What a debate you've started . . .
And I've quite a lot to say on the subject as I've access to both and use both on a regular basis. I'll also apologise to those people who I bore to tears but hopefully, someone will find some of my ramblings useful...
Anyway, I'm going to agree with a lot of what was said but (and if you look at my list of synths) there are some things that need clarifying.
Synths generally fall into four categories:
Original Analogue: These use VCOs for wave generation (I unfortunately don't have any of these anymore)
Digital Analogue: These use DCOs for wave generation (I have a number of these)
ROM Players: These use Samples for wave generation (I have a number of these too)
Virtual Analogues: These use a computer program to simulate the sound of an analogue synth (I have a few of these also)
And to be honest, I can't really tell them apart as they all have their own character and I can't say that one sounds worse than the other as the same patch played across them, sounds different but not worse or better so it's a matter of taste really. The only one that I'm still not convinced of its worth yet is my EX-8000 and I always wanted one of them but the jury is still out on that synth ATM. It's also a slightly weird one as it uses a ROM oscillator to generate the waveforms in real-time which are then filtered via analogue circuity including a VCA so it has a warmish analogue sound but I can't put my finger on what it is I'm not thrilled about.
It's also important to realise that software synths can only fall into the last two categories; ROM Players and VAs!
I also agree that most modern computer hardware is more powerful than most synths that have been made and are being currently made.
However, there is one overriding fact in favour of a "real/hardware synth!. As it has been designed for a sole purpose, the processor need not be as powerful as it only needs to perform a set of given instructions and that is coded from the basecode upwards whereas a computer has to multi-task and perform loads of instructions which all compete for resources and therefore this requires a more powerful processor so that argument is irrelevant really as long as the devices processor can handle everything required of it, the latest and fastest CPU would be a waste of money anyway!
Lets look back in time. My Roland JV-2080 made in 1997 to take an example, only has a 32-bit processor running at only 20Mhz! But it's a RISC processor and not a CISC bassed processor like Intel and AMD chips. It means that it has a reduced instruction set so that it's more efficient than a CISC based processor and therefore doesn't need to run as fast or be as powerful as a CISC processor.
By nowadays standards, you could say that it's laughable but how many computers from 1997 do you know are still able to run and perform the same job? Probably very few yet my JV still performs as quick as it did the day I bought it and is still able to do its original job.
Even when you compare it to modern day ROM players, it's still very powerful and capable as it can output 64-voices across 16 MIDI channels and has 5 FX sub processors that can handle all of this simultaneously. It's fully editable as every parameter can be changed and you can even have stereo patches too! It also has a drum sound module built in and it stills sounds as good as anything out there now because Roland used very efficient algorithms to compress those sounds back then so it needed minimal memory to store those samples as a result! If you get bored of the inbuilt sounds, you can expand them with space for 8 expansion boards.
I think it's still a monster and stands up against most other things well enough to still use! It also has dedicated audio circuitry which was D/A converters that were 18-bit even back then and again that's more than enough as you always end up converting downward to 16-bit for CD anyway. As you can expand it using 8 expansion cards and a memory card, it can store close to 3000 wavwforms/samples!
These facts alone make it a monster but think inside your current computer, how much processing power you would need and space on your disk to store all of that.
My point is, as a dedicated ROM player synth, it's very powerful and efficient at what it does but anything you buy as a softsynth will consume loads of resources to match that head-to-head spec wise as it's likely you'd need to run multiple instances to achieve the same!
The other thing is note how a lot of computer-based producers spend a lot of money and time tracking down plug-ins which give their productions a more organic feel, less precise, warmer sounding, less digital, etc, etc.
Familiar words? Of course, you'll hear loads of producers mention that as guess what? They're all trying to emulate analogue synths and analogue recording techniques because their productions are done entirely within their DAW! How sterile and you'd be correct for saying that. This is especially true in the Progressive House and Trance genres.
So, why would established producers like Mike Koglin (I only mention him as you can find plenty of quotes he's made online) ditch all or most of their hardware (I know he used to use a Nord Lead 2, JP-8080, Juno 106, D-50, etc) for software? Because it's convenient, instant re-call, portability, he takes his laptop around with him so that he can work on productions all over the world. He loved his hardware but he did it and successfully too!
Personally, I think it's a matter of choice! I grew up salivating all-over Roland and Korg brochures (Yamaha never did it for me, especially the DX-7, sorry). I dreamed of owning a Roland Jupiter 8, TR-909, Korg Mono/Poly, DW/EX-8000, etc and when the JP-8000 arrived, I almost passed out!!! As a result, over the years I developed a bad case of G.A.S and as a result, the last decade I've spent re-building my home studio after moving and re-acquiring an arsenal of synth hardware again because for me, there nothing like twiddling knobs and sliders and seeing flashing lights all over my studio.
I also frequently fall off my chair laughing when I see demonstrations of people trying to program a VST synth by only using a mouse which results in difficult movement when in fact you could use a control surface but you'd have to program that of course which is why it led me to conclude, I still like hardware thank you!
On to my last points now, I recently tried to get the same white noise sound from a bunch of VSTs as my Supernova II was producing and could I? Well, yes but was it as lush, smooth, bodied, phat, etc? Nah, it wasn't and I could hear the difference.
I tried the following:
PSyn II, Pentagon I, Triangle II, Superwave P8 and the Pro, MiniMogueVA and a few other free VSTs - gave up in the end.
All I was trying to do was produce a white noise sweeping sound. Simple right? You'd think so but it wasn't.
Therefore, I recently purchased a Novation V-Station as this is meant to be the VST version of my A-Station. We'll see how true that is when I try the same patches on both and I'll let you know!
I've therefore concluded that if you want decent sounds, you probably need to pay for a synth that's been programmed up from the ground and not one based on the SynthEdit tools as I noticed all the free ones based on that toolkit, pretty much sounded the same. Also, don't get me wrong as there's nothing wrong with free synths based on SynthEdit as it's a wonderful tool but all I'm saying is that while I've got those synths to produce some wonderful sounds, they do have limits to what they can do.
I'm also aware that the bundled Cakewalk synths are originally from RGC Audio and not made from SynthEdit which is why I had high hopes for those!
There is one VST that really impresses me is Lennar Digital's Sylenth 1 as the demos I've heard are indeed extremely impressive.
That's it folks, that's my essay for that topic...
post edited by RickJP909 - 2013/04/29 16:22:33