The Fight for FM

[z15]

Replicating a VOPM preset won't address the issue at all, but a completely different sound that has that firm, "real" quality to it would address the issue precisely.

But if someone makes a sound with VOPM (or any reference synth x people think is awesome) that people agree has that "real" quality (whatever it is), and the replication of that sound in another synth sounds the same and has the same quality, doesn't that prove the point that it isn't about the intrinsic quality of the synth but about being able to program a synth? Or am I missing something?

That is, there's a reference sound that has qualities that some people are after, and some kind of recipe to get that sound programmed - and if the end result is like the reference sound, mission accomplished? And if it the end result doesn't sound the same, it proves that there's something in the signature sound that the reference synth has that the other one doesn't.

The problem with the "completely different sound" test is that there is no reference that people agree on. Someone can say "well, this sounds good" and someone "well, this doesn't have the quality I'm after". But if at first some people agree that some reference sound is what they're after, replicating it with another synth proves the point that the same sound is achievable on that synth.

[living sounds]

Because that's how it always ends up: someone will eventually sell you something you believe in, it doesn't matter if it's true or bs.
There's panic about cellphones being harmful? Here, take this little piece of plastic for $50, its magic counter-rays will protect your brain from all possible cancers. And it will make your calls sound better.

I saw a recent thread about someone saying he tested expensive EQ plugins and concluded it was the biggest con. Well, it's HIS fault in the first place. He thinks/thought that EQ's can sound better? There will obviously be people to sell him a more expensive EQ.

Having actual mixing and producing experience with all these tools for years and a very refined sense of listening for it (through training and upgrading the monitoring) has nothing to do with gullability. As a skeptic I'm actually sorta insulted to be put in the same categority with that kind of thing.

Again, all you bring to the table here is ridicule. So there's not much point in discussing anything with you. If you don't know and don't have any means to find out what is different yourself, why complain that others "don't want to find out"?

If you're a software developer, sit down and try to come up with something that sounds as good as that ancient hardware, don't complain. If you don't hear these differences, maybe developing music software is not for you.

BTW, the envelopes in my old Roland samplers sound WAY better (punchier, faster) than any plugins. So you've all still got some work to do. Don't come up with straw men like "it's all subjective". Because you're not being scientific here, you're being ignorant.

[Aroused by JarJar]

Replicating a VOPM preset won't address the issue at all, but a completely different sound that has that firm, "real" quality to it would address the issue precisely.

But if someone makes a sound with VOPM (or any reference synth x people think is awesome) that people agree has that "real" quality (whatever it is), and the replication of that sound in another synth sounds the same and has the same quality, doesn't that prove the point that it isn't about the intrinsic quality of the synth but about being able to program a synth? Or am I missing something?

You lost me there- if another synth has that intrinsic quality to the sound, then it does and that's it, it doesn't matter what the "picture" of the sound
is like. A synth or programmer could completely fail to replicate a sound either because the features or the skill level aren't there, that's not relevent to the essential quality of the sound.

As evidenced by praise for the VOPM, which is clearly far, far less capable as far as features than the FM8.

That is, there's a reference sound that has qualities that some people are after, and some kind of recipe to get that sound programmed - and if the end result is like the reference sound, mission accomplished? And if it the end result doesn't sound the same, it proves that there's something in the signature sound that the reference synth has that the other one doesn't.

The signature sound is evident regardless of the patch (within reason, you could program any synth to sound like garbage I suppose).

[z15]

BTW, the envelopes in my old Roland samplers sound WAY better (punchier, faster) than any plugins. So you've all still got some work to do. Don't come up with straw men like "it's all subjective".

But isn't this a chance to prove him wrong? Get a sample, run it through an envelope in a Roland sampler, post it and the original sample, and let tony run it through a software sampler of his choice. It doesn't take too many minutes for you and you can both post the results for people to hear. Then you'll have some solid basis for making conclusions, other than "there is a difference, whether you believe it or not" or "there can't possibly be any difference".

Ditto with FM, program a patch and record it, then send the patch data in some form to replicate. And any other points of debate.

[tony tony chopper]

THAT's sound that sounds like sound and not a picture of sound

No idea what you're talking about.
Is this philosophy now?

[nuffink]

Living Sounds, and others, have specifically pointed out, a number of times, VOPM

It rather proves my point, that VOPM is supposed to be an emulator. FM7/Sytrus/others are not emulators, they're synths with a lot more capabilities, that have rough import of DX7 settings.
So you surely CAN achieve the same as the original patches you compare to, but it will require tweaking.

I know I couldn't import DX7 features properly by the lack of documentation. Algo's & levels, no problem. But the envelope times were encoded in a special way, I found no info about this. Meaning: if you import a DX7 preset into Sytrus, it will certainly never sound like the original. You can either conclude that it's because it's software & lacks of pixie dust, or you can investigate & find out that it will be the same with some envelope tweaks.
Now there's one DX7 envelope feature that can't be replicated, though. I remember when I read the specs, that the DX7 release was not for a defined length, but down to a level, thus a lower velocity would sustain less (I don't remember the details), and this was not doable with my envelopes.

But yes, I'd like a challenge. We'd need a minimal VOPM preset that supposedly can't be achieved in FM7 or others. I'll try to replicate it in Sytrus.

But your point is sidestepping the real issue. I'm sure every person here who is disappointed in the sound of the FM8 for example will agree with me when I say that the real point is: a sound should sound like a sound, and not like a picture of a sound.

You can always make any sound more pale or pastel, or puffy or distant or weak or whatever. What's desired is a sound that is solid and somehow "real", which can on one hand function with zero effects and on the other endure all kinds of manipulation before fuzzing off into feebleness. The sort of "platonic ideal" or "world champion" of this, and once again I'm dead sure that everyone with years of experience will agree, is a clean mono acoustic recording with a strong midrange. That's about as strong and straight as it gets, and the experienced know that it's the most durable and flexible as well. THAT's sound that sounds like sound and not a picture of sound, and THAT's what we want from synths as well (and it's been done countless times so it sure is doable).

Fundamental, intrinsic quality.

It's quite simple.

Replicating a VOPM preset won't address the issue at all, but a completely different sound that has that firm, "real" quality to it would address the issue precisely.

Edit: Happily when using my new USB interface, my browser doesn't grab the audio driver so I can WWW as the dancer practices, heheh! So I'm downloading the Sytrus demo now and will check it out ASAP. I don't give a flying f**k as to whether it sounds like a DX7 or not, but whether it has that essential kernel to the sound (and whether the color is to my taste).

So, a non-visual sounding synth that reminds you of a clean mono acoustic recording with a strong midrange?

I can understand why you're disappointed, but I doubt you should be presuming to claim you know what "we" want.

[z15]

You lost me there- if another synth has that intrinsic quality to the sound, then it does and that's it, it doesn't matter what the "picture" of the sound is like. A synth or programmer could completely fail to replicate a sound either because the features or the skill level aren't there, that's not relevent to the essential quality of the sound.

What I mean is that we need the same "picture" of sound to compare with, otherwise it's a subjective argument and we can go arguing about it for the rest of the year or just stop because "the other guy just doesn't get it". It may be psychological, about marketing or internet hype, or a number of other factors, or it may be a real difference. People can also always dismiss a single sound (without a reference) by not wanting to admit they were wrong, by associating it with a certain piece of gear they do not like, et cetera.

The way I understood it, the challenge would be thus:
1) Someone makes a patch that people agree on sounding awesome, on a chosen FM synth. A wav file and a description of the patch (the parameters) is posted.
2) Tony replicates it with another FM synth (Sytrus?) and posts the replication in wav format, possibly the Sytrus patch too if people want to try it out.

So if the two patches don't sound the same, the difference can be in either the synth or the programmer's abilities. In former case, it has been proven that the synths sound different. In latter case, the burden of proof is still tony's - the sound should be replicated in a satisfactory way, and if it isn't, it can't be proven that the synths are capable of making the same timbre, which was the point argued by many people it seems. If the difference is supposed to be evident even in simple sounds, then the test should probably be started with recreating a simple sound.

The signature sound is evident regardless of the patch (within reason, you could program any synth to sound like garbage I suppose.

Yeah, but that's a claim that can only be backed up by replicating a patch. That's my point. You can say "it's evident" or "it's obvious" but if you want someone who doesn't agree to actually agree with you (or prove him wrong properly), those can be considered subjective arguments. If you don't want the other side to see your point of view, it's not much use arguing.

To clarify, I'm not debating "for" or "against" the perception that a certain FM synth (or other piece of gear) has a magical sound. I actually do have the same perception as you. I'm just saying that "well, it's obvious that it does and this one does not" is a subjective opinion (possibly shared by a number of people). Comparing the same patch on two synths leaves room for subjectivity only in whether both of the sounds sound different to someone or not. That is, whether a difference in two versions of the same patch can be perceived. It's much more fair than to just dismiss the other opinion with "well, it's obvious that..."

[TristezaOrange]

This thread has lost it.

Can I - as a person that has never used hardware synths in any *real* capacity (other than messing around with some of them occasionally) - ask if it really matters that much what the differences are in the context of a mix? Having heard the sound demos for the Transient Attacks sound-pack on NI's site, I would be glad to own an FM8. Really, what exactly is wrong with it?

Threads like this blow my mind.

[Aroused by JarJar]

THAT's sound that sounds like sound and not a picture of sound

No idea what you're talking about.
Is this philosophy now?

It's not philosophy, it's a basic issue of working with sound. No recording or mixing engineer, or musician, I've talked to over the years has had any trouble understanding or talking about this, and I highly doubt that you actually don't understand the idea, maybe my words are confusing.

When people say "but it sounds like it's made of plastic", they are not deluded, and they are talking exactly about this issue.

This isn't voodoo or philosophy, it can be explained by simple science. "Movement is life". Data compression in video makes things look "fake", as it usually simplifies and makes static areas of the picture that the viewer allegedly won't notice (yeah right, my five year old notices such things even more quickly than I do, and I'm attuned to it for technical reasons).

[tony tony chopper]

Btw, before testing, let's agree that a sine out of a DX7 is a sine like one out of FM7 (and not a colored one due to something else, otherwise we're not talking about the synthesis anymore, and this distortion can be achieved with post-effects in software too).
The DX7 should be producing slightly noisy sines (I assume) but not one with overtones, that is.

Otherwise, if a sine out of the DX7 isn't a sine, let's first question the recording method/tools.

[z15]

When people say "but it sounds like it's made of plastic", they are not deluded, and they are talking exactly about this issue.

That's actually my point, "when people say" means that lots of people agree on something, it doesn't mean that something is a fact. Lots of people think M-Audio is professional quality gear, and lots of people think Britney Spears is talented and rocks. Lots of people think all DCO synths must always sound "warm and analogue" and DSP synths can't.

There's just no other way to make sure that perceptions are right, than to put those perceptions into test and share the results so others can verify what has been tested. Unless you don't want to try to question a subjective opinion, in which case it isn't much use trying to debate about it or try to make people who don't have the some opinion to agree.

[z15]

Btw, before testing, let's agree that a sine out of a DX7 is a sine like one out of FM7 (and not a colored one due to something else, otherwise we're not talking about the synthesis anymore, and this distortion can be achieved with post-effects in software too).
The DX7 should be producing slightly noisy sines (I assume) but not one with overtones, that is.
Otherwise, if a sine out of the DX7 isn't a sine, let's first question the recording method/tools.

In case of a softsynth like VOPM, I guess the recording method can't be questioned, but in the case of a hardware synth you're right, it should be done straight to good converters and cleanly since otherwise you could as well run a software synth through the same process.

After eliminating the problem of recording... Now, to be honest, I haven't followed the discussion all the way through. Is the point of the debate about core synthesis algorithms or the actual perceived output of different FM synths? If it's the latter case, it would make more sense to me that the exact sound with all the possible imperfections should be replicated - by using additional software effects if needed.

As I understood, the whole point of the debate wasn't that something is not possible with DSP at all, but that something isn't possible with certain pieces of software / hardware (or more specifically, most pieces of software some posters have tried?). Thus it should also be demonstrated that it is actually possible to achieve the same end result using post-effects available to general public, and not barely claiming "you can always add all the imperfections using effects" without actually proving it.

[Aroused by JarJar]

You lost me there- if another synth has that intrinsic quality to the sound, then it does and that's it, it doesn't matter what the "picture" of the sound is like. A synth or programmer could completely fail to replicate a sound either because the features or the skill level aren't there, that's not relevent to the essential quality of the sound.

What I mean is that we need the same "picture" of sound to compare with, otherwise it's a subjective argument and we can go arguing about it for the rest of the year or just stop because "the other guy just doesn't get it". It may be psychological, about marketing or internet hype, or a number of other factors, or it may be a real difference. People can also always dismiss a single sound (without a reference) by not wanting to admit they were wrong, by associating it with a certain piece of gear they do not like, et cetera.

The way I understood it, the challenge would be thus:
1) Someone makes a patch that people agree on sounding awesome, on a chosen FM synth. A wav file and a description of the patch (the parameters) is posted.
2) Tony replicates it with another FM synth (Sytrus?) and posts the replication in wav format, possibly the Sytrus patch too if people want to try it out.

So if the two patches don't sound the same, the difference can be in either the synth or the programmer's abilities. In former case, it has been proven that the synths sound different. In latter case, the burden of proof is still tony's - the sound should be replicated in a satisfactory way, and if it isn't, it can't be proven that the synths are capable of making the same timbre, which was the point argued by many people it seems. If the difference is supposed to be evident even in simple sounds, then the test should probably be started with recreating a simple sound.

The signature sound is evident regardless of the patch (within reason, you could program any synth to sound like garbage I suppose.

Yeah, but that's a claim that can only be backed up by replicating a patch. That's my point. You can say "it's evident" or "it's obvious" but if you want someone who doesn't agree to actually agree with you (or prove him wrong properly), those can be considered subjective arguments. If you don't want the other side to see your point of view, it's not much use arguing.

To clarify, I'm not debating "for" or "against" the perception that a certain FM synth (or other piece of gear) has a magical sound. I actually do have the same perception as you. I'm just saying that "well, it's obvious that it does and this one does not" is a subjective opinion (possibly shared by a number of people). Comparing the same patch on two synths leaves room for subjectivity only in whether both of the sounds sound different to someone or not. That is, whether a difference in two versions of the same patch can be perceived. It's much more fair than to just dismiss the other opinion with "well, it's obvious that..."

Yes comparing very similar shapes would be a good idea. But they shouldn't be "Yamaha emulation" kinds of patches, but "classic" things that pretty much any FM synth could do, like "sines, 3:1 M:C with mod index from 3 to .1 exponentially over 6 seconds". Bonehead simple stuff. Not a patch programming competition, but test of the intrinsic sound quality of the synth.

By the way, as far as confirmation bias and all that: because this all this shit is relevant to my work, and though I make a living I'm far from rich, I for one am actually more than willing to accept, and maybe even biased towards, the idea of inexpensive (or free! ) software equalling or outdoing hardware, yeah! I am ecstatic when I hear "the sound" from software, not prejudiced against it's very existence.

[tony tony chopper]

Wow I'm trying VOPM, I don't get anything of those confusing settings. Can't manage to make it output just a sine. I also get stuck tones, or timbre randomization while I'm not seeing something that looks like randomness.
I guess someone will have to help. And also post a good preset bank because it doesn't come with any (well, 5 presets that suck).

Edit: ok the blank patches seem to output a normal (little noisy) sine.

[z15]

Yes comparing very similar shapes would be a good idea. But they shouldn't be "Yamaha emulation" kinds of patches, but "classic" things that pretty much any FM synth could do, like "sines, 3:1 M:C with mod index from 3 to .1 exponentially over 6 seconds". Bonehead simple stuff. Not a patch programming competition, but test of the intrinsic sound quality of the synth.

Yeah, I completely agree on that. The main point of debate doesn't seem to be "nobody has made an emulation of DX7 that is accurate right down to the last bit of noise" but "there is something in the sound of several different FM synths that manifests itself even in simple patches, and is missing from a lot of other FM synths". So I guess VOPM would be a nice starting point IF people do agree that it's something that has a great overall timbre.

I can't help with that part unfortunately, since I haven't really gotten deep enough into different FM synths that I could make any arguments about them. Used to have a DX7 years ago and used some software FM synth briefly, but that's about it.

By the way, as far as confirmation bias and all that: because this all this shit is relevant to my work, and though I make a living I'm far from rich, I for one am actually more than willing to accept, and maybe even biased towards, the idea of inexpensive (or free! ) software equalling or outdoing hardware, yeah! I am ecstatic when I hear "the sound" from software, not prejudiced against it's very existence.

Same here! I guess lots of people actually do have an open mind and do prefer inexpensive plugins / hardware that are subjectively as good / better sounding than more expensive ones.

Strangely often on bigger pro studio context or with people who buy the hype and don't evaluate things carefully themselves, you do see "gear snobs" that have some sort of bias for boutique priced stuff (whether software or hardware) since "it probably sucks if it doesn't cost more than that" - and of course there is the point about old or analogue gear having to be necessarily better because of "mojo" or "magic". That isn't to say that expensive, vintage or analogue gear couldn't be subjectively better (there are a lot of those things too), just that some people do have a bias for them whether the differences are audible in reality or not.

FWIW and unrelated to the debate, I'm really hyped about the D-Cam Synth Squad based on the features + demos so far, it's probably the first time after the last G-Force VA that I've been really anxious to get a certain piece of "analogue modeling" synth.