=2C=20--_412babcc-059a-4942-932f-95324755d331_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable> > It doesn't mean the synth will not sound good=2C there are just some=20
The Matrix 1000 users also complained of the quantisation of the mod matrix
for this beast=2C admittedly only when being used for very wide/wild modula=
tion
where again the steps could be heard (*). Perhaps this is getting off the p=
oint: if
work is going to be done on a CV proposal (which is admittedly debatable) t=
hen
quantising to some level at which people have been known to complain of in =
the
past is not a good starting point.
The nice thing about the tuxfamily proposal is that CV is completely 'analo=
guous'
by which I mean analogous to the native sampling parameters. That is badly
phrased=2C I am likely to get hit on for saying it but it does not suffer f=
rom any worse=20
quantisation than the underlying audio signal. The analogue monsters took a=
lot=20
of flac when they started quantising their parameters and the tux design wo=
uld=20
avoid such issues.
Another reason this is getting off the point is that defining a separate CV=
port in
Jack would obfuscate the issue=2C probably. The people discussing ambient c=
ontrol
would want CV at the native rate and that can be delivered now anyway over
audio ports - the only question is how the application uses the data on tho=
se ports.
If a new port type is defined then it will carry some quantised representat=
ion of this
data and then if any app talked about supporting CV you would not know whic=
h
type was being supported. I don't think the overhead in jack would be that =
high=20
for distributing native/audio CV and if people wanted to quantise down to f=
ractions
of the sample rate or resolution because that is how their apps work then t=
his could=20
still be applied as an interface into the native data. The reverse is natur=
ally not true.
Nick.
* The Matrix-1000 (and related 6 and 12) are amongst the best synths ever=
=2C don't
misunderstand my critisism. When they were first released several of the ha=
rdware
magazines refered to them as the most over the top analogue synths imaginab=
le.
If anything=2C a new implementation of CV should take into account the smal=
l number
of complaints that were made against it and parameter quantisation is one o=
f them.
I still want to build an emulator for one of these however I need to finish=
the EMS
Synthi first as the Oberheim mod matrix was loosely based on its patch pin =
panel.
=20
_________________________________________________________________
Hotmail: Free=2C trusted and rich email service.
https://signup.live.com/signup.aspx?id=3D60969=
--_412babcc-059a-4942-932f-95324755d331_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
>=3B >=3B It doesn't mean the synth will not sound good=2C there are ju=
st some >=3B >=3B things that it>=3B >=3B will not be able =
to do.>=3B >=3B>=3B >=3B Regards=2C nick.>=3B >=
=3B Indeed a B3 emulated by an Oberheim Matrix-1000 might be no good choice=
=2C >=3B OTOH because of the automation=2C you won't have any synth w=
ith a good B3 >=3B emulation to change 20 parameters during a song=2C=
while for any >=3B "synthetic" sound e.g. from an Oberheim Matrix-10=
00 you might wish to >=3B change a lot of parameters during a song an=
d changing filter parameters >=3B for "synthetic" sounds 128 steps ar=
e enough.>=3B The Matrix 1000 users also complained of the qu=
antisation of the mod matrixfor this beast=2C admittedly only when bein=
g used for very wide/wild modulationwhere again the steps could be hear=
d (*). Perhaps this is getting off the point: ifwork is going to be don=
e on a CV proposal (which is admittedly debatable) thenquantising to so=
me level at which people have been known to complain of in thepast is n=
ot a good starting point.The nice thing about the tuxfamily proposa=
l is that CV is completely 'analoguous'by which I mean analogous to the=
native sampling parameters. That is badlyphrased=2C I am likely to get=
hit on for saying it but it does not suffer from any worse quantisatio=
n than the underlying audio signal. The analogue monsters took a lot of=
flac when they started quantising their parameters and the tux design woul=
d avoid such issues.Another reason this is getting off the poin=
t is that defining a separate CV port inJack would obfuscate the issue=
=2C probably. The people discussing ambient controlwould want CV at the=
native rate and that can be delivered now anyway overaudio ports - the=
only question is how the application uses the data on those ports.If a=
new port type is defined then it will carry some quantised representation =
of thisdata and then if any app talked about supporting CV you would no=
t know whichtype was being supported. I don't think the overhead in jac=
k would be that high for distributing native/audio CV and if people wan=
ted to quantise down to fractionsof the sample rate or resolution becau=
se that is how their apps work then this could still be applied as an i=
nterface into the native data. The reverse is naturally not true.Ni=
ck.* The Matrix-1000 (and related 6 and 12) are amongst the best synths=
ever=2C don'tmisunderstand my critisism. When they were first released=
several of the hardwaremagazines refered to them as the most over the =
top analogue synths imaginable.If anything=2C a new implementation of C=
V should take into account the small numberof complaints that were made=
against it and parameter quantisation is one of them.I still want to b=
uild an emulator for one of these however I need to finish the EMSSynth=
i first as the Oberheim mod matrix was loosely based on its patch pin panel=
. Hotmail: Free=2C trusted and rich email servic=
e. Get it now.
=
--_412babcc-059a-4942-932f-95324755d331_--
LINUX® is a registered trademark of Linus Torvalds in the USA and other countries.
Linuxaudio.org logo copyright Thorsten Wilms © 2006.
Hosting provided by the Virginia Tech Department of Music and DISIS.