Emu XL-7 & MP-7 User's Group

--- In xl7@yahoogroups.com, "Ravi Ivan Sharma" <noision1@h...> wrote:
> A layer in a patch is a completely different thing than having more 
than one osc or dco in a voice. So I don't think your casio 
comparison is correct. The casios you mention never had more than 8 
voice polyphony no matter how many dcos or "lines" per voice were 
available.

Actually, on the CZ-5000 and CZ-3000 I could play 16 different notes 
<*live*> if my preset was a one-layer [aka one-line] preset. In fact, 
I created a couple of patches to take full advatage of this feature.  
Nevertheless, Casio didn't try to deceive consumers.  They were 
clear: 8 voice polyphony with a 2-layer preset; 16 voice polyphony 
with a 1-layer preset.
 
> Perhaps E-mu could have said, "4-layer patches" rather 
than "voices" but there is nothing new about the fact that most 
modern synths allow you to layer voices, but that such layering eats 
up polyphony. For example the Roland JV units, etc.

Excellent suggestion for revised manuals (calling them 4-layer 
presets). That would be better than what they did.  They could also 
call the gear 128-layer sound modules or command stations and have an 
chart and/or explanation of how flexible the layering capabilities 
are.

As for Roland, let me quote from a sub-section of "How the M-GS64 Is 
Organized."  The sub-section is entitled, "Voices and Maximum 
Polyphony."  

"The sounds of the M-GS64 are produced by units called 'Voices.' 
There is a limit to how many of these voices can sound at once, and 
in the case of the M-GS64, up to 64 simultaeous voices can be used.  
Some sounds (Patches) use one voice and others use two voices. etc.'

Although I would argue Roland's use of voices here (1) means what 
they used to call partials and (2) differs from what polyphony 
traditionally means, they went the distance to explain to the user 
that they will not always have 64 voices.

I could be wrong, but E-Mu manuals (1) don't offer such an 
explanation (except in reference to turning on chorus within a 
preset), and (2) seem to conceal the limitation with their 128-voice 
statement so prominently in the introduction to their manuals.
 
> Also remember that layering does not necessarily have anything to 
do with polyphony, but only does depending on how one uses it.

Excellent point.

> One could have a 4 layer patch that assigns each layer to exclusive 
ranges of midi note number, or exclusive ranges of key-on velocity 
for example. I.e. velocity 0-40 sounds layer 1, vel 41-61 sounds 
layer 2, vel 62-100 sounds layer 3 and vel 101-127 sounds layer 4. In 
such a patch, one could still hold down 128 separate and discrete and 
differently tuned voices.

Excellent example.  All the more reason to call the gear 128-layer 
sound modules/command stations and to include a section showing the 
flexibility of configuring the 128 layers.  (BTW, I like your use 
of "voices.")

> On the other hand, if one programs a patch to layer two voices at 
all time, you are essentially playing two completely different sounds 
(each with their own adsrs, filters, amps, etc) at the same time, as 
if you midied two synths together, so yes a single keypress sounds 2 
voices, but the fact that it does so simulataneously does not mean 
they are now converted to 1 voice . . .

Ooops.  Not a good example in my opinion. Here, it seems you are 
using voice and layer interchangably.  I would argue that on a 128-
layer sound module a 4-layer preset spead over the full range of the 
keyboard yields 32 voice polyphony.

Consider your last example.  If an orchestra hit sample consists of 
64 different instruments played simultaneously and placed onto a ROM 
as a single instrument, and if that sample is assigned to L1 in a 
preset spread over the full range of the keyboard (and L2-L4 are set 
to none), you have 128 voice polyphony, not 8192 (128 * 64) voice 
polyphony. 

In other words, if you can press and hold only up to 32 keys at a 
time (or have a sequencer send the sound engine only up to 32 note-on 
messages at a time) without either additional note-ons not sounding 
or previously sounding note-ons cut off, you are operating with 32 
voice polyphony.  Even if the sampled ROM instrument being used on 
each of 4 layers was created with a chior of 72 human voices (18 
each - SATB).

I can hear it now: Expand your Proteus/Command Station to 9216 voice 
polyphony with the E-Mu CosmoChior ROM (MSRP $279).  [Each ROM sample 
to contain 72 voices.]  <---Tongue-in-cheek

---Steve

>   ----- Original Message ----- 
>   From: steve_the_composer 
>   To: xl7@yahoogroups.com 
>   Sent: Wednesday, June 18, 2003 9:06 AM
>   Subject: [xl7] Re: polyphony, was "arpeggiator out useless?"
> 
> 
>   --- In xl7@yahoogroups.com, aeon <aeonlux@a...> wrote:
>   > On 6/17/03 9:44 PM, "steve_the_composer" <smw-mail@p...> wrote:
>   > 
>   > > As long as we're talking about polyphony, let me ask, does
>   > > 128 voice polyphony = 32 notes x 4 layers per note?  Or do
>   > > I have 128 4-layer voices?
>   > 
>   > think of it as 128 total layers
>   > so 32 4-layer voices...
>   > or 16 4-layer voices with voice-chorus
>   > or 16 4-layer dual-link voices
>   > or 10 4-layer tri-link voices (for that killer dozen-detune-saw 
>   patch)
>   > 
>   > etc.
>   > 
>   > 
>   > cheers,
>   > aeon