Why Not Synths with Unlimited Polyphony?

[jah3]

I've got a question for you all.

Electronic musical instruments these days are judged among other things by how much polyphony they have. For example, 32, 64, or even 128 note polyphony, many of the voices being cut down when layering many sounds together.

The question is this:

1.Why not just make them with full or unlimited polyphony, as in an organ, or acoustic piano. Would it be too cost prohibitive to justify developing them this way, or would it not be possible to do it very easily?

If they had full polyphony, then you wouldn't have to worry about running out of voices when layering, or sequencing, etc.

Just wondering.

[Daniel Buck]

Is this possible?!?!

I've never heard about it but, if it does exist, I want one for me, please! [o=

hahaha

[jah3]

No, as far as I know, they don't exist. I'm just wondering why none of these electronic instrument manufacturers like Korg, Roland, Yamaha, just to name a few, haven't thought about doing this.

I mean it's good to keep improving on how many notes or voices one can play at a time on a synth, for example. It just seems that if they could develop synths where you wouldn't be limited to so many notes at a time, and could have full polyphony like on an acoustic piano for example, it would make more sense.

It seemed like an interesting thought. Maybe it wouldn't be practical, I don't know.

[Jeebus]

It's simply not practical. Polyphony requires hardware and software resources. Like any other feature, it's all part of a balancing act, with price being a key factor these days.

[jah3]

I guess that's true.

No telling what the prices of synths would go up to if it were possible. Anyway, I guess we're still doing pretty well with today's standards as far as polyphony goes. I think the most currently is around 128 notes, if I'm not mistaken. That ought to be more than enough, I guess. Although the synth I have that has the most polyphony currently is the N264, which has only 64 notes. That's plenty, I guess.

[keynote]

One of the 'ploys' used by Keyboard Manufacturers is to squeeze every ounce (dollar) out of the current technology. A few years ago it was 16 notes of Polyphony then you had/have 32 notes of Polyphony, then 48, 64, 128 and so forth. One of the crafty ways Keyboard Manufacturers induce sales is to slowly and incrementally increase the amount of Polyphony on their Boards. For instance let's say the current technology sports 16 notes of Polyphony like the Flagship Keyboard's of just a few years ago had. When suddenly Boards start hitting the shelves having 32 notes of Polyphony! People get excited about the prospect of having an Axe with 32 notes of Polyphony and many and dare I say most Keyboardists will upgrade to the newer technology Keyboards that have the increased Polyphony. So you have this great wave of people trading up or selling their old 16 note polyphony Boards and getting the 32 note polyphony Boards.

 

It's all about marketing my friend. Plain and simple. And Money. Don't forget about money.

 

They (the Keyboard Manufacturers) will then squeeze every ounce out of 32 note Polyphony (and by ounce I mean - Money) and over time they'd introduce Boards with 48 note polyphony, etc. Then squeeze every dollar out of that and slowly go to the next incremental increase, i.e., 64, etc.

 

If they gave it to us all at once they would lose all of that profit from those 'induced' sales from incremental updates. And Polyphony is just one scenario. Many other features on Keyboards are incrementally updated also. And in doing so the Keyboard Manufacturers assure themselves of maximized profits in the interim, i.e., (each phase of technology). All from the likes of you and me and most other dedicated Musicians who want and need the latest and greatest.

[Jeebus]

Keynote is right on all counts. Just like the computer industry, or any other technology based industry, we continue to buy as the manufacturers slowly improve their product. They have a carrot on a string and they're dangling it in front of us. Every few years, the carrot gets a little bit bigger, in order to hold our interest.

[fisheye]

With sample-based instruments, unlimited polyphony is impossible in a finite sized box. At least, for piano sounds. Organs are limited in their polyphony by their number of keys, piano's are limited.

 

But I guess some large number, like 128, would do, as long as you do not play full piano glissando's fast after each other (still then, no-one would notice).

[felix.]

A piano doesn't have unlimited polyphony - it has 88-note polyphony. An organ only has as much polyphony as it has keys (and pedals too, I suppose).

 

Some time soon, standard polyphony on most synths and keyboards will be enough that it is no longer an issue. 128 note polyphony is already a bunch, and this is pretty much the current measuring stick.

 

Keyboardists are already very, very spoiled. Most other instruments are 1-6 note polyphony.

[Dave Bryce]

Originally posted by keynote:

It's all about marketing my friend. Plain and simple. And Money. Don't forget about money.

 

They (the Keyboard Manufacturers) will then squeeze every ounce out of 32 note Polyphony (and by ounce I mean - Money) and over time they'd introduce Boards with 48 note polyphony, etc. Then squeeze every dollar out of that and slowly go to the next incremental increase, i.e., 64, etc.

 

If they gave it to us all at once they would lose all of that profit from those 'induced' sales from incremental updates. And Polyphony is just one scenario. Many other features on Keyboards are incrementally updated also. And in doing so the Keyboard Manufacturers assure themselves of maximized profits in the interim, i.e., (each phase of technology). All from the likes of you and me and most other dedicated Musicians who want and need the latest and greatest.

Actually, it's simpler than that - it's much more about processing power.

 

The reason the polyphony of synths increases over the years is the same reason computers continue to get faster and more powerful while not becoming more expensive - the technology is improving and therefore becoming more affordable.

 

I believe that has more to do with the polyphony issue than any of the factors stated above.

 

One more thing - do we really need more than 128 voices? Guitar players get by with six. Most of the instruments of the orchestra use one, as do singers. Think about 1t - 16 channels of MIDI means 128 notes of polyphony gives you eight voices per channel; and, if you're using all 16 channels, chances are excellent that many of the sounds/parts you'll be using may not use eight voices each...

 

dB

[resigned]

Yeah, unlimited polyphony may be impractical; but then somebody should explain why high-end home keyboards like Yamaha's CLP-series instruments can have 196 or even 256 notes of polyphony but their pro stuff doesn't.

 

This is where softsynths and computer-based music production systems are certainly outpacing hardware keyboard development by offering polyphony that is only limited by the computer's CPU and RAM, both of which are often user-upgradeable.

 

But what's really irritating is when companies like Korg tell us their latest synths offer a "generous 62 notes of polyphony" that's non-expandeable... like we're supposed to be impressed...

 

Keyboardists are spoiled by polyphony? Bunk.

[Dave Bryce]

Originally posted by The Pro:

Yeah, unlimited polyphony may be impractical; but then somebody should explain why high-end home keyboards like Yamaha's CLP-series instruments can have 196 or even 256 notes of polyphony but their pro stuff doesn't.

It can, if you want. If you take a 128 voice MOTIF ES and add the GM and piano expansion boards, I believe that'd give it 256 voices.

 

The beauty of optional expansion is that the people who don't need it don't have to pay for it if they don't want it.

 

dB

[tarkus]

A little off topic -

 

I would layer 10 voices (like an orchestra patch on the Triton) and do a Jerry Lee lewis gliss, or a Keith Emo whoosh - and I would get the weirdest midi lag - as if the cpu couldn't keep up with what I was doing. The first layer of voice would react immediatly then in an almost slap-back the next set of voices would follow. Does this have anything to do with polyphonics?

[cnegrad]

There's another angle to the marketing angle as well. The thinking is that when you start to consistantly run out of polyphony in your home studio, then it's time to BUY ANOTHER NEW KEYBOARD to add to the arsenal!

[Mike Martin]

Yeah, unlimited polyphony may be impractical; but then somebody should explain why high-end home keyboards like Yamaha's CLP-series instruments can have 196 or even 256 notes of polyphony but their pro stuff doesn't.

Perhaps thst "somebody" is me?

 

Have you taken a look at the PRICE of the CLP/CVP products that offer that kind of polyphony? Some are over $10k.

 

dB hits the nail on the head. You can add much more polyphony to a Motif ES with things like the PLG expansion boards. I could be mistaken but I expect that the Clavinova's are reaching these polyphony levels with seperate engines for different types of sounds.

 

All kinds of polyphony is possible..but it all comes at a price. Having worked for a few manufacturers now, I can tell you that we're all trying to deliver the maximum performance at a price that the market will take and that we can be profitable in producing.

[ITGITC]

Remember, a stereo sample takes TWO notes of polyphony. So, in my PC2x, with the addition of the polyphony expansion board, I've got 128 notes available to play at any one time... unless I'm playing a stereo voice like many of their piano programs. Then I'm back at 64 notes of polyphony.

 

Playing glissandos and 'two-fisted' Dave Brubeck inspired piano chords plus using the sustain pedal can eat up quite a bit of polyphony. I added the polyphony expansion board to the PC2x because I just didn't want to hear any notes dropping out (if it could be helped).

 

Kurzweil is known for its intelligent note-stealing algorithym. My 1000PX and K2000 instruments have only 24 notes of polyphony. Sure, back when I was using them for piano voices there was the occasional dropping of notes, but it really wasn't too bad. These are still useable instruments in that regard.

 

Imagine if we had 128 notes of polyphony back in the early seventies when the OB-8 and Prophet synths were in their heyday! Of course, back then if they did have that much polyphony, they would be too expensive for most of us to consider. AND they would be way too large to carry around without a transfer truck!

 

I guess I'll have to agree with "The Pro" from down Hilton Head way - us keyboard players ARE a bit spoiled. But ain't it great to be living in 2004? As Dave Bryce said "The reason the polyphony of synths increases over the years is the same reason computers continue to get faster and more powerful while not becoming more expensive - the technology is improving and therefore becoming more affordable." I STILL think that's pretty cool!

 

Is There Gas In The Car?

[eric]

I still remember playing piano on my Roland JX8P back in 1985 with 6 note polyphony. I had a borrowed Korg SG1D digital piano in the early 90's that was pretty much the gold standard in digital pianos back then. It had 12-note polyphony, if I remember correctly. So I feel a bit spoiled with these 64 and 128 voice keyboards nowadays.

 

Regards,

Eric

[c4]

the promega 3 has 320 note polyphony and isnt exorbitantly priced. Makes you wonder about some companies charging 2 grand and up for pro boards with only 64 note ployphony..

[jerrythek]

I empathize with you - I'd love that as well.

 

But there's no such thing - it doesn't exist and likely never will. Unlimited, I mean. There's always a limit. Depending on the technology you use/develop you maximize the spec to the best ability, considering price and performance. And these elements keep on growing and getting cheaper over time.

 

It is NOT a manufacturer game of holding back, at least not in my 20 years doing this. We try to give you as much as possible balanced with the total product spec and system performance capabilities.

 

If you are using a chip/ASIC-based system you develop the chip to provide certain capabilities. Polyphony is only one factor, the voice architecture is another. The "deeper" the voice architecture the less notes of polyphony you will be able to produce - it's always a balancing act.

 

For instance, when developing the Triton-Rack we found that by stealing 2 voices we could expand the bus architcture to allow the system to "see" MUCH more ROM waves than before.

 

As CPU power increased we found we were able to overclock the system in the TRITON-Studio and new Extreme to double that polyphony. The chip was the same, but with newer/faster CPU's we could push it more at a reasonable price. No "evil" games, just the march forward of technology.

 

In software products it is a balancing act of code - the depth/complexity of the synth egine and sound-producing method versus the computer CPU power requirements. A simple synth voice will yield WAY more polyphony than a more complex one given the same CPU. And a simple form of sound generation (basic, low-quality analog simulation) will yield more voices then a more complex method (deep physical modeling).

 

Here's an observation: a Hammond isn't unlimited polyphony - it has 91 tonewheels. That's it. Why not more? They didn't feel it needed any more and it was expensive enough already.

 

A Clavinet isn't unlimited, nor is a Rhodes. How about a guitar? Nope, nope, nope. They were the right balance for what they were intended for and what their "technologies" could do.

 

:-)

 

But I agree - in synths more is better.

 

My personal "take" is that:

 

"With increased polyphony comes great responsibility."

 

You want more and more effects so that each sound can be fully processed to sound "complete". So I always choose to balance polyphony with effects capability, so that in the end you can play multiple sounds and they all sound good.

 

Regards,

 

Jerry

[Dan South]

This reminds me that I have a question about voice allocation.

 

If you play a piano or an organ, if you repeat a note, the original note stops playing, and the newly played note takes it's place. This is because there is only on set of strings for each note on a piano, and that set is momentarily damped and then restruck even if you're holding down the sustain (right) pedal.

 

Same thing with any acoustic instrument that I can think of. On guitar, pick a note, then re-pick it on the same string. You don't get two strings playing the note; only one does.

 

But electronic synths and samplers, unless they're monophonic, assign a second voice to play the second sounding of the note. Now you have two "voices" playing the same note. This (a) sounds unnatural, especially on ROMplers, and (b) wastes polyphony.

 

Do any synths allow you to control the retriggering of the same note so as to restrict it from grabbing a second voice or set of voices?

[M Peasley]

With 128-note polyphony we have to be getting far out in the realm of diminishing returns. Who has 128 notes sounding at once in their music?

 

I suppose some experimental music might want to goof around with a zillion notes all sounding at once, but that would be a tiny, tiny minority the gear manufacturers would be foolish to try and please.

 

When RAM was really expensive, like when the Kurzweil K2000 came out, polyphony came at a heavy price. They weren't just teasing us into the market with 24-not polyphony so they could later tempt us to upgrade to the K2500 with 48-note polyphony.

 

Someone tell me what in the world they would do with 1024-note polyphony. Or even 256-note for that matter. Something worthwhile, I mean.

 

M Peasley

[ITGITC]

Originally posted by M Peasley:

Someone tell me what in the world they would do with 1024-note polyphony. Or even 256-note for that matter. Something worthwhile, I mean.

 

M Peasley

Ummmmm, perhaps we could string all the oscillators together to play in unison and FINALLY get their attention:

 

http://wyoming.pacificnorthwestmovies.com/CloseEncountersOfTheThirdKind/close15.jpg

 

Of course, we'd have to be set up to play in STEREO !

 

Is There Gas In The Car?

[Dave Bryce]

Originally posted by Dan South:

Do any synths allow you to control the retriggering of the same note so as to restrict it from grabbing a second voice or set of voices?

The Alesis QS synths do - the parameter is called Sound Overlap.

 

I imagine others must allow this option, but I'm not certain. I think the Kurz VAST engine does...

 

dB

[jah3]

This is certainly interesting. I hadn't thought about some of what goes into the process of developing and getting the most out of synths, and other electronic musical instruments in general like that before.

I'm certainly doing fine for the most part with the polyphony of my instruments, although to be fair, they are older gear. (32 notes for TS-12, 64 notes for N264, 30 notes for SY-85).

It is amazing how much manufacturers put into the products they develop.

Quite a very interesting discussion.

It's also great to hear from those in the industry who have keen insights about the developments of the latest technologies.

Maybe one day, when my budget allows it, I can also add some of the latest and greatest gear to my setup as well.

It's always good to keep improving on things.

[fisheye]

When you hit piano strings with the hammer, the hammer will not stop the strings from vibrating completely. The new hit will add new energy and new vibrations to the strings. So adding two samples is a nice approach of hitting a key twice. Therefore, to simulate a piano really good, you need unlimited polyphony.

 

Program your sampler to use a replacing algorithm and it will play bad. Just try it.

[Togakure99]

Originally posted by fisheye:

So adding two samples is a nice approach of hitting a key twice. Therefore, to simulate a piano really good, you need unlimited polyphony.

 

Program your sampler to use a replacing algorithm and it will play bad. Just try it.

I don't agree. Playing two samples together doesn't equal the same as hitting a key on an acoustic piano twice. By playing two samples, you now have two instances of the same note playing, just at different phases in the amplitude/frequency content envelope. On an acoustic piano, for each key hit, you're adding to the energy of the already vibrating string, not starting a new string vibrating at the same pitch, which is what playing two samples would do, in my opinion.

[burningbusch]

I think it was very fortunate the first synths were monophonic. It forced us, piano/organ players, to think outside of the box we've been taught to play within. For the better part of the 1970s, a most creative period, synths were largely monophonic and whole new styles of lead and bass playing evolved as a result. I don't believe this would have taken place had the synthesizer been polyphonic from the start; we would have simply apply our old techniques to these new devices.

 

Some of my favorite synths are mono/duophonic.

 

Busch.

[resigned]

Originally posted by Mike Martin:

Yeah, unlimited polyphony may be impractical; but then somebody should explain why high-end home keyboards like Yamaha's CLP-series instruments can have 196 or even 256 notes of polyphony but their pro stuff doesn't.

Perhaps thst "somebody" is me?

 

Have you taken a look at the PRICE of the CLP/CVP products that offer that kind of polyphony? Some are over $10k.

 

dB hits the nail on the head. You can add much more polyphony to a Motif ES with things like the PLG expansion boards. I could be mistaken but I expect that the Clavinova's are reaching these polyphony levels with seperate engines for different types of sounds.

 

The "over $10k" Clavinova's you speak of all come with a hefty investment in piano-style wooden cabinets. I suspect that if Yamaha ever wanted to produce a portable version of it's high-end Clavinovas for professional use, it could do it for less than $10k.

 

You really shouldn't have gone there with the PLG's Mike because as an owner of both a 9000 Pro and a Motif ES 8, both fully loaded with Yamaha's PLG expansion cards, I can tell you that there are few products on the market that are as poorly supported and documented than the Yamaha PLG-series expansion cards. Motif owners have the Motifator.com forum as a backup resource for help when using the PLG's but 9000 Pro owners don't have any resources at all. The PLG's don't come with any documentation on how to use them in the 9000 Pro, or most of the Yamaha synths they fit in for that matter. I had a hell of a time trying to get Yamaha support to explain to me how to use the PLG-150DR card in a 9000 Pro for example, and it was only after I bought this card and got Yamaha support to ask your engineers in Japan that I found out the limitations of the DR card with my instrument (won't work with arranger styles for example). This is inexcusable because the PLG150-DR card came out well after the 9000 Pro so there should have been some kind of supplemental documentation on how to use it with not only the 9000 Pro but EVERY Yamaha keyboard that the PLG fits in.

 

My point is this: I need mega-polyphony and to get it I've spent a lot of money on my 9000 Pro, Motif ES 8 and five PLG's only to have to spend weeks with tech support to find out what I could and couldn't do to make these products work together as they were designed to. Even if the Clavinova's cost more than what I've invested so far (not by much) at least they would provide the polyphony I need along with full documentation. I'd be very interested indeed to see a portable pro-style version of the Yamaha CVP-900 (nice color display, 256 note polyphony, etc.).

 

If the PLG expansion cards were such a great idea then why hasn't Yamaha followed through and created new PLG's with additional sounds since the DR and PC cards came out? I know I've seen Motif owners asking for new PLG's on the Motifator.com forums but Yamaha hasn't responded with any new versions. I think expansion cards are a great idea but I don't think Yamaha has done such a hot job of fulfilling their potential.

[RABid]

Originally posted by Dan South:

..... Do any synths allow you to control the retriggering of the same note so as to restrict it from grabbing a second voice or set of voices?

The early Roland JV line used to work this way. I'll have to check to see how the Fantom/XV line retriggers the same note. I remember this from hooking up my SP-11 pad to a Roland rack unit to trigger drum sounds. Sloppy playing on the pads would result in ghost notes muting some sounds. The Roland electronic kits trigger new notes while letting old notes ring. While not really working like a drum kit, it does prevent soft notes from having a choke effect when played after a hard note.

 

Last year I asked a few developers about retrigger methods and got disappointing responses. Z3ta+, one of my favorite soft synths, uses the trigger new note method. Rene claims this more closely mimics hardware synths but that is not my experience. If for no other reason I would think that the increased CPU efficiency would be a good reason to trigger new notes. If you are trying to emulate many analogs you not only trigger the same note, you also take into account the effect of the egs. On a Moog a eg driving a filter sweep does not start from zero on a new note, a new note builds on the current state of the eg.

 

For anyone who wants to see what method their sound source uses there are two methods. For software you can hold the sustain pedal and play the same note over and over while watching the polyphonic counter. That is, if your VSTi has a polyphonic counter. The other method is to hit a hard, loud note followed by a very soft note. If the sound mutes the synth retriggers the same oscillators. If you continue to hear the loud note then it triggers new oscillators while letting the old note continue.

 

Robert

[synthetic]

We've had GigaStudio 3 running on some systems with more than 600 voices of polyphony, is that enough for you?