New Gaia... Weird Oscillators!

Forum for SH-01 GAIA
Post Reply
andyr1960
Posts: 55
Joined: 09:46, 5 March 2009

New Gaia... Weird Oscillators!

Post by andyr1960 »

Hi All, I've just bought a Gaia to replace my ailing JP8000. While I love the Gaia for it's immense sounds, I've noticed a quirk with the oscillators in that they are not "free running" like the JP8000.

For instance set, up a basic two osc patch and slightly detune one of them... play some notes, and you'll find that the oscillators start phase is the same for each note, I know it's difficult to explain, but you'll hear a distinct phasing sound that is NOT the typical two VCO "beating" that the JP8000 produces.

I've also noticed that the basic waveforms are brighter at certain points on the keyboard, which makes me believe that samples are being used, which would also account for the above.

Apart from that... it's a great little board.
User avatar
V-CeeOh
Posts: 3956
Joined: 18:13, 28 September 2004
Location: Portugal

Re: New Gaia... Weird Oscillators!

Post by V-CeeOh »

Sorry if this a disapointment to you but, yes, you just discovered that the Gaia is sample based and has not a "true" VA engine.
andyr1960
Posts: 55
Joined: 09:46, 5 March 2009

Re: New Gaia... Weird Oscillators!

Post by andyr1960 »

Not a disappointment as such, but I'd be worried about the Jupiter 80, as it seems to have a very similar VA engine from what I've seen... Ah, hold that thought, I've just noticed a similar thread on the JP80 forum.

Maybe the 80 has a random sample start function to make the oscillators more "analogue".
User avatar
V-CeeOh
Posts: 3956
Joined: 18:13, 28 September 2004
Location: Portugal

Re: New Gaia... Weird Oscillators!

Post by V-CeeOh »

Well, I'm not worried at all. And yes this lead to several discussions on the Jupiter forum. But the Jupiter has NOT a Gaia similar VA engine for the simple fact that the Gaia has no VA engine. The structure may seem similar but the engine is not.
We're all expecting to see it, however ;-)

BTW
Yes, the Gaia is a little beast :-)
User avatar
Artemiy
Site Admin
Posts: 19754
Joined: 13:00, 17 April 2003
Location: Ukraine
Contact:

Re: New Gaia... Weird Oscillators!

Post by Artemiy »

Both the SH-01 and Jupiter-80 have sample-based waveforms. You can tell this from pitching a waveform down and you will hear it loose brightness since all harmonics are shifted unlike with realtime-gerenated VA waveforms. In fact the Juputer-80 sounds the same as SH-01 to my ears, I bet it's the same engine albeit with more improvements. For example the Jupiter has free-running waveforms which removes the initial phase lock issue on key press.
Synthtron
Posts: 690
Joined: 16:24, 26 November 2011
Location: USA-East Coast
Contact:

Re: New Gaia... Weird Oscillators!

Post by Synthtron »

Do you all think it is misleading for Roland to label these as virtual analog if they use sample playback and there is no modeling taking place?
User avatar
Artemiy
Site Admin
Posts: 19754
Joined: 13:00, 17 April 2003
Location: Ukraine
Contact:

Re: New Gaia... Weird Oscillators!

Post by Artemiy »

Synthtron, the thing is there is no actual definition of "virtual analog". Whether you use it for math-modeled oscillators only, as a purist would do, or use it as a marketing gimmick, it's up to you. I can't blame Roland for doing the latter, although I'd prefer them not to because it seems that JP-8000, SH-201 and SH-01 are all virtual analogs, whereas in reality the SH-01 is different, so putting them under one roof is not right technically. But once again, "virtual analog" is just a label. Like an "R" or "S" badge on a car, it doesn't necessarily mean it's any faster.
Synthtron
Posts: 690
Joined: 16:24, 26 November 2011
Location: USA-East Coast
Contact:

Re: New Gaia... Weird Oscillators!

Post by Synthtron »

I would prefer them not to use the term as well, unless the sound engine was modeling technology, I guess the end user has to be discerning.
Grauw
Posts: 46
Joined: 04:29, 8 December 2010

Re: New Gaia... Weird Oscillators!

Post by Grauw »

I don’t really get why you would say the SH-01 does not fit under the VA nomer, or that Roland is supposedly using it as a marketing term only and thus not being truthful to its customers.

A VA is more than just an oscillator, what matters is that the sounds are composed in a way mimicking an analog synthesiser. Wikipedia:
An Analog Modeling Synthesizer is a synthesizer that generates the sounds of traditional analog synthesizers using DSP components and software algorithms to simulate the behaviour of the original electric and electronic circuitry, in order to obtain the sound in a more precise manner from the simulated inner working of the circuitry, instead of attempting to recreate the sound directly.
Who says what algorithms a VA must implement to achieve this? In the end a wave table is just another algorithm with a lot of data points. In a way you could say it may produce a more accurate analog oscillator representation than other algorithms with less data points (as an algorithm increases in complexity its data points will approximate those of a wave table).

If some manufacturer implements a cutoff filter with a different algorithm than usual, does that make it ‘not a true VA’ either? Who even cares how exactly a cutoff filter is implemented?

What matters is that the oscillator produces one of a common set of elementary waves, which then get processed and combined into a sound, as opposed to just recording the final sound. As long as the sound is composed the same way as analog synthesisers do, it is virtual analog.

Besides, all talk about the underlying mechanism is rumour, based on some obscure observations. No definitive confirmation. Sure it is an interesting theoretical discussion topic, but in the end I don’t see why it matters, or why this would mean Roland shouldn’t be calling the Gaia a VA and would be misleading its customers.
Re-Member
Posts: 114
Joined: 02:11, 4 June 2007
Location: Portland, OR

Re: New Gaia... Weird Oscillators!

Post by Re-Member »

Grauw wrote:Besides, all talk about the underlying mechanism is rumour, based on some obscure observations. No definitive confirmation. Sure it is an interesting theoretical discussion topic, but in the end I don’t see why it matters, or why this would mean Roland shouldn’t be calling the Gaia a VA and would be misleading its customers.
It's no rumor or obscure observation, the waveforms are multi-samples. Proof of this is how the GAIA allows you to select "variations" of the waveforms. The high polyphony count is also another indication of this. I was considering buying it when it was first released, but found a few things in the manual that stuck out like a sore thumb:

* If you’ve selected SYNC (oscillator sync), the MONO
setting will be forcibly chosen.

* PW and PWM cannot be used for tones 1 and 2 if you’ve
turned on SYNC (Oscillator Sync) or RING (Ring Modulator).
Also, if the OSC section’s asymmetric rectangular wave
is selected, variations for the OSC become invalid, and
the resulting tonal quality will be somewhat different
compared to what it would be with the original wave
selection.


Those two things matter. There's also no way to modulate the width of the Supersaw wave in real time, you're only allowed to choose between sampled variations of it. Had I just assumed everything functioned as a regular VA should, I would have overlooked this and ended up having to return it. Also,the lack of being able to create split patches turned me off as well. It seems to function a lot like the SH-32 which was classified as "Wave Acceleration Synthesis," not virtual analog.
Post Reply