DSP comparison request

[Warp69]

Some high end reverbs use ray tracing..................

Oh, which high-end reverbs use raytracing?

[greenbluegold]

VSS4 algorithm in the TC electronics Reverb 6/4000 is the ray tracing reverb that I was thinking of. Not sure as to the details.

[astroman]

ok, I'll elaborate my point of pointless a bit because that's what it really is about... pointless
afaik there are no acoustic transmitters that allow a precise 'positioning' of sound in space similiar to the dots on screen which represent a fraction of a real world beam of light. Even the latter takes an enormous amount of effort to be convincing.

Our ears are used to decipher a mess of waveforms regarding content, location and motion of the (supposed) source - for obvious reasons
As we 'function' by indirect 'processing' it really doesn't make sense to apply absolute geometric methods.
I'm not as familiar with the details as Warp, but imho the approach of algorithmic reverbs is feeding our ears exactly that mix of signals that 'triggers' the desired sensation.

cheers, Tom

ps: GPUs are only 'powerful' if you have a lot of data which needs a relatively simple processing, so you can deal with a few hundred elements or so in one cycle. When the data is continually changing, handling of the elements may suppress a significant part of the chip's processing power.

[siriusbliss]

Actually, the GPU developments are not entirely off-topic, and DO point to the increased desire to get more dedicated processing power off the CPU.

But nevertheless, the main advantage of our lovable Scope is having dedicated processing that isn't sideline processing other tasks.

Now, which one of you synth designers will create something equivalent to Omnisphere to show off the real horsepower?

Greg

[astroman]

I'd rather say it's less about the power to create such sounds, but to handle the environment.
Omnisphere is not exactly powerful, but extremely convenient... imho

cheers, Tom

[siriusbliss]

I'd rather say it's less about the power to create such sounds, but to handle the environment.
Omnisphere is not exactly powerful, but extremely convenient... imho

cheers, Tom

I agree

Greg

[greenbluegold]

ok, I'll elaborate my point of pointless a bit because that's what it really is about... pointless
afaik there are no acoustic transmitters that allow a precise 'positioning' of sound in space similiar to the dots on screen which represent a fraction of a real world beam of light. Even the latter takes an enormous amount of effort to be convincing.

Ah... see, my interest is not in some kind of point to point translation of physical acoustical data into graphical data matrices, but to use the same kind of data structures that are used for say vector graphics and manipulate them using vector graphics algorithms and use those data sets for synthesis. You could set the parameter correlations to whatever you like. This way you could still present a graphical representation of whatever the parameters are.

What I'm thinking of could be as simple as just putting your synthesis parameters (say sample start time ,filter cutoff, delay time, stereo spread etc...) into multi-dimensional matrices then use graphical processors to do matrix math on them using gpu libraries. You could display the data graphically, and send the output back to a synthesis algorithm for audio output. Inversely, you'd be able to take graphical date and convert it into parameters for sound synthesis.

[astroman]

well, just edit your posts and remove that blank behind the 'closing' slash
same applies to the 'bold' typeface - no whitespace allowed between the slash and the tag it's closing. Never mind, such stuff happens

cheers, Tom

[Warp69]

VSS4 algorithm in the TC electronics Reverb 6/4000 is the ray tracing reverb that I was thinking of. Not sure as to the details.

Im afraid not. They just used an enhanced VSS3 algorithm. The VSS4 does not use raytracing.

But both Lexicon and TC Electronic uses acoustic auralization software when designing reverbs :

TC Electronic : www.odeon.dk
Lexicon : www.catt.se

[kylie]

ps: GPUs are only 'powerful' if you have a lot of data which needs a relatively simple processing, so you can deal with a few hundred elements or so in one cycle.

SETI

[greenbluegold]

VSS4 algorithm in the TC electronics Reverb 6/4000 is the ray tracing reverb that I was thinking of. Not sure as to the details.

Im afraid not. They just used an enhanced VSS3 algorithm. The VSS4 does not use raytracing.

But both Lexicon and TC Electronic uses acoustic auralization software when designing reverbs :

TC Electronic : http://www.odeon.dk

A ray tracing program was used to analyze early reflections. The data collected was used for the odeon project, from which the auralization program was developed. So the algorithm is several steps removed from the ray tracing.

http://www.tcelectronic.com/Default.asp?Id=12266

[siriusbliss]

Interesting breakdown of DSP/native/FPGA and Pyramix' Masscore.

http://www.merging.com/products/show?product=1&page=54

Greg

[throbert/thor]

"I'm trying to avoid sound comparisons because frankly, no offense, but I think the younger generation has been listening to crappy mp3's for so long, that they wouldn't know good quality audio (or be able to feel the grit of a good filter) if it hit them in the head."

You're more right than you might think.

http://technology.timesonline.co.uk/tol ... 847674.ece

Oh yea it's every where
A child is feed crap will eat and lke it and wounder why their elders are pissing and moaning while others are LOL liting a cigar with $100 bill

[nightscope]

[siriusbliss]

Hi guys,

I'm bumping this old thread to help give support to SC and our beloved (pending) Scope 6 developers - to help us distinguish between DSP and CPU horsepower in today's systems.

I'm tired of the haters and naysayers on other forums constantly poo-pooing what Scope / Xite (DSP-based) gives us vs. the exclusive use of VST/VSTi's (no matter how GOOD these VST's might be).

I'm still using a mix of VST (mostly Samplitude's native plugins), and Scope for live tracking/mix/mastering, so I can appreciate any value-added data to support how we use and develop Scope going forward.

Thanks!

Greg

[dante]

The main advantages of (Scope) DSP over Native as I see (and hear) them :

1) Sounds better (although this is arguably subjective).
2) DSP designed for Audio (wider data paths, more parallelism).
3) Low Latency (even if native could ever 'sound as good' it would still lag timewise).
4) Developers don't have to compromise sound quality to obtain performance ( eg fidelity, zippers etc).
5) Load is pre-allocated (less chance of processor overload mid performance - you know your limit before you start).

Any more ?

[Warp69]

I'm bumping this old thread to help give support to SC and our beloved (pending) Scope 6 developers - to help us distinguish between DSP and CPU horsepower in today's systems.

Wouldn't it better to bump/create threads on other boards to support SC?

There will always be some people that don't like or don't understand the concept of the Scope environment. I have a lot of outboard and Im used to connect my units via multiple patchbays, so Scope is absolutely perfect for me - the flexibility is awesome. Most people writing on different boards have never ever worked in an environment with patchbays and therefor simply doesn't acknowledge the benefits in the Scope concept.

Regarding sound quality - I believe that native have surpassed Scope in some ways - not in all areas, but quite a few. That's absolutely not the same as saying native is better sounding than DSP in general - my most loved units and best sounding are of course DSP based, but they're also way ahead of Scope and native in quality.

[siriusbliss]

There are already threads going on Gearslutz, harmony central, KVR, and Xited.

The response has been positive overall, but there are still ignorant haters that don't care to do some research.

Greg

[dante]

Screw em. Like Jimmy says, makes it easier for us to smoke em

[YISH313z]

Man I remember passing over Creamware for so many years, I had no idea what the h*ll it was, there was very little info other than, dsp card and external boxes, these I stopped looking at because I was all native (I started with hardware before VST though) , then after straining my ears and trying to get native to sound real(relative to digital audio hardware), I just researched everything that was out there, because I wasn't convinced. I was pissed off at myself for not getting in earlier. It jumped out and bit the hell out of me. And to think I almost forgot to come back after I sold my first home card to attempt to buy a car(Fail).

To be fair to the OP request, We are running our DSP chips at 66mhz and 333mhz, strictly for signal processing, the audio portion of the DSP system is Front and Center, and is hardwired to the converters.(You know these things) It doesn't do anything else. But it does this better than any native system. The sound is different, because there is no need to compromise for other resource allocations.

With ASIO and all the other audio drivers its a work around to a system that will not bend for the audio and musician's sake.
DSP's should have been apart of the Motherboard to begin with, but since these people are in the business to save and make money the quickest way, the minorities in the market(Audio Enthusiast and Musicians) are largely ignored and with a workaround are forgotten. Meanwhile the native platform moves forward because its cheap and easy. Too easy. Too Cheap Sounding(IMHO!!)

If someone was to make a DSP system based on Intel Core 2 Duo chips or whatever, I'm sure it would smoke DSP's in a pipe and smile. But(and those who know please correct me), it would take a lot more for CPU chips to equal the simplistic nature of these DSP chips and let us not forget the heat factor.

Raw power is relative to its intended use, Power therefore is useless until you use it.

Okay, sorry for my ranting...
Just my 2 cents






OOPS...I went smiley crazy there.