Why PSR-S serie don't use AMM2 technology ?

[pquenin]

The PSR-S970 or S975 are quite expensive but they still use the old AWM technology.
I have play with a "cheap" MX 49 that is AWM2 and the sounds were "bigger"...

[panos]

Hi pquenin,
As long as the Yamaha MX49 is a synthesizer I guess the synthi sounds should sound better there.
It needs a pc and a daw to produce sounds, right?
And probably it is using another technology than the s970/s975 were the sounds must be more "realistic" to acoustic organs.

https://www.youtube.com/watch?v=Juy7ZeAfIhY
https://www.thomann.de/gb/yamaha_mx49_v2_black.htm


I could find this article about what WMA technology means
https://en.wikipedia.org/wiki/Linear_Arithmetic_synthesis

Very nice synthi sounds by the way for such a low price.

[SeaGtGruff]

I don't know what the differences between AWM and AWM2 are, but Yamaha's arrangers use AWM, while their synths use AWM2:

PSR-S3000: AWM Stereo Sampling
PSR-S975: AWM Stereo Sampling
Tyros5: AWM Stereo Sampling
Genos: AWM Stereo Sampling, AEM technology

MM8: AWM2
MOTIF XF8: AWM2, with Expanded Articulation
MOXF8: AWM2, with Expanded Articulation
MX88: AWM2
MONTAGE8: Motion Control Synthesis Engine, AMW2: 8 Elements, FM-X: 8 Operators, 88 Algorithms

The following page says a number of things-- some incorrect-- from various people who don't really know:

http://www.harmonycentral.com/forum/forum/Keyboards/acapella-18/341405-

But one comment that might be correct is that the difference has to do with how the soundwave samples are compressed:

AWM uses a constant algorithm to encode the wave information. AWM2, on the other hand, allows the programmer (sound designer) to roughly optimize the compression for a given sample. For example, a triangle lead may be able to suffer much more compression (due to the fact that it has very little bandwidth and harmonics) when compared with a steel-string guitar, or harpsichord, both of which contain a great deal of non-static harmonics.

The following page on Yamaha's site doesn't really explain the difference:

https://usa.yamaha.com/products/contents/music_production/synth_40th/history/chapter03/index.html

But it seems to say that AWM2 uses improved digital filters:

At that time, Yamaha had actually developed a digital filter capable of reproducing the behavior of an analog one, a feature that made its long awaited debut in our SY77 digital synthesizer in 1989. The SY77 was equipped with both an AWM tone generator and an FM tone generator, both of which could be used together with the digital filter to sculpt sound for remarkable levels of expression. These two new approaches to tone generation were christened Advanced Wave Memory 2 (AWM2) synthesis and Advanced Frequency Modulation (AFM) synthesis, respectively.

Anyway, it seems that AWM must be sufficient for "ROMpler" keyboards which are intended to play back sampled instrument sounds without much need or desire to create new sounds the way a synthesizer can. Yes, you can tweak the voices by making adjustments to the attack time, release time, filter cutoff frequency, filter resonance amount, etc. But apparently AWM isn't sufficient for digital synthesizers, which is why Yamaha's synths use AWM2.

[Joe H]

My MU100R and MU128 XG sound modules use AWM2 and they are not synths.  I suspect AWM Stereo may very well be AWM2 by another name. Or it is possible AWM2 uses a different kind of compression than AWM allowing larger samples or a different sample rate. 

The Motif features AWM2 Voices and I honestly don't hear any real significant difference in the sound quality between the S970 and the Motif. When I play them together and they blend very well with one another. But the S970 sounds better than the MU128.  I suspect it is the improvement in DSPs or how they are adjusted that make the S970 sound better in that case. I believe the S970 inherited some the Motif DSPs so maybe that's why the Motif and S970 sound better than the 1990s MU sound modules.

Having said all that, I must admit the when I put the MU128 into Performance Mode the DSPs make it shine because we can layer up to 4 Voices and use up to 5 DSP processors at once plus Channel EQ and Master EQ.

Joe H

[Joe H]

The MX49/61 are the last in the line of recycled Motif technology without the great display of the Motif XS or XF. What you do get is the sounds of the original Motif with only 999 arpeggios instead of thousands, only 2 arpeggiators instead of 4, and fewer User Voice Banks. It is the most scaled-back Motif.  The MX is the last attempt by Yamaha to milk the Motif cow for all it's worth before releasing the Montage.

Yamaha has recycled much of their technology such as vintage synths like the DX7 (FM), AN1x (Analog) and Grand Piano (from their first digital piano) in the form of plug-in cards, beat boxes (for FM and AN) then again as mini keyboards they call Reface.

We can get the Motif sound out of our arrangers with the right synth Voices. And Yamaha is adding those Voices with every new arranger keyboard it releases.

Joe H

[SeaGtGruff]

I don't think AWM2 is the same as AWM with stereo sampling, because Yamaha specifically says "AWM2" when describing the tone technology used in most of their current synths, but specifically says "AWM Stereo Sampling" when describing the tone technology used in their arrangers.

Also, they specifically say "AWM (Organ Flutes)" for the reface YC, but "SCM + AWM2" for the reface CP. So it seems clear that Yamaha considers AWM and AWM2 to be different in some way.

I don't think the difference is just the use of digital filters, because AWM uses digital filters, too, doesn't it? But according to Yamaha's article about the history of their synths, the digital filters introduced with AWM2 were an improvement over the earlier digital filters.

And I wonder if the claim about a difference in how the samples are compressed-- fixed algorithm versus varying algorithm-- might have some bearing on this as well, because if a variable compression method can allow better reproduction of complex harmonics in the samples then perhaps this improvement could work in conjunction with the improved digital filters to achieve a more analog-like manipulation of the sound?

[pjd]

Hi --

Not too much to add to what Michael and Joe have said. Thanks, guys!

From the hardware perspective, the original "AWM" hasn't been deployed in a Yamaha product for a zillion years. (Please see the Yamaha history in Michael's post). I remember reading (Yamaha synth forum?) where marketing has been using "AWM" and "AWM2" interchangeably in recent years, both meaning AWM2.

The sound generation chip (SWP51) in the first gen MX is the same as the Motif XS and XF, MOX and MOXF, Tyros 4 and 5, S750 and S950. The newer Yamaha products use the newest chip (SWP70). That would be Montage, Genos, S770 and S970, S775 and S975.

I have not seen a service manual for the second generation MX, but I suspect that it uses the SWP70, too. A hardware refresh was required because the SWP51L is out of production and the SWP70 is the mainline tone generator chip.

Much of the AWM2 technology is incorporated at the chip level -- the sample decompression is done on the fly from dedicated waveform memory. Filtering and other magic is done on the fly, too, in order to reduce latency.

I play and have compared MOX6 and S950. If there are sound differences, it's due to voice and DSP programming. Architecturally, the Motif XS, XF, MOX, MOXF  allow two insert DSPs in series/parallel while arrangers are limited to one insert DSP per voice. That is one of the few fixed factors that are different between the software+hardware engines.

Hope this helps -- pj

P.S. As Joe mentioned, the MX is the value product in the three tier Montage, MOXF, MX product stack. BTW, the MOXF is also due for a refresh.

[pquenin]

Thank you guys for all these informations.
I have played again with the MX 49 today in a store (but don't know if it was a MX I or MX II) because I was about to buy a used blue MX 49 II  (250 € here in my town).
But it was not as magic as the first time. I think that the first time I played it I have been surprised by the sounds of this little and cheap synthi.
I have tried to replicate a sound on my PSR-S670 by layering 2 voices and adding harmony and it sounded good too.
So I will not buy the MX 49 and it's wise because I don't have too much money and I just bought a used but like new PSR-S670 (250 € too).

[Lisa]

I had the MX49 for a short time here but it sounds really thin when I take the MOXF for example.
The Piano's on it are a bit better then my S670.
And the internal sound device is just 16 bit resolution so it does not sound very professional in my ears.
The sound improved when I recorded it thru the Line Outputs and my Focusrite Audio device..
Another thing but that's with more Yamaha synths the problem is that is has trouble to keep up with the MIDI clock of Cubase..
At some point while recording it's starts running behind the BPM.
That error occurred more then one time while testing it.

Offcourse the amount of programming on the device it self is a bit on the small side.
Only 2 voices and 4 knobs with 3 functions..
You can have more then 2 layers but you have to be handy in Cubase itself.
But programming the more added layers are a bit of a hassle.

So very good you bought a S670 keyboard in my opinion.

Lisa,

[SeaGtGruff]

Another thing but that's with more Yamaha synths the problem is that is has trouble to keep up with the MIDI clock of Cubase..
At some point while recording it's starts running behind the BPM.
That error occurred more then one time while testing it.

Someone else once mentioned having similar problems with their Yamaha keyboard-- Tyros or PSR-S, I forget-- and the recommended solution is to either sync the Yamaha's clock to the other hardware/software (External Clock on), or sync the other hardware/software to the Yamaha's clock.

However, I was investigating the Tempo and other settings on my PSR-E models-- in particular, the MIDI events and data values which get generated when specific panel controls or settings are used-- and I discovered that the keyboard doesn't use the full precision for the Tempo's data value, nor does it round to the nearest value for the degree of precision that it does use. I'll have to look back into this so I can give some specific examples, but since changing the Tempo doesn't generate a "live" MIDI event, what I was doing was setting the Tempo to a specific value, recording a User Song, saving the User Song to a standard MIDI file, looking at the MIDI file to see what value was used for the Tempo, computing the value that should have been used (i.e., was actually closest to the requested Tempo), and comparing those two values.