The No.1 Website for Pro Audio
 All  This Thread  Reviews  Gear Database  Gear for sale     Latest  Trending
UAD X coming soon ... ?
Old 2nd September 2018
  #31
Gear Maniac
 

Thanks!

bb




Quote:
Originally Posted by ~ufo~ View Post
I don't know, what does GOOGLE stand for?

Field Programmable Gate Array..

That doesn't tell you much more, does it?

An FPGA is a chip used in many interfaces, usually to route audio (implement software input mixing functionality etc).
Recently, they've also been used more to run effects to process audio signals.

They seem to be gaining popularity in favour of DSP chips.

If I were UAD I wouldn't put the R&D into another platform unless they are absolutely sure it will continue to have significant R&D behind it itself, in order to keep up with the development of the Intel architecture.
But obviously, I'm not UA and they've been extremely stubborn with sticking with DSPs so far.

The fact that they are hiring an FPGA engineer, does not necessarily mean they will use FPGAs for processing though. I wouldn't be surprised if FPGAs were already in their interfaces, running the IO. Anyone in the know?

I'll be surprised if they go FPGA for processing, but if they decide to stay with hardware based processing, a move to FPGA makes sense. I just wonder how much R&D it costs to port their stuff to this platform, I'm not even sure it's possible to do so with 100% null. FPGA is a very different beast than DSP as I understand. It basically acts like a circuit board with components (which obviously makes it interesting to mimic hardware), it's not just porting your code to a different language/platform, I'm afraid.
It's possible, but I would guess it's quite the investment to move UAD to FPGA.
They'd better be sure...
Old 2nd September 2018
  #32
Lives for gear
 
Marando's Avatar
Quote:
Originally Posted by MGA View Post
I think the answer is in the Hardware Test Engineer's job description currently advertised on the UA website.

Universal Audio Jobs

FPGA is coming to UAD.
They use FPGA in their Apollo rack units since day one. Not for running plugins though.
2
Share
Old 2nd September 2018
  #33
Lives for gear
Are we sure X just doesn't mean a new UAD with ten DSPs? Have they surpassed eight in any one product yet?
Old 2nd September 2018
  #34
Lives for gear
 
~ufo~'s Avatar
Quote:
Originally Posted by Funkybot View Post
Are we sure X just doesn't mean a new UAD with ten DSPs? Have they surpassed eight in any one product yet?
Oh ****.... that would be the most lame upgrade ever! That’s just too depressing to even consider. I guess if it’s 10x dual core it could be acceptable on the short term providing dual core means it could share DSP load between cores and not be limited to one core max per plugin.
Old 2nd September 2018
  #35
Always entertaining reading know-it-all musicians talk about digital hardware engineering like they know what it is.

Marando is correct. FPGAs are not typically as efficient or best option at calculating Algortihms such as UAD as DSPs are, but that doesn’t mean they aren’t good for other things.

Quote:
Originally Posted by Marando View Post
They use FPGA in their Apollo rack units since day one. Not for running plugins though.
3
Share
Old 2nd September 2018
  #36
Lives for gear
Quote:
Originally Posted by ~ufo~ View Post
Oh ****.... that would be the most lame upgrade ever! That’s just too depressing to even consider. I guess if it’s 10x dual core it could be acceptable on the short term providing dual core means it could share DSP load between cores and not be limited to one core max per plugin.
If it went from 8 to 10 DSPs, that would equate to a 25% increase in available power. So not exciting by any means, but I wouldn't rule it out. I don't know anything about DSPs though, are there dual core SHARCs (or whatever UA is using)? If there are, dual core chips would be a big upgrade.

I wish they went native, even if the card acted as a dongle, but I don't find that likely. Would love to be wrong on that though. I've been staying away from new UAD plugins for years, and would consider them again in a second if they didn't run on the hardware.
Old 2nd September 2018
  #37
Lives for gear
 
~ufo~'s Avatar
Like I said. Dual core would only be a significant upgrade if there was core sharing and plugins could exceed one core. Otherwise... meh.
I mean, to me dual core is meh akready, so I guess it would be double meh.
Old 2nd September 2018
  #38
Lives for gear
 
Quetz's Avatar
I don't think there are any inherent problems with base code when using FPGA chips instead of ASIC chips.

What I took away from that Wikipedia article is that they're both ICs, and they both rely on Hardware Description Languages.
If you read it all, you'd have seen that what someone said above about FPGA chips being inefficient was old hat, and the real nugget is here:

Quote:
Originally Posted by Wikipedia
Historically, FPGAs have been slower, less energy efficient and generally achieved less functionality than their fixed ASIC counterparts.
An older study had shown that designs implemented on FPGAs need on average 40 times as much area, draw 12 times as much dynamic power, and run at one third the speed of corresponding ASIC implementations.

More recently, FPGAs such as the Xilinx Virtex-7 or the Altera Stratix 5 have come to rival corresponding ASIC and ASSP solutions by providing significantly reduced power usage, increased speed, lower materials cost, minimal implementation real-estate, and increased possibilities for re-configuration 'on-the-fly'.
Where previously a design may have included 6 to 10 ASICs, the same design can now be achieved using only one FPGA.
This sounds like one UAD X card with 4 FPGA chips could be the equivalent of four Octos in UAD-2 terms.
For little to no increase in price.
That would piss a few people off though, I reckon, to say the least. Just as going native would, basically wiping out the value of every dsp card in existence overnight.
So it has to be hardware/dongle, it's a successful system business-wise, but now they do need to step up with some real processing grunt, as this is now the sticking point for most people.
Four Octos on one card for the same price, or even less - That is the kind of hardware upgrade that will make people invest in the system. Long term.
Old 3rd September 2018
  #39
Being an engineer that has designed circuit boards and code for both DSP based projects and FPGA projects (and projects that have both, which is probably the most common these days) ....you guys really with all due respect have no idea what you are talking about. I'm not sure what else to say

I would be surprised if they abandoned their SHARC based platform. This would require a big effort to port the code to another processor as they are probably using many Analog Devices specific DSP Structures and Architecture. This is equivalent to something like when Apple switched from PowerPC to Intel. It doesn't just happen willy nilly. (but who knows)

Comparing an FPGA to a DSP doesn't really make too much sense to be honest. Although FPGAs are going more the ways of floating-point, that isn't their advantage and why they rose to becoming popular. It is the parallelism, custom re-programmable logic, and clock speed with low-latency that make them attractive for certain designs.

I've been out of the game for a few years now doing other things, but I just want to shed some light when people start talking about FPGA vs DSP like it is some Spec sheet from the latest iMac or something. That isn't it at all.



Quote:
Originally Posted by Quetz View Post
I don't think there are any inherent problems with base code when using FPGA chips instead of ASIC chips.

What I took away from that Wikipedia article is that they're both ICs, and they both rely on Hardware Description Languages.
If you read it all, you'd have seen that what someone said above about FPGA chips being inefficient was old hat, and the real nugget is here:



This sounds like one UAD X card with 4 FPGA chips could be the equivalent of four Octos in UAD-2 terms.
For little to no increase in price.
That would piss a few people off though, I reckon, to say the least. Just as going native would, basically wiping out the value of every dsp card in existence overnight.
So it has to be hardware/dongle, it's a successful system business-wise, but now they do need to step up with some real processing grunt, as this is now the sticking point for most people.
Four Octos on one card for the same price, or even less - That is the kind of hardware upgrade that will make people invest in the system. Long term.
Old 3rd September 2018
  #40
Lives for gear
 

Quote:
Originally Posted by contramark View Post
I would be surprised if they abandoned their SHARC based platform. This would require a big effort to port the code to another processor as they are probably using many Analog Devices specific DSP Structures and Architecture. This is equivalent to something like when Apple switched from PowerPC to Intel. It doesn't just happen willy nilly. (but who knows)
UA would not have to abandon SHARCs if they find a way to interconnect a new unit with older ones. Would make sense in regard to latest add-on cards for older Apollos.
Old 3rd September 2018
  #41
Lives for gear
 
clonewar's Avatar
 

You guys are reading too much into the mention of FPGA in the job ad. Every UAD-2 card and Apollo already use FPGA, just obviously not to execute the plugin DSP code.

Here's an FAQ about older UAD-2 cards being incompatible with new Mac Pros because they use a previous generation FPGA. The issue is with the expanded PCIe addressing scheme, which means that at least one of the things the cards use the FPGA for is PCIe communication.
Why are the UAD-2 SOLO, DUO, and QUAD PCIe cards incompatible with the new Mac Pro? – Universal Audio Support Home
Old 3rd September 2018
  #42
Lives for gear
 
~ufo~'s Avatar
Quote:
Originally Posted by contramark View Post
I would be surprised if they abandoned their SHARC based platform. This would require a big effort to port the code to another processor as they are probably using many Analog Devices specific DSP Structures and Architecture. This is equivalent to something like when Apple switched from PowerPC to Intel. It doesn't just happen willy nilly. (but who knows)
Exactly. Which is why I think that if they spend that R&D into a port to a new platform, they'd better be damn sure that platform itself has enough R&D to carry it forward with steady performance bumps to allow them to continue to develop the UAD line with any sort of bang for buck.

Right now they've been in a territory for years where the quality of their plugins has never been up for dispute but they've been writing more DSP power hungry plugins while they are stuck on the same old sharcs.
The instance chart on them is embarrassing.
UA are left with no option but put their fingers in their ears and chant weird things like 'RIGHT NOW THEY DO WHAT WE NEED THEM TO DO' and the fanboys are going like 'WELL BEING ABLE TO RUN 16 NEVE CHANNELS ON MY OCTO IS STILL SO MUCH CHEAPER THAN BUYING 16 NEVE CHANNELS, SO....LUCKY US".
It's bizarre.

If the sharc development has stalled (which sees the case, certainly compared to Intel), I guess they could just use the same chip in a massive price drop and offer 16 and 32 chip cards.
But that would still leave them with a major problem, right: They cannot write 'bgger' plugins... Plugins that exceed one chip.
They have certainly been flirting with that as a few plugins almost eat up a whole chip. It stands to reason that this lack of DSP headroom, has hindered them already and will hinder them in the future.
Being a DSP/FPGA engineer, perhaps you could answer whether a dual core sharc could allow for plugins that exceed one one core?

I'd say they either stick with sharc and just up the hip count, for the same or less money or they should port to intel.
They could even port to run on the Waves intel based DSP servers.

Porting to Intel is probably also a substantial R&D load, but it would at least provide a platform with a steady track record, that's not going anywhere, with a powerful level of bang for buck and with a regular performance increase.

We'll see. They'll either sticking to their current platform with all its problems or making a final port, I'd say.
I doubt it will be FPGA for processing. It might make sense for some newer developers like Antelope, but porting an existing line of plugins to run on FPGA and then potentially developing new plugins to run on both.... Ugh.
You gotta be real certain it's gonna be worth the hassle...
Old 3rd September 2018
  #43
Lives for gear
 
GeneHall's Avatar
Regarding UAD SHARC and plugs being dsp hungry, I have 12+2( pcie/c +TwinDUO) and never regretted what I paid for those nor have I never run out of power during a mix.
Seems the gripes come when folks have like a quad or less and expect to have access to ridiculous amounts of plugin instances.
For me, the SHARCS work just as they are supposed and as I was told they would work when I bought into the platform.
I have many buyer regrets in my time with pro audio, but never once had a single moments regret about my UAD2 purchases.
Expensive to some, unattainable for many but well worth any cost or physical limitations.
If they ran native, I can only imagine the increased difficulties that might create.
I know the limitations are there and they are real, they just don't affect me and I'm certain that I'm not alone in this.
Only thing I can really ask of UA, based on my own experiences would be if they would include a clean up program with the updates software, so I didn't have to do it manually, and give me the option of only installing the x64 plugins I use. It would be cool if we could trade or sell plugins we don't use much with other UAD users, too.
They could call that an upgrade, hell.. call it X

2
Share
Old 3rd September 2018
  #44
Lives for gear
 
GeneHall's Avatar
Quote:
Originally Posted by ~ufo~ View Post


They could even port to run on the Waves intel based DSP servers.
Ugh...please don't encourage this!
1
Share
Old 3rd September 2018
  #45
Here for the gear
 

Hi all,

I wonder where the informations + preview pic about version X come from. The only source seems to be our GS newsletter, couldn‘t find anything on uaudio channels.
So my question who brought it in?

As the current sw version is 9.6 the next logical step at first glance is just a sw version 10 soon … but of course I‘d be surprised if UA won‘t use the magic number X for something special
1
Share
Old 3rd September 2018
  #46
Lives for gear
 
~ufo~'s Avatar
Quote:
Originally Posted by GeneHall View Post
Ugh...please don't encourage this!
Haha, I mean to run on Waves DSP servers for those who really need/want it (like in the live market), ADDITIONAL to running natively.


Regarding the previous poster.
Sure there's many people happy with their UAD and for some the plugin count is ample. That's all fine. No-one's telling you to not be happy with your system.
It's obviously a popular system and the plugins are ace.

The hardware side is underpowered for today's standards, however, to the point where the processor offload of a duo is laughable, a quad is negligible.
You really need an octo to have a substantial processor offload AT LOWEST buffer.
This cannot be emphasised enough.
The plugins are great, the benefits of it being DPS based over native are for most users, negligible. This means that for most users, their instance count is hindered/limited BECAUSE they are running them on DSPs.

I challenge any UAD user to download demoes of native versions of plugins that are the same on UAD and native, like VSM-3, Maag EQ, VSC-2.

Put your system at the highest sample rate and lowest buffer your system runs stably at and test how many native instances you can run stably vs how many UAD instances. Say you cap your CPU meter at 70% to allow for some headroom.

Chances are that even at lowest buffer, the native count will beat your UAD count, often embarrassingly so unless you have a really old computer with a weak processor, or a particularly hefty UAD setup (16 chips or more).

My 2014 2.8 quad MBP can easily run 2 octos worth at 96k 64 samples, and my 2010 3.33 6-core MP can easily run 3 octos worth at the same setting.
Those are powerful, but ageing computers.
Expect modern similarly specced machines to do 30-50% better.
Then compare a 1k modern processor (10 core!) to a 1k Octo.
The difference will be ridiculous AT LOWEST BUFFER.
Come mixdown, when you up the buffer, you'll have even more power. Not so with UAD, it's fixed. This used to be an advantage over native, it's started to be a disadvantage.

It used to be that you had to raise the buffer to max to get close to DSP performance, but it's been the opposite for quite a while. You 'have' to lower the buffer of your native system to get close to DSP performance.
It is true that DSPs have the advantage to behave very predictably and dependably and that's a great advantage, but considering the aforementioned, it's simply mostly irrelevant to most users.
There are situations where only DSP systems will cut it, but these situations are becoming less and less common. So much so that DSPs can be more of a hindrance than a help.

All i'm saying is: keep using your DSP chips to help you, for some people they will especially be handy on the way in, but set those puppies free so that you can harness the power of your CPUs too.
Use your chips for rock solid finite instance count, use your CPU for scalable varying plugin count.

That's how most plugins run and it's really not much of a problem at all.
Most people just need to watch out a little with using crazy heavy plugins at lowest buffers. If you really need to run those at lowest buffer: I have nothing against the OPTION of running some things on DSPs.

But having to schlep around a UAD box while you have a quad in your laptop....
Come on man... it's 2018... this is absolutely unnecessary and it has been for years.
I'm not a DSP hater, I've used them for many years. They have upsides and downsides and at some point you have to make up the balance.
For years now that balance has been: No way, not worth it to me.
which is a shame, because I and with me scores of others producers and engineers would gladly buy UAD plugins if they weren't forcibly tied to DSPs.

Let's all enjoy those UAD plugins, some on hybrid, some on native. Everyone happy.
Right, I've preached my sermon again, I'll hopefully leave it at that.
We'll see what Universal Audio comes up with.
They'll possibly continue with DSP and this hideous cycle will repeat itself :D

Hope we all get what we want.
Giday!
Old 3rd September 2018
  #47
Lives for gear
 
Quint's Avatar
Quote:
Originally Posted by ~ufo~ View Post
Exactly. Which is why I think that if they spend that R&D into a port to a new platform, they'd better be damn sure that platform itself has enough R&D to carry it forward with steady performance bumps to allow them to continue to develop the UAD line with any sort of bang for buck.

Right now they've been in a territory for years where the quality of their plugins has never been up for dispute but they've been writing more DSP power hungry plugins while they are stuck on the same old sharcs.
The instance chart on them is embarrassing.
UA are left with no option but put their fingers in their ears and chant weird things like 'RIGHT NOW THEY DO WHAT WE NEED THEM TO DO' and the fanboys are going like 'WELL BEING ABLE TO RUN 16 NEVE CHANNELS ON MY OCTO IS STILL SO MUCH CHEAPER THAN BUYING 16 NEVE CHANNELS, SO....LUCKY US".
It's bizarre.

If the sharc development has stalled (which sees the case, certainly compared to Intel), I guess they could just use the same chip in a massive price drop and offer 16 and 32 chip cards.
But that would still leave them with a major problem, right: They cannot write 'bgger' plugins... Plugins that exceed one chip.
They have certainly been flirting with that as a few plugins almost eat up a whole chip. It stands to reason that this lack of DSP headroom, has hindered them already and will hinder them in the future.
Being a DSP/FPGA engineer, perhaps you could answer whether a dual core sharc could allow for plugins that exceed one one core?

I'd say they either stick with sharc and just up the hip count, for the same or less money or they should port to intel.
They could even port to run on the Waves intel based DSP servers.

Porting to Intel is probably also a substantial R&D load, but it would at least provide a platform with a steady track record, that's not going anywhere, with a powerful level of bang for buck and with a regular performance increase.

We'll see. They'll either sticking to their current platform with all its problems or making a final port, I'd say.
I doubt it will be FPGA for processing. It might make sense for some newer developers like Antelope, but porting an existing line of plugins to run on FPGA and then potentially developing new plugins to run on both.... Ugh.
You gotta be real certain it's gonna be worth the hassle...
There's what UAD may WANT to do and there's what UAD may HAVE to do. It obviously will take a lot of extra work for them to take the next step, whether that next step be newer/more Sharcs, FPGA, native or whatever else they might cook up. But what is becoming more indisputable as time goes by is that UAD is reaching a point where their current DSP power and DSP hungry plugins have created a situation where UAD will HAVE to do something.

So I wouldn't discount the move to FPGA solely because it might be difficult. All options for UAD are probably difficult at this point. UAD just has to pick the best option at this point. Who knows?

Last edited by Quint; 3rd September 2018 at 06:45 PM..
Old 3rd September 2018
  #48
Old 3rd September 2018
  #49
Gear Addict
current UAD software is 9.6 - seems almost certain that UA is hinting at a “UAD 10” software update, hence the X.

my guess is probably an updated version of Console as well as new plugs
2
Share
Old 3rd September 2018
  #50
Old 3rd September 2018
  #51
Lives for gear
Quote:
Originally Posted by citrusonic View Post
current UAD software is 9.6 - seems almost certain that UA is hinting at a “UAD 10” software update, hence the X.

my guess is probably an updated version of Console as well as new plugs
I think this is the most likely explanation or some 10-DSP hardware.
Old 4th September 2018
  #52
Here for the gear
 

Thanx, it wasn’t online yet when the GS newsletter was sent.

IMHO def a new software (... more than just added plugins) but my guess is that a version 10 / X is way too alluring to present a new hardware concept too. Otherwise it would be 9.7 etc until there are also HW news or new concepts.

Let’s sit back, wait and see


Quote:
Originally Posted by Screwup View Post
Old 4th September 2018
  #53
Here for the gear
 

Quote:
Originally Posted by mucmusic View Post
Thanx, it wasn’t online yet when the GS newsletter was sent.

IMHO def a new software (... more than just added plugins) but my guess is that a version 10 / X is way too alluring to present a new hardware concept too. Otherwise it would be 9.7 etc until there are also HW news or new concepts.

Let’s sit back, wait and see

And twitter ...
Universal Audio (@UAudio) | Twitter

Would have been a little more fun if each social media platform had a new hint ...
Old 4th September 2018
  #54
Lives for gear
 
Jim Rosebrook's Avatar
 

Control Surface??

Just a wild guess....

Control Surface.
Old 4th September 2018
  #55
Lives for gear
 

Quote:
Originally Posted by Jim Rosebrook View Post
Just a wild guess....

Control Surface.
What for? For UA HD with an X instead?
Old 4th September 2018
  #56
Lives for gear
 
DrSax's Avatar
VST3 support would be a logical and soon to be necessity for UAD. I’m guessing “X” is v10 of the software. I think it’s more likely we’d see new hardware announcements at NAMM in January
Old 4th September 2018
  #57
Gear Addict
 
sirthought's Avatar
The teaser does NOT say UAD. It says Universal Audio X. I don't think it's UAD specific.

We've been told by UA folks it's the letter "X" and NOT 10. So, I don't think it's version 10 of anything, at least not as the main feature.

TOP FIVE guesses:
1. X refers to expansion. A larger Satellite or a rack unit for Sharc chips. An UA hardware controller, a la Control One, would be pretty sweet, though.

2. X refers to some sort of cross communication protocol that's a hybrid. While current UA interfaces lack modern protocols competitors have, namely AV Bridge (AVB), last year they did release UAD-2 Live Rack with MADI. Maybe this product will allow for greater expansion of this type of usage for both studio and stage.

3. Last year UA released the OX box for guitar players. I could see them continuing down this path with a X box for a keyboard controller, or maybe an actual keyboard controller with OX-like DSP processing for old-school synths. They could partner with someone like Arturia to have their soft synths be used live or in the studio without the need of a computer or plugin on site.

4. X refers to 'X marks the spot.' Maybe UA X is something you hit...or stomp, like a foot pedal that can load certain amp sims or plugin qualities.

5. X refers to their own brand of microphone like the Townsend Labs Sphere, Antelope, Slate, etc. It's Microphone X.
2
Share
Old 4th September 2018
  #58
Quote:
Originally Posted by ~ufo~ View Post
Being a DSP/FPGA engineer, perhaps you could answer whether a dual core sharc could allow for plugins that exceed one one core?

I'd say they either stick with sharc and just up the hip count, for the same or less money or they should port to intel.
They could even port to run on the Waves intel based DSP servers.

Porting to Intel is probably also a substantial R&D load, but it would at least provide a platform with a steady track record, that's not going anywhere, with a powerful level of bang for buck and with a regular performance increase.
When you are trying to do essentially real-time intensive calculations (such as in Apollos) Intel processors may not even be a viable option. They don't have the low-latency architecture that something like a SHARC can get you. (but that gap is changing more everyday)

I don't think a Dual Core will help the situation for plugins that exceed the core limit. ( But I could be wrong, haven't worked with that specific SHARC device before.) It seems the interface to the DMA is the same for 2 cores on 1 Chip vs. 2 Cores each in an individual package, thus not improving things that much for this specific application.
Old 4th September 2018
  #59
Lives for gear
 
Progmatic-Studios's Avatar
Quote:
Originally Posted by sirthought View Post
The teaser does NOT say UAD. It says Universal Audio X. I don't think it's UAD specific.
there are a lot of UAD plugin screenshots in the teaser and nothing else so I think it is about UAD
2
Share
Old 4th September 2018
  #60
Gear Addict
 
sirthought's Avatar
Quote:
Originally Posted by Progmatic-Studios View Post
there are a lot of UAD plugin screenshots in the teaser and nothing else so I think it is about UAD
I get that, but if you think about it, they wouldn't have an image to show of anything new.
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump