OASYS vs. KRONOS: sound quality

shawnhar · Post by **shawnhar** » Sat Sep 03, 2011 11:38 pm

[quote="Megakazbek]Digital is completely deterministic, there is no possibility at all of "slightly different build" in particular OASYS or "a millionth of an amp being different in one of the circuits".[/quote]

Absolutely true. Changing the CPU type, voltage, etc, does not change the output of deterministic digital algorithms in any way.

But there is broad category of possible changes where the code running on this CPU may not be entirely the same on different devices, or even different minor OS revisions. Even if the algorithm is broadly the same, Korg may have optimized it to run more efficiently. Perhaps they upgraded some DSP code from using SSE2 to SSE3 SIMD instruction sets. Perhaps they didn't change their code at all, but recompiled it using a more recent version of the gcc compiler, which has a different codegen implementation and so produced different program binaries. There are any number of reasons why the specific instructions running on the CPU may not be 100% identical.

Thanks to how floating point math works in computers, even a tiny change in codegen may make tiny changes in the output values. Even though the algorithm remains basically the same, rounding behavior will be different. This is a good article if you are new to this area of computer science: "What Every Computer Scientist Should Know About Floating-Point Arithmetic" - http://download.oracle.com/docs/cd/E199 ... dberg.html

I personally feel these differences (if they exist at all) are so minor as to be irrelevant at least for my purposes. But there are many totally likely and innocent reasons why a tiny difference could occur.

Dany · Post by **Dany** » Sun Sep 04, 2011 1:01 am

maphill wrote:Okay, this is not conclusive, but here are some recordings of the same named patch (native, I did not load the OASYS patch on the KRONOS). The patch is played using the pads, and for this Combi it uses a fixed velocity default for the pads, so velocity is not a factor.

I do not know for sure that I am using the same cables. Probably not.

Velvety Strings: OASYS Combi Int-A 070 and KRONOS Combi Int-D 070

http://music.doctorobvious.net/VelvetyStrings6.mp3
http://music.doctorobvious.net/VelvetyStrings5.mp3

Variables to consider: Two recordings based on KARMA or any patch that has LFOs that may or may not retrigger will be slightly different. The resulting waveform will be different. What is important to consider is if the character and tonality is the same, and overlook small variations in performance (if there are any random expressive elements in KARMA) and phase of LFOs (where in a cycle a LFO is when a given note starts may matter).

My initial conclusion: These sound awfully similar to me. The character doesn't seem obviously different. Though I'm interested in others thoughts.

All it takes is a SINGLE case that shows the sound to be the same (as long as it that exercises a majority of the frequency spectrum) to indicate with high probability that the the sound production quality is basically the same. If you don't get this, then we can discuss it.

If this isn't a convincing example, then we can proceed to control more variables (listening to simpler sounds without LFOs through the same cables or digital recordings, using the same OASYS patch on both machines). But to do that, I'll need the cooperation of an OASYS owner...

Mark

Thanks for the comparison Mark! The first obvious difference I can hear is, that the VelvetStrings5 example seems to have a different reverb level or a larger or more audible reverb room, than the VelvetStrings6 example. It is especially audible between second 23 to 28. There are probably other reasons for this subjective difference in the perception of the audible room part of the sound, which doesn't say a lot, because of the suboptimal conditions of your test.

I'd like to help you, to achieve a reliable comparison. Please send me a PM with your email adress.

-

leo200 · Post by **leo200** » Sun Sep 04, 2011 6:22 am

Dany wrote:
Dany wrote:
Akos Janca wrote:...Interesting thread. Thanks for the examples, I hear the differences. Do you want to reveal the sources?
(Spoiler alert for all Hungarian readers...)

Ákos Üdvözöljük!

Leo elküldte nekem a megoldás e-mailben. Nehéz elhinni, de a tizenegy és tizenhárom az O.

Best wishes, from a Swiss half Hungarian...

-
After Leo's Post, where he reveals the solution of his comparison test, I've realized that I accidentally gave the wrong information in hungarian language to Akos. Sorry for the confusion.

BTW: To Korg: Why don't you release the new Kronos KRS-02 Str-1/Mod7 Phase II Sounds for the OASYS , as well?

-

SORRY! It's I accidentaly gave WRONG information- 12 and 14 is KRONOS!!

SanderXpander · Post by **SanderXpander** » Sun Sep 04, 2011 7:35 am

Haha, really?
That makes me happy, I thought those sounded a bit better

leo200 · Post by **leo200** » Sun Sep 04, 2011 3:44 pm

SanderXpander · Post by **SanderXpander** » Mon Sep 05, 2011 9:24 am

Sorry to be dense, but what does the wink mean? That you were kidding or that you're being happy with me?

leo200 · Post by **leo200** » Mon Sep 05, 2011 10:00 am

Surely happy with you!

Hedegaard · Post by **Hedegaard** » Mon Sep 05, 2011 3:30 pm

leo200 wrote:
Megakazbek wrote: I'd still like to hear any comments from Korg about this. They should know better what exactly they've changed, or what settings could produce the difference.
Very interesting point. I'm wondering why Korg still dont answer to this questions? There is any reason or we are simply discontinued product?

Dan from Korg did indeed answer these questions about audio differences between the Kronos and the Oasys.......NONE!

Its awfully simple really, the DAC's and circuits are the SAME, hence there is no difference.

I imagine people are just trying to fish for some kind of difference, which incidentally dosen't exist.

If the circuit is the same, therefore the output is the same. The "only" differences you'd likely to notice, is the small 0.0001% fluctuations in the output jacks, the impedance of resistance in the wires e.t.c.

Making 2x MP3 files (MP3 of all things!) and listening to both, will of course yield unreliable differences.

The only way to test is with pro spectrum analyzers that show GRAPHICALLY and computationally any differences.

But why do that?
Dan from Korg has already written about this.

Akos (welcome back by the way) will gladly find it and post the link to that conversation, because this is what Akos does best

leo200 · Post by **leo200** » Mon Sep 05, 2011 5:17 pm

Hedegaard,
did you compared files personaly?

leo200 · Post by **leo200** » Mon Sep 05, 2011 5:33 pm

Hedegaard wrote:
leo200 wrote:
Megakazbek wrote: I'd still like to hear any comments from Korg about this. They should know better what exactly they've changed, or what settings could produce the difference.
Very interesting point. I'm wondering why Korg still dont answer to this questions? There is any reason or we are simply discontinued product?
Dan from Korgdid indeed answer these questions about audio differences between the Kronos and the Oasys.......NONE!

Its awfully simple really, the DAC's and circuits are the SAME, hence there is no difference.

I imagine people are just trying to fish for some kind of difference, which incidentally dosen't exist.

If the circuit is the same, therefore the output is the same. The "only" differences you'd likely to notice, is the small 0.0001% fluctuations in the output jacks, the impedance of resistance in the wires e.t.c.

Making 2x MP3 files (MP3 of all things!) and listening to both, will of course yield unreliable differences.

The only way to test is with pro spectrum analyzers that show GRAPHICALLY and computationally any differences.

But why do that?
Dan from Korg has already written about this.

Akos (welcome back by the way) will gladly find it and post the link to that conversation, because this is what Akos does best

1. output jacks not involved in any way- it's internal DIGITAL recording
2. it is not mp3 files-it is wav 48(!)kHz files
3. Dan at Korg (with all respect) can say that difference not exist, but I (as many here) still believe to my ears.
4. Spectrum analyzer- I agree- will show exactly where the difference and in which amount. But why to do that? Qestion is very simple-does ANY difference exist. And answer simple too-SURE!

Hedegaard · Post by **Hedegaard** » Mon Sep 05, 2011 5:49 pm

Hello Leo,

While I can respect your enthusiasm for this project

I however don't see a need to continue to do research in this regard.

The electronics speak for themselves really.
No, I didn't audition anything from any links posted here, because its plain to see that the setup is not within any scientific calibration procedure, so therefore you can deduce automatically that the outcome will FAIL even before it starts.

Just an example, to "audition" with ears is purely coincidental, because your ears are not an accurate measuring tool. One day you will be in a good mood, or be tired, or stressed from work, and the tones can change. So what you once thought sounded good, sounds slightly different.
This is the wrong approach.

Another example:
Typically, people without an education, will try and try and try and try to build something and 99% fail, whilst people that have higher degrees, will already from the beginning know WHAT can be done/not done.

Also, the CPU, be it a Pentium 4 or a dual core, will have absolutely NO reference to the sounds, its not the CPU controlling anything "audiowise" in that regard, a CPU is just a processor, that must process things shown on the screen and handle inputs/outputs e.t.c.
The Audio DAC's (which is 100% the same in Kronos as the Oasys) is where the "real audio" is happening and this is what you hear as such.

Reading these posts, I am beginning to think that some future post will sound like...."and because the Kronos has black plastic and the Oasys has aluminium end pieces, therefore the Oasys will sound better, because aluminium is heavier than plastic"

To me, it sounds like Oasys users are absolutely desperate to fish for "something" that will make their Oasys better than a Kronos(?)

No offence

(I also have an Oasys - and am a little envious of the Kronos - but not so much that I want to try to distort facts)

McHale · Post by **McHale** » Mon Sep 05, 2011 5:50 pm

Finding the same patch names on both synths isn't good enough for a side by side comparison. They tweaked a lot of patches and the Kronos has more parameters. The best way to test is to take a PROG from the OASYS then import it into the KRONOS. And even that doesn't guarantee the same settings if the KRONOS has more parameters.

We do know that the direct to digital algorithms are new since they increased the specs there.

Personally, after years of owning analog gear and finding that nearly no two analog synths of the exact same type will sound exactly the same, I don't see all the fuss. To any listener, they will sound exactly the same.

markuk · Post by **markuk** » Mon Sep 05, 2011 5:59 pm

leo200 wrote:
Hedegaard wrote:
leo200 wrote: Very interesting point. I'm wondering why Korg still dont answer to this questions? There is any reason or we are simply discontinued product?
Dan from Korgdid indeed answer these questions about audio differences between the Kronos and the Oasys.......NONE!

Its awfully simple really, the DAC's and circuits are the SAME, hence there is no difference.

I imagine people are just trying to fish for some kind of difference, which incidentally dosen't exist.

If the circuit is the same, therefore the output is the same. The "only" differences you'd likely to notice, is the small 0.0001% fluctuations in the output jacks, the impedance of resistance in the wires e.t.c.

Making 2x MP3 files (MP3 of all things!) and listening to both, will of course yield unreliable differences.

The only way to test is with pro spectrum analyzers that show GRAPHICALLY and computationally any differences.

But why do that?
Dan from Korg has already written about this.

Akos (welcome back by the way) will gladly find it and post the link to that conversation, because this is what Akos does best
1. output jacks not involved in any way- it's internal DIGITAL recording
2. it is not mp3 files-it is wav 48(!)kHz files
3. Dan at Korg (with all respect) can say that difference not exist, but I (as many here) still believe to my ears.
4. Spectrum analyzer- I agree- will show exactly where the difference and in which amount. But why to do that? Qestion is very simple-does ANY difference exist. And answer simple too-SURE!

Dan doesnt really give a toss what oasys owners think because korg dont sell the oasys.
I will be totally honest, i dont really mind if they dont bring anything else out for the oasys but i do think its discusting that the have used oasys owners a guinea pigs then fobbed them off.
Korg are only a name, they dont really care what you think unless it affects their business.
They sent letters to a selected few people on here to keep everyone happy. I didnt get one by the way. although im not really arsed about a piece of paper.

leo200 · Post by **leo200** » Mon Sep 05, 2011 6:02 pm

Hedegaard wrote:Hello Leo,

No, I didn't audition anything from any links posted here

I advise you compare it by yourself
http://www.sendspace.com/filegroup/JSUE ... v4eoYmIVJg

http://www.sendspace.com/file/rb9sp4

to be fair I've being ready that K will sound even better. I'm not looking for reason why O is better.

Hedegaard · Post by **Hedegaard** » Mon Sep 05, 2011 6:11 pm

Leo, ok I downloaded both clips.
And?
Sorry, I don't hear any comparative differences.
(I was just playing it on my PC speakers - which are quite good Pioneer system - not the cheap WallMart edition plastic junk type things)

But what am I trying to listen for?
Theres so much fill in's going on, its hard to hear exactly what I'm supposed to be listening to.

Ok, good to know you're not "fishing" for things to make the Oasys seem better