UAD Spark Subscription model opinions?

[billcarroll]

Today licenses for my API 2500, Galaxy Tape Echo, and Lexicon 224 appeared in my account. I'm pretty happy to have the API 2500 native. I'm still waiting on other licenses, like the Studer A800 to show up.

[zvenx]

Well I got my first three: API Vision Console, API 2500 and Neve 1073...
when you load it the gui starts blurry but focuses within a second.. still no resize alas... what is great though is the cpu usage on my system is decently low, compared to other similar native plugins from other developers, what is surprising and disappointing as a Cubendo user is that there is no sidechain in the compressor. This is always what I look forward to most in vst3 plugs in Cubendo..

Overall I am pleased thanks UAD. Bring them on
rsp

[electro]

I always understood all UAD plugins were developed natively and only then ported over to SHARC DSP.

I think it has to be an emulator. Im not sure if this is somehow connected, but if you want to test this out yourselves

1. Open the minimoog UA plugin in your DAW, try it in Ableton, or FL Studio as they both have Qwerty keyboard enabled options without an additional Qwerty pop up window.

2. Start to play the synth with your computer keyboard

3. Start to click on some parameters within the synth while playing the plugin with your computer keyboard

4. You will notice that Your Mac starts to produce a bleep sound every time you press a key on your computer keyboard

I think the all the native plugins must be running some sort of wrapper or something because OSX doesnt recognized the UI, so your mac emits a bleep noise everytime you hit a key. Clicking away from the plugin with your mouse gets rid of the bleep sound and you can continue to play your synth with your computer keyboard without a bleep

I dont know if this is somehow connected, but its very annoying

[jamcat]

what is great though is the cpu usage on my system is decently low, compared to other similar native plugins from other developers

That confirms to me that UAD's move is an admission their DSP "accellerator" hardware is a boondoggle that cripples users with an expensive, glorified dongle that can't keep up with native multicore CPUs.

[noiseboyuk]

what is great though is the cpu usage on my system is decently low, compared to other similar native plugins from other developers

That confirms to me that UAD has finally admitted their hardware DSP is a boondoggle that cripples users with an expensive, glorified dongle that can't keep up with native CPUs.

Well now. Half true?

UAD-2 was released in 2008, boasting "ten times the power of UAD-1". It is therefore based on 15 year old technology. I doubt anyone believed that the UAD-2 units in 2008 were dropped from Vega with technology that didn't exist anywhere else on planet Earth - it was just (at the time) a cost-effective way to increase audio processing power rather than using a whole other computer to do the number crunching.

Since then there have been hardware units offering double, quadruple the grunt, but it's still small beer compared to the improvements in host CPU. The M1 series in particular should leave it for dust.

So any talk of the hardware only ever being a dongle or that there was some great fraud perpetrated I think is a bit silly. They had a good product and a good model that worked brilliantly, still works perfectly well but yes with each successive year the performance - relatively - has decreased. In 2022, the hardware offerings don't make much sense any more. I think they've done well to keep it going this long - longer than Avid managed with their own DSP solutions - but it's high time to move on.

EDIT - forgot to say that the hardware platform still allows for much better real time use when tracking.

[jamcat]

It always served a dual purpose as a dongle. The difference is now that’s all it’s good for and they can’t pretend otherwise.

[noiseboyuk]

It always served a dual purpose as a dongle. The difference is now that’s all it’s good for and they can’t pretend otherwise.

Except the real time thing. The irony is that they've stopped manufacturing some of the hardware that was best suited for that (eg Live Rack). But the interfaces with quality preamps etc are also still good I guess. It's the Accelerators that should be made redundant by migrating everything to native.

[trackbout]

The reason I’ve always avoided UA is that they’ve always been economically incentivized to write inefficient code. Any software application can be implemented in such a way that it consumes more or less resources, without changing its ultimate effect. That is a simple statement of fact, not opinion.

Let’s say your first attempt at a plugin that emulates a Neve console consumes X amount of CPU. You could very easily add some inefficient code to that plugin such that it would consume 2X CPU, without changing the way it sounds at all. Likewise, with some effort, you might possibly be able to refactor some code in that plugin such that it would consume 1/2X CPU, without changing the way it sounds at all.

The basic point being, a software plugin’s impact on the audio signal, and its impact on the CPU, are two completely different things that can change independently of one another. They are related obviously, but they are in no way perfectly correlated.

Up until now, UA has had an extremely powerful economic incentive NOT to optimize their code. The less resource efficient their plugins are, the more DSP boxes you need to buy to run their plugins, the more money they make.

Now their incentives are more properly aligned, as with all native developers, the less optimized their code is, the fewer native licenses they sell, the less money make. That is potentially a huge win if you’re interested in their plugins.

I do think they owe people transparency about whether the Spark plugins are running in a SHARC emulation, or whether they are true native code ports, not because I particularly care one way or the other, but just in the interest of truth in advertising. I find them completely untrustworthy at the moment, but I’d prefer not to.

[zvenx]

Though I don't agree necessarily that they are doing this, I do follow your logic. Which leads me to my question (pun unintentional)... do you think Apple does the same with their native apps like Logic, or Final Cut Pro? Intentionally not optimize it as best as possible to make one buy a faster and more expensive Mac?

rsp

[trackbout]

To be clear, I am not saying I know that UA intentionally writes inefficient code, I'm just pointing out the facts of their incentives. As for my opinion about what they might do, I will quote a very successful investor who famously said, "show me the incentives and I will show you the outcome."

I think Apple's incentives are infinitely more complex, and therefore harder to draw conclusions about. Their "pro" apps drive a tiny fraction of their hardware sales.

[jamcat]

Though I don't agree necessarily that they are doing this, I do follow your logic. Which leads me to my question (pun unintentional)... do you think Apple does the same with their native apps like Logic, or Final Cut Pro? Intentionally not optimize it as best as possible to make one buy a faster and more expensive Mac?

Yes, sort of. I have a 2012 Mac mini with the latest OS it's capable of running, which is Catalina. the previous couple Mac OS versions started getting pretty sluggish. And it goes for Windows, too. Windows 8 was alright, but Windows 10 is painfully slow on it. Worse than Catalina. I have Ubuntu on it too, and it's really snappy.

As computer power increases, there is less incentive for developers to stringently optimize their code. Almost no one hand-codes or fine-tunes in asm anymore. But that used to be extremely common.

And importantly, the optimization they are doing is targeted for the latest systems, which have new routines and instruction sets not found in the old hardware.

[v1o]

Spark also proves that all UAD accelerator cards are mostly dongles

That's a really bizarre conclusion which would only be the case if the Native versions used zero CPU. They don't use zero CPU.

It's possible these native ports are not as efficient as possible.

I say that because the Brainworx plugins that started life as exclusively DSP based UAD plugins, and were later ported to native x86 plugins, they use very little if not zero CPU.

[Virgil]

Explained in the FAQ. They are working through depositing licenses for everybody, but it could take up for four days to complete. You’ll need to be patient, but they will show up…

Yeah, I got the 1176 collection already, so niceee.
I am still waiting for another but they said they are doing transfer per plug, not per account.

[_leras]

With several caveats currently though.

1. Only 17 of them are currently available.
2. They require ilok (and you know how that drives some crazy ).
3. Their installer is even worse than ikmm.. it installs everything. you buy one plugin and it installs all is it 300+ of them.
4. Mac only currently.
5. They are traditionally very expensive (with list prices of $200 and up, so when they put them on sale for $99 some people think God has just dropped manna from heaven..)

those are the current caveats I think.

rsp

oh... I think you missed a few:

6. some of the older GUIs are a bit on the small side
7. They are rock solid reliable
8. It might make you think all your other plug ins sound sh.te

[zvenx]

Definitely don't agree with 8 at all... I think their 1176, la2a and 175/176 are probably their best plugins ..

I think lots of other developers have caught up. fab filter, softube, fuse audio, ddmf etc.... But the beauty of native UAD is people who never had uad hardware can now test for themselves if this supposed dsp exceptionalism is fact or fiction.. I think it was real 10 years ago, not at all now, and as I have mentioned I have been using UAD's since 2004 and still use them today. But to think that in 2022 only they have hired or have access to the top dsp engineers well...

I did mention 6 in an earlier post but I do agree with 6...

rsp