u-he SATIN as Reason Rack Extension?

Official support for: u-he.com

u-he SATIN as Reason Rack Extension?

yes
36
54%
no
31
46%
 
Total votes: 67

RELATED
PRODUCTS
Satin$149.00Buy

Post

TheoM wrote:Is it that uhbik analog code is so high quality that other RE simply aren't as high quality? I for example never imagined a flanger using so much CPU (but uhbik f does sound ridiculously warm).
All Uhbiks are oversampled between 2x and 8x internally to achieve a certain level of quality. To make this usable in the realm of VST/AU their dsp parts as well as the upsampling/downsampling sections are coded with a method that entirely compiles to vectorized SSE on Intel, AltiVec on PowerPC (even though we don't support PPC anymore). As this requires assembly level development, and as assembly level is not available for RE development, we had to make the same code compile to standard serial instructions for RE.

Now, in theory, a compiler can restructure this code easily and generate vectorized SSE optimized code. I doubt though that the compiler used by Propellerheads does this automatically, and so it would appear that the result will be up to 4 times slower.

As for mathematical functions… all common programming languages come with a set of mathematical functions such as sin(), cos(), exp(), log(). These are generally considered too slow for realtime processing. Hence we have written our own versions of those, which are approximations. They are good enough for tuning oscillators and stuff, and extremely fast. I presume that if the RE SDK 2.0 provides implementations of those that's certainly good for some developers, but it won't help in our case.

- Urs

Post

[DELETED]

Post

TheoM wrote:I get that.. that's fair enough i suppose..but i doubt any reason user would have voted no as the RE is what's important to them. I am sure in return if there was an RE developer that did a poll at PUF whether to go VST, they'd all vote no to keep the dev concentrating on RE
yeah, pretty understandable i would say. nothing to wonder about. subtractor as vsti would be great btw. :)
Whoever wants music instead of noise, joy instead of pleasure, soul instead of gold, creative work instead of business, passion instead of foolery, finds no home in this trivial world of ours.

Post

TheoM wrote:I am just not sure where you get "up to 4x" from.
Vectorized instructions are doing stuff in batches of 4. Essentially doing four calculations for the price of one. So if no vectorized instructions are used, the same thing needs 4 times the CPU time to execute. AFAIK.

Post

TheoM wrote:I am just not sure where you get "up to 4x" from.
Well, SSE can perform the same operation on 4 samples at the same time. This is how we use it. If those operations have to be done one after the other it might take up to 4 times longer. Hence my guess.

Thing is, we have no means to inspect the resulting code. Hence all I can do is guess.

It's a downside of an otherwise phantastic concept.

Post

EvilDragon beat me to it :)

Post

[DELETED]

Post

TheoM wrote:SO basically, you've done all you can from your end for now for uhbik, and that's the best the performance will get as an RE for now.
I truely believe that we've done what we could.

Post

Thanks for the update, Urs. We'll keep patiently waiting for Satin RE. I agree that it'll be a huge hit as well.

Cheers!
Win 10 | Ableton Live 11 Suite | Reason 13 | i7 3770 @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro| Akai MPC Live 3 & Akai Force | Roland System 8 | Roland TB-3 | Roland MX-1 | Dreadbox Typhon | Korg Minilogue XD

Post

Thanks for the detailed response, another Reason user here eagerly awaiting Satin RE.

As all my needs are being met by rack extensions I've been gradually selling off my plugins. I much prefer working within Reason.

Right now the only thing I go outside of Reason for is tape/console emulations. Just happen to be using Slate Digital at the moment, but as soon as Satin RE drops I'm thinking this will change!

Agree with Theo that Satin RE will be unrivalled, right now there are some sweet EQs and compressors appearing in the RE store so this would really be the icing on the cake.

Thanks again for posting and hope all goes well with the bug fixes.

Post

murnau wrote:
Tronam wrote:
spktkpkt wrote:It is okay to vote "no", but it would be interesting to know why they are have voted "no". I guess they are knowing that "Satin" will look and sound totally great in "Reason" and are unhappy that they don't have it. :hihi:
Yeah, I'd be curious too. :)
it's easy, i voted "no" because i don't use reason and want that urs concentrate on the important things. :)
This is a clear statement. :)

Post

Clear and selfish, but I like the honesty. :tu:

Post

Urs wrote:We have a Satin RE with nearly finished UI, but it currently doesn't work. Crashes Reason.

So we're on a global bug hunt (reflected in the Latest Builds thread) and then we hope that Satin RE will run properly. If not, then we'll try to figure it out. Unfortunately I'm the only developer at u-he dealing with this, but I'm also in charge with a lot of other things. Thus unreleased software such as Satin RE has a lower priority than, say, fixing bugs in released software.

The changes in the latest RE SDK have no influence on how our stuff works. We've always used faster maths than whatever the standard libraries provide. The new SDK still does not allow us to code on assembly level. Thus there won't be a Diva RE anytime soon.

The new UI capabilities are cool, but they are not compatible with our UI code out of the box. A complex control such as a Zebra MSEG needs to be coded from scratch, scrolling panels still seem impossible. Thankfully neither Uhbiks nor Satin need this, but a Zebra RE is still out of the picture (as well as Filterscape and MFM2 I'm afraid).

Furthermore, cables are still confined to the backside, which rules out an ACE RE or a Bazille RE.

Lastly, the UI design process is tedious and can't be done for free. This rules out our freeware offerings.

Thus, so far, Uhbik's and Satin are the only u-he plug-ins that can go RE.

Our main problem with RE is human resources. We've hired 3 more people within half a year, but getting everything up to speed takes time. That's our main bottleneck, but I hope I'll find the time to do this soon.

- Urs
As much as I love existing u-he stuff, I can't help feeling that ports of existing synths is an idea that should be abandoned for the time being. I'm feeling a bit ambivalent because as much as I love Zebra, Diva, and ACE, I think that an unique u-he RE would be better for both u-he and the RE format.

I don't have any hopes that any such synth will happen in the near future, or even long term for that matter, but I hope that once you are ready for a synth, that you will consider a non-port.

My 2c anyway, best wishes.

Post

eXode wrote:As much as I love existing u-he stuff, I can't help feeling that ports of existing synths is an idea that should be abandoned for the time being. I'm feeling a bit ambivalent because as much as I love Zebra, Diva, and ACE, I think that an unique u-he RE would be better for both u-he and the RE format.
I totally agree with this. Synapse Audio Antidote RE is a good example.

Post

eXode wrote:As much as I love existing u-he stuff, I can't help feeling that ports of existing synths is an idea that should be abandoned for the time being. I'm feeling a bit ambivalent because as much as I love Zebra, Diva, and ACE, I think that an unique u-he RE would be better for both u-he and the RE format.

I don't have any hopes that any such synth will happen in the near future, or even long term for that matter, but I hope that once you are ready for a synth, that you will consider a non-port.

My 2c anyway, best wishes.
spktkpkt wrote:I totally agree with this. Synapse Audio Antidote RE is a good example.
I completely agree.
Win 10 | Ableton Live 11 Suite | Reason 13 | i7 3770 @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro| Akai MPC Live 3 & Akai Force | Roland System 8 | Roland TB-3 | Roland MX-1 | Dreadbox Typhon | Korg Minilogue XD

Locked

Return to “u-he”