Music Player Network
Previous Thread
Next Thread
Print Thread
#434172 03/08/00 01:19 PM
A
Anonymous
Unregistered
Anonymous
Unregistered
A
are there any plans to make your EQ and Dynamic package, designed for oxford, sonic and mackie available for sony's new "baby oxford" as well?
have you had a chance to see this console, or get an impression on how it performs, sonically? if so could you coment how it compares to other digital consoles you know.

#434173 03/08/00 03:22 PM
Joined: Feb 2000
Posts: 2,184
gm Offline
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Feb 2000
Posts: 2,184
In my never-ever-so-humble opinion, the "sound" of digital is almost directly related to available processing. up to a point, of course. for example, the dbx Quantum box pushes 500mips and still falls on it's face in several areas.

The R100 has but a slight connection with the Oxford research, and I'm not sure if "baby Oxford" is an appropriate term. The R100 is based on a handful of SHARC's; the Oxford on a custom processor deployed in a large, flexible array. The Sony Japan design team decided to "re-invent" the software, and declined to utilize by far the most powerful feature of the Oxford: the graphical, hierarchical design Tool.

I haven't heard it yet, but look forward to listening to it to see how they did.


George Massenburg

http://www.massenburg.com
#434174 03/09/00 11:57 PM
A
Anonymous
Unregistered
Anonymous
Unregistered
A
you're right about "baby Oxford" not beeing appropriate, it's just a term i read somewhere on the web. of course i am aware, that the R100 plays in a totally different league.

in an ideal world - would a handful of sharc's provide sufficient power to sound good?

is it appropriate to assume that this graphical, hierarchical design Tool enables you to theoretically use your algorithms originally developed for the Oxford?

whenever you get a chance to listen to it and see how they did, could you please post your impressions on this board?

---

please excuse the spelling of some of the words i use, but i guess you can forgive foreigners.

#434175 03/10/00 06:51 AM
Joined: Feb 2000
Posts: 2,184
gm Offline
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Feb 2000
Posts: 2,184
Dear Mischa,

First, it's clear that if you *like* 32 bit floating point, then SHARC's would sound o.k., and a large array of them would probably provide the power necessary to do some real work.

Second, I believe that the hierarchical design "TOOL" would make transfering algorithms much easier, although some attention would have to be paid to the architectural differences (32bit float in SHARC's vs 32 bit fixed in the SPIC's).

Will let you know how it sounds when I hear it.

George


George Massenburg

http://www.massenburg.com
#434176 03/10/00 01:16 PM
A
Anonymous
Unregistered
Anonymous
Unregistered
A
Dear George,

To a novice in the discussion about architectural differences in DSP design, could you please explain the major differences of 32 bit fixed vs 32 floating and how this presumably translates to the "sound" of a system utilizing fixed vs one containing 32 bit floating DSP's.

Mischa

#434177 03/10/00 11:42 PM
Joined: Feb 2000
Posts: 2,184
gm Offline
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Feb 2000
Posts: 2,184
Dear Mischa,

The audibility questions between 32bit floating and fixed are many, and have not really been answered in any conclusive way. I know of no listening protocol to evaluate and conclusively quantify the differences.

We are learning more, however, about how one wants all of the resolution one can possbily get, particularly in the extended math you find in the innards of digital filters. 32 bit float utilizes a 24bit mantissa and 7 bit exponent, and multiplies are usually done 24 x 24, although because this 'floats' it has a very wide dynamic range for a given artifact level. 32 bit fixed (such as the Sony/Oxford Digital product) has that much more resolution at a given level, but somewhat less dynamic range (160dB vs 1000dB, approximately). The Motorola 563xx processors (nominally a 24bit fixed point format) have a mode which concatanates the A & B accumulators and allows 48 x 48 bit math. You take a big processing hit, but it's got resolution to spare.

As you limit the word width, the two artifacts that increase most obviously are noise and distortion.


George Massenburg

http://www.massenburg.com

Moderated by  gm 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.5