Native vs. UAD, comparisions....

[eduardo_b]

You're too opinionated to be curious. The tools don't matter anywhere near as much as the person using them. They can't replace talent. Spending more money won't change this. There's nothing to quantify.

Once again, missing the point. Also I'd like to know how I'm "too opinionated to be curious". That's blatantly false. There is a very simple truth of reality that tools do indeed matter. Whether you choose to place their level of importance low or high, none can argue that they matter. That is all I was saying.

I presented my first question to you because you said the music will be good regardless of what tool is used. While it may be true on paper, the tools affect how the music sounds, and conversely how we enjoy them. A song recorded with a 10 dollar microphone on a 4 track mixer with an out of tune squier would not sound nearly as good as one recorded with superior tools. Are you going to argue that this is not true? A simple yes or no will suffice.

Look, you're thinking that the tools matter so much because you're on this end of the process. The listener at the other end doesn't give a rat's ass about the tools. They don't know, don't care. They will or will not enjoy the music regardless of whether you used a $400 virtual compressor or a free one. That's my point. Yes, it's nice to have good tools, but that is not going to have any significant influence on how good the listener perceives the music to be. That is and remains my point. Your point is trying to prove otherwise...even though a talented user can make good music with any tool that is used with skill, whereas the most expensive plug-ins available can't make up for a lack of talent and skill. It's really that simple.

[dirty oscillators]

i don't think rexxz inferred any of that.

[Cap'n Spanky]

I have a a highly nuanced position that I want to argue about!

[adydub]

Whilst I certainly agree that musical ideas are more rather more important and fundamental than the tools used to express them, I'm still interested in in the tools. That's why I sometimes come on KVR .

As far as I'm concerned, the more processing power I can get my hands on, the more potential there is for for making interesting noises. And I am a geek. So I found this very, very interesting:

UAD2 Quad - 6,4 Gigaflops - ~1500€
http://www.gearslutz.com/board/music-co ... ost3239925

Intel Core2Duo E8400 - 24 Gigaflops - ~140€
http://www.intel.com/support/processors ... 023143.htm

GForce 8800GT - 504 Gigaflops - ~100€
http://en.wikipedia.org/wiki/GeForce_8_ ... al_summary

This illustrates why general purpose computing on GPU's (GPGPU), is a massively compelling proposition. CUDA, and now OpenCL, are finally making this something that's now not far 'round the corner.

As pretty much every newish machine has an OpenCL compatible graphics card, a plugin running on a GPU over OpenCL has most of the convenience of a native plugin, but with access to huge amounts of untapped grunt. Of course laptops with integrated graphics like my Macbook get left out of the party .

There's a good chance this will render the current dsp plugin platforms obsolete sooner rather than later. Of course, people have been saying this for years with reference to CPU power increases vs dsp, but GPGPU could well be a game changer.

Having said this, I have no plans to chuck out my LiquidMix or any of my dsp powered hardware just yet! But I'd certainly be wary of investing in something like the UAD2 Qaud and a load of UAD format plugins.

EDIT: doh - OpenCL not OpenGL - damn force of habit.

[lavoll]

if this thread was a pluging, it would be a bit reducer, automated to slowly reduce bits over time.. i think we are down to maybe 4 bits now.

[eduardo_b]

i don't think rexxz inferred any of that.

He asked a "question" while saying my point was contradictory. His focus on the tools to the exclusion of any other considerations -- certainly they weren't stated by him if there were any -- is incorrect. I gave him a larger picture to view the tools in.

Had he worded his first reply differently, I wouldn't have taken him to task for it.

[adydub]

if this thread was a pluging, it would be a bit reducer, automated to slowly reduce bits over time.. i think we are down to maybe 4 bits now.

Well, if we can get it down to 1-bit @ 2.8224 MHz everything will be wonderful.

[bmanic]

Whilst I certainly agree that musical ideas are more rather more important and fundamental than the tools used to express them, I'm still interested in in the tools. That's why I sometimes come on KVR .

As far as I'm concerned, the more processing power I can get my hands on, the more potential there is for for making interesting noises. And I am a geek. So I found this very, very interesting:

UAD2 Quad - 6,4 Gigaflops - ~1500€
http://www.gearslutz.com/board/music-co ... ost3239925

Intel Core2Duo E8400 - 24 Gigaflops - ~140€
http://www.intel.com/support/processors ... 023143.htm

GForce 8800GT - 504 Gigaflops - ~100€
http://en.wikipedia.org/wiki/GeForce_8_ ... al_summary

This illustrates why general purpose computing on CPU's (GPGPU), is a massively compelling proposition. CUDA, and now OpenCL, are finally making this something that's now not far 'round the corner.

As pretty much every newish machine has an OpenGL compatible graphics card, a plugin running on a GPU over OpenGL has most of the convenience of a native plugin, but with access to huge amounts of untapped grunt. Of course laptops with integrated graphics like my Macbook get left out of the party .

There's a good chance this will render the current dsp plugin platforms obsolete sooner rather than later. Of course, people have been saying this for years with reference to CPU power increases vs dsp, but GPGPU could well be a game changer.

Having said this, I have no plans to chuck out my LiquidMix or any of my dsp powered hardware just yet! But I'd certainly be wary of investing in something like the UAD2 Qaud and a load of UAD format plugins.

Which is why I've been saying all the time that the future is with plugins like Nebula. They realized early on the potential of CUDA and already have their plugin running pretty well on the Nvidia platform. Once it gets some more polished over the coming years it'll be pretty hard to beat both in terms of sound and performance (it already does by far the best hardware reverb emulations I've ever heard).

Cheers!
bManic

[The Chase]

Then study harder. Cause right now you're failing 101.

Like what? So far no one has really provided any compelling falsifications to his claims.

the point was that he was comparing processing power between two processors which are not designed to do the same set of tasks. i could be wrong but it seemed that he was in fact comparing apples to oranges to make his point.

That's why gigaflops are such a good comparison though. Had he been comparing partially architectural things like Ghz then it wouldn't be a fair comparison, but FLOP/S is purely a performance spec and therefore a reliable comparison between different types of systems.

If you're into gigaflops and whatever, why not get into gaming, extreme overclocking and stuff ? They got nice fps, too.
Last time I checked it was notes, lyrics and sound that mattered in music.
Oh well,
susiwong

Importantly, the responses comparing the processors are in no way making the tools any less about their sound. It's still about sound to everyone.

Everyone here seems to agree that they sound great. The thing is that there are still people who claim that the reason they sound great is on the basis of CPU vs. DSP, which is on the whole incorrect, misleading, and IMO detrimental to how people think about their tools.

Notes, lyrics and sound are important to music, but when making music, knowing the tools you use to portray the notes, lyrics and sound is important as well. Bench tests are one of many tools to see which ones will be the most effective regarding what we use to portray these things.

Also, no matter how important the sound is over technical specifications, an invalid will always be rightly subject to correction. So even DSP vs CPU doesn't matter at all, no one is in the wrong for correcting someone when they say that a particular example of one is superior to another, when it isn't.

[Nokenoku]

price that a recent UAD-1 card went for on Ebay:

My father bought a fully working 12000$ laser printer on eBay once ... for 16$.
--> Argument failed

What I'm saying is, that for most user the power of their CPU is simply totally enough. And if they'd actually need more power, they could buy a better CPU for less money than a UAD-card (which is offering much more additional processing power than such a card).

which is WHY you use the card to free your computer's CPU to run additional native plugs, ...

Do you actually read, what you quote?

Just a quick little note here that native versus DSP versions of seemingly identical plugins can sound different at identical settings. Good examples are the Waves RTAS versus Waves TDM plugins. I've posted the wave files here several times. The differences are quite subtle but they are there. Also notable is that doing a phase cancellation revealed small "spikes" which means that the transients were not processed the same in both plugins.

Do you still have some links to those files?

[WillieJenkins]

Then study harder. Cause right now you're failing 101.

Like what? So far no one has really provided any compelling falsifications to his claims.

Writing generic C code on x86 that you cannot optimize because you don't know what exact hardware is running underneath vs. writing assembler optimized code for a single platform that you total knowledge of alone makes a huge difference in performance. Never mind the fact that you're quoting theoretical GFLOPS, not real world GFLOPS.

That said, a modern cpu with that kind of power consumption is going to destroy a SHARC chip, but it's just not as wide as theoretical GFLOP performance makes it seem. Also, consider that some people do like the fact using the UAD's are light on actual computer CPU. Knowing that with your UAD-2 Duo, you get 32 Neve Consoles, no matter what else is going on, on your computer is useful for planning.

Another thing to consider is power consumption. Another argument is that these GPUs that draw like 100+ watts of power alone can beat out SHARC chip performance. Yeah, no kidding, but what happens when UA releases a quad core card with 4 of those GPU's in it, and you tell the consumer, oh yeah, and you need to buy a 350 watt power supply bigger then the one you have just to run this one card.

When designing/picking an architecture to use, there are tons of factors involved, not just 'how fast is it', there's tons of other considerations that must be evaluated. availability of chips, power consumption, scalability.

Thats just from a pure engineering standpoint.

The other thing is that the UAD hardware *is* a dongle, and from a real world perspective, that why they can make the emulations they can. DSP Engineers cost money, and since people are forced to buy the hardware, it ensures they can pay the engineers and take their time really modeling the hardware. As other people have pointed out, it's attractive to hardware manufacturers knowing they'll get their licensing fees, as opposed to everyone in the free world downloading it.

UA is an actual company, they have real world business decisions to consider. This has to be taken into account when discussing what they do. The theoretical rambling is nice and all, but it doesn't map into the real world. Engineering types above all else should know this.

If the initial guy I was responding to really is focusing in engineering, he should use UA as a case study, there are alot of real world concepts he could take away from it, and it would apply some real world examples to couple with the theory he's learning in school.

-E

[AndrewSimon]

The other thing is that the UAD hardware *is* a dongle, and from a real world perspective, that why they can make the emulations they can. DSP Engineers cost money, and since people are forced to buy the hardware, it ensures they can pay the engineers and take their time really modeling the hardware. As other people have pointed out, it's attractive to hardware manufacturers knowing they'll get their licensing fees, as opposed to everyone in the free world downloading it.

Exactly, I'm sure UA could offer all their plugs in native format if they wanted too.... but they don't want too!

Read the company history on the UAD site, it's all there, it's not a secret,
they attribute their success to the willingness of companies
to license their products to be emulated because it's a secure format.



Now... nobody likes dongles... and I'm dongle free.
I am however a UAD costomer for many years.
A UAD "dongle" helps my speed.... all other dongles just slow it down.

[siriusbliss]

I'll take Sonic Core Scope synths and effects over native effects/synths - even though I own and still use native plugs - I usually run native synths through Scope anyways during a mix in order to get more 'beef'.


Greg