Stereo-oscillators?

[siebenhirter]

Hello,

can anyone explain what is meant by the Pa5x's 24 stereo oscillators?

In my opinion, an oscillator can only produce one oscillation at a time (mono), so there is no such thing as a stereo oscillator in the technical sense.

A popular tool in sound design is to slightly detune several oscillators from each other. This gives an engaging, wider sound than a single oscillator - but that's not subject to stereo oscillators.

When one speaks of stereo multisamples - as with KROME - one means the "Double" oscillator mode, in which twice as many voices are called up in contrast to the "Single" oscillator mode (in "Single" mode Only one oscillator is used, in "Double" mode two).

This is nothing new in and of itself, but has been around with Pa-arranger keyboards for a long time.

In order to achieve a stereo effect, one oscillator is assigned the right sample and the other oscillator is assigned the left sample of a stereo recording. Although it wasn't called "stereo multisample", for stereo it also required an oscillator for the output of the left and another for the right sample.

What is different about the Pa5x?

[showwhite]

Bump!

I am also curious about the stereo OSC. Does it support stereo pcm sample now? Or we still need to export the left and right channels separately and pan them left and right in order to achieve the stereo signal?

[siebenhirter]

I am also curious about the stereo OSC...

Following answer I got from Jürgen Sartorius - Assistant Brand Manager of Korg&More, Germany (translated German to English):

Manufacturer's website states that you no longer have to divide the oscillators or your own samples into left and right,
in order to later reassemble them into a stereo image in Sound Edit.
For example, if a stereo sample is imported, it remains a stereo sample and is not split into L and R.
Here are screenshots for clarification.
Greetings Juergen Sartorius

[musiccankill]

I am also curious about the stereo OSC...

Following answer I got from Jürgen Sartorius - Assistant Brand Manager of Korg&More, Germany (translated German to English):

Manufacturer's website states that you no longer have to divide the oscillators or your own samples into left and right,
in order to later reassemble them into a stereo image in Sound Edit.
For example, if a stereo sample is imported, it remains a stereo sample and is not split into L and R.
Here are screenshots for clarification.
Greetings Juergen Sartorius

Exactly!
And ,even though i haven't compared a split one to the same but stereo imported, i think i hear a better response on the stereo imaging now that the oscillators are stereo compared to the mono split and panned one.
Also it is way easier to just configure 1 oscillator instead of doing the work 2 times..

[Asena]

Thank you all for this information,
I think the idea behind is similar to the Kronos.
In Kronos we can import for example a sf2 file and all Oscar s loading correctly with the necessary information of the layering.
I hope they can ad this import function to the Pa5x since everything out there can do this.

I love to see this keyboard to come
Alive very soon, for their business and reputation it’s important that they do it.

[siebenhirter]

... now that the oscillators are stereo ... it is way easier to just configure 1 oscillator instead of doing the work 2 times..

I think Pa5x oscillators are not stereo, but editors software will handle that link within a stereosample. As far as I know, the hardware of the sound engine has not changed compared to its predecessor, so the stereo sound comes from two different oscillators, just like before.

Stereo sound contains two channels and these are separated as individual samples and linked only to the editor software as one stereo sample.

This has the advantage that various parameters are automatically assigned to the two samples and no longer have to be edited separately.

.. is similar to the Kronos. In Kronos we can import for example a sf2 file ....

Same with significantly improved import functionality of Kronos - in case of stereo sound two oscillators are required. With Kronos up to 16 stereo velocity zones are supported within a single HD-1 Program, by using stereo Multisamples and also with both OSC1 and OSC2 (Image Kronos-Osc1-2.jpg).

Stereo Multisamples use twice as many voices as Mono Multisamples - should be the same for the Pa5x.

If you want to edit each sample of a stereosample separately, just simply unlink them so they will behave like separate monophonic samples.

[musiccankill]

I think Pa5x oscillators are not stereo, but editors software will handle that link within a stereosample. As far as I know, the hardware of the sound engine has not changed compared to its predecessor, so the stereo sound comes from two different oscillators, just like before.

The hardware HAS changed (the newest that has ever been used was AM5718 and now we have AM5728 processor).
Although that has absolutely nothing to do about mono/stereo oscillators as all these are software side things.They could implement a stereo oscillator even on pa3x and pa2x if they wanted to.

Stereo sound contains two channels and these are separated as individual samples and linked only to the editor software as one stereo sample.

Stereo Multisamples use twice as many voices as Mono Multisamples - should be the same for the Pa5x.

Again not really (although they could be written as 2 mono samples in the KMP file.I will have to investigate how they are written as internally).The stereo performance is handled in DSP mode as first tests of some of my friends shown that the polyphony is the same if you use mono or stereo samples (a stereo sample doesn't need 2 voices as with the older split Multisample way).
I haven't tested that personally so don't take it for granted!

If you want to edit each sample of a stereosample separately, just simply unlink them so they will behave like separate monophonic samples.

Are you sure there is a way to unlink a stereo sample? (need to check that).

[Scott]

Yamaha made a similar shift. Motif XF had 128 polyphony. A mono sample would take 1 unit of polyphony; stereo would take 2 units of polyphony since it was actually playing two samples. That changed with the Montage... its sample engine has 128 polyphony, but it is 128 regardless of mono vs. stereo. Stereo is somehow implemented as a "stereo waveform" rather than as a pair of mono waveforms.

Roland Fantom kind of does the same thing, though they describe it differently. They have max 256 polyphony, but they say that, if you're using their PCM (sampled) sounds, each sample you play takes two units of polyphony for stereo... and it takes those two units even if the sample is mono! Which means it's effectively the same as what Yamaha calls 128. I'm guessing that Roland describes it that way because there is some other (non-PCM based) sound engine in there that could actually produce the 256, but I don't know.

[siebenhirter]

.. hardware HAS changed ...

I already know that a different Sitara SoC is used in the Pa5x and with same Pa4x/1000 hw-components CortexA15 subsystem and C66x DSP etc. Regarding stereo/mono, I mainly intended to state that the sound engine has not changed, but that the linking of the parameters from mono samples to a stereo sample mainly takes place in a more comfortable editor software.

Whether someone wants to understand a sound engine as the method originally named as tone generation as software or hardware of an SoC is up to them.

For me, the thread "How to overcome Pa4X/1000 SET loading issues in Pa5X" is an indication of an unchanged method of sound generation from samples.
http://www.korgforums.com/forum/phpBB2/ ... 244#815244
If you copy PCM and multisamples from the Pa4x to the KST file by copying a PC system and this can then be used properly in the Pa5x, then this is not proof, but an indication of the consistent sound engine, which uses the same methode and same PCM -Samples like the Pa4x/1000 works.

For the time being, only the information from Korg & More is reasonably reliable in this matter:
that with a stereo sample import you no longer have to divide the samples in the editor into the oscillators and then combine them into a stereo image afterwards.

[siebenhirter]

.. if you're using their PCM (sampled) sounds, each sample you play takes two units of polyphony for stereo... and it takes those two units even if the sample is mono...

I think a stereo signal can be generated from a sample in container format with a header, but stereo data is missing from raw PCM data without a header, without info about channels etc.

PCM raw data, i.e. a time- and value-discrete representation of the chronological progression of a signal, depends on two values, the sampling rate (number of samples per unit time) and the resolution (bit depth).

But it is not known what kind of sample format and what method Pa5x uses for its stereo sounds.

[musiccankill]

.. hardware HAS changed ...

I already know that a different Sitara SoC is used in the Pa5x and with same Pa4x/1000 hw-components CortexA15 subsystem and C66x DSP etc. Regarding stereo/mono, I mainly intended to state that the sound engine has not changed, but that the linking of the parameters from mono samples to a stereo sample mainly takes place in a more comfortable editor software.

Whether someone wants to understand a sound engine as the method originally named as tone generation as software or hardware of an SoC is up to them.

For me, the thread "How to overcome Pa4X/1000 SET loading issues in Pa5X" is an indication of an unchanged method of sound generation from samples.
http://www.korgforums.com/forum/phpBB2/ ... 244#815244
If you copy PCM and multisamples from the Pa4x to the KST file by copying a PC system and this can then be used properly in the Pa5x, then this is not proof, but an indication of the consistent sound engine, which uses the same methode and same PCM -Samples like the Pa4x/1000 works.

For the time being, only the information from Korg & More is reasonably reliable in this matter:
that with a stereo sample import you no longer have to divide the samples in the editor into the oscillators and then combine them into a stereo image afterwards.

As somebody that can look inside and have analyzed all the files till pa4x and some basic things in 5x pcm/kmp structure , i can safely say that although the OLD pcm form does LOAD directly when you put it inside a kst, when you resave to a KST from the internal memory, the PCM and KMP format get updated to the new format which has many changes compared to the old format.

I will do some tests on a clean (restored) keyboard to inform everybody here about how the stereo samples are being handled once i feel a bit better (i got covid)

[siebenhirter]

.. As somebody that can look inside and have analyzed all the files till pa4x and some basic things in 5x pcm/kmp structure ....

Big thanks for your information and your efforts - and most of all I wish you a speedy recovery and full recovery post Covid - all the best

[showwhite]

Much appreciated for the information. Can't wait for the detailed sample import tutorial. So sad I did not find a word mentioning that in the performance guide from KORG.

[davekalbach]

Can't wait for the detailed sample import tutorial. So sad I did not find a word mentioning that in the performance guide from KORG.

-- OFF Topic --

Deleted - Moved to a new thread

[Asena]

Music can kill thank you.
Thanx to you we are getting a better understanding about this.

Take care and be recovered quick.