using all the (virtual) semi-real properties of those instruments.

But I have to admit sometimes I get almost exact sound using only FM that surprises me, the most important aspect for the dynamics part in FM (as I see it) is done only with envelopes and filters. you want to do a (lets say something similar to sampling here) round robin there is no other way than using random velocities, phases, different ADSR, and pitch for each operator.

Cheers.

[realmarco]

the whole point of physical modelling was to emulate real Acoustic instruments back when Samplers costed a fortune and had only
1meg of ram memory

now we have Multi-gigs of Ram

So PM is quite limited compared to sampler VSTies

[pdxindy]

So PM is quite limited compared to sampler VSTies

I'd much rather have a PM synth than a multi-gb library... and I think the sample library is much more limited...

[V0RT3X]

I'd rather record actual instruments.

[stimresp]

Blue II allows you to switch between 'pure' FM and 'PM' FM. Nice option - they do sound different. Subjective of course, but I find PM to provide more 'musical' results.

Blue II is for me is the current pinnacle of FM synthesis, but it doesn't stop there. The demo has tons of Rob Papen patches - all neatly organized - that give an excellent impression of what this synth can do.

[Xenobioz]

Blue II allows you to switch between 'pure' FM and 'PM' FM. Nice option - they do sound different. Subjective of course, but I find PM to provide more 'musical' results.

Blue II is for me is the current pinnacle of FM synthesis, but it doesn't stop there. The demo has tons of Rob Papen patches - all neatly organized - that give an excellent impression of what this synth can do.

But does that PM really stand for Physical Modeling or is it just Phase Modulation?

[stimresp]

Blue II allows you to switch between 'pure' FM and 'PM' FM. Nice option - they do sound different. Subjective of course, but I find PM to provide more 'musical' results.

Blue II is for me is the current pinnacle of FM synthesis, but it doesn't stop there. The demo has tons of Rob Papen patches - all neatly organized - that give an excellent impression of what this synth can do.

But does that PM really stand for Physical Modeling or is it just Phase Modulation?

Oops my bad - just re-read the OP. Yea, it means Phase Modulation - not what the OP is looking for.


(Though there are 2 comb filters )

[Kel]

Multigigabyte sample libraries may sound incredible, but if you want an especific articulation that is NOT inside the lib, you are dead. If you are lucky you may alter the sample using somekind of script (like kontakt) and get a good enough result.

That is why I prefer FM, PM, or whatever... expresiveness.

Thanks for the idea about Blue 2. Comb filters, yep, those are interesting. I will have a look, Rob Papen got my respect

[himalaya]

So PM is quite limited compared to sampler VSTies

I love the way I can edit various aspects of the sound on a PM synth, like the way the finger affects the fret board on a string instrument, which is not possible on a sampled sound. With a synth like the String Studio VS-1 (or VS2 now), it's so satisfying being able to control these elements.

This gives way to cool new hybrid sounds, which is PM's strongest aspect.

However, I really enjoy layering sampled and PM sounds as this gives a richer sound in the end.

There was a recording where I needed a harp-like sound. The PM harp sound was too 'cold' or 'flat' on its own, and the sampled one was too 'real'. But layering the two worked a treat.