KVR Audio

SGA · Post by **SGA** » Sun Feb 01, 2015 4:19 pm

VariKusBrainZ wrote:Is the new commercial MKII demo really fully functional?

What are the new 'major enhancements'

I went ahead and bought it anyway being a No Brainer price, was just wondering.

I really like the free one, is great for adding that analog vibe and beefing up the low end

Thank you for your purchase! Hope you like SGA1566 MkII.
Briefly, enhancements are:
- performance improvements (we shaved off about 6-10% of CPU usage as measured on core i7 processor)
- Added MID CPU mode. You can save up to ~30% of the CPU utilization when compared to the HIGH CPU mode. Although MID CPU mode is strictly speaking less accurate than the HIGH CPU mode most of the time you won't be able to hear the difference.
- EQ can be taken out of the signal path completely so it does not affect the phase
- Additional oversampling filter. The additional filter has much sharper cutoff and preserves more of the frequencies higher than 15kHz.
- 8x oversampling that you can use for the final rendering. Good if you crank up SGA1566 MkII to the max. with a hot signal full of mid and treble frequencies.

You can read more about all the features and differences between CPU modes and oversampling filters in the user manual for SGA1566 MkII.

Ultrasonic · Post by **Ultrasonic** » Sun Feb 01, 2015 4:59 pm

@Ultrasonic
All you have to delete is SGA1566.dll. The installer is there so that you can accept the license and to create uninstall. SGA1566 uninstall completely removes it from the system.

I have no problem with the installer. I prefer it...but if you think you can go around deleting program files (not you in particular - any person) , or files within program files - including .dll files, and there wont be any consequences, well all i can figure is that youve been lucky. Maybe ive been UNlucky, but Its one surefire way to screw up an OS.

I can see how an installer would be a better option than a .dll. What i cant understand is how a .dll file could be any better option than an installer? I see more risk.

ObsoleteAcc99 · Post by **ObsoleteAcc99** » Sun Feb 01, 2015 5:25 pm

[DELETED]

Naenyn · Post by **Naenyn** » Sun Feb 01, 2015 5:54 pm

This little guy sounds pretty great! After a quick demo, it does seem to make things sound warmer/richer. Nice job!

SGA · Post by **SGA** » Sun Feb 01, 2015 6:29 pm

TheoM wrote:IMO this is honestly one of the best plugins of this type I have ever heard. The reason I can't use it as I would like to is the CPU on my now old 2.2ghz i7 quad sandy bridge macbook pro. It takes almost an entire thread per instance, so i could get 7 instances of this over 7 threads (Logic always reserves the last core for live inputs/rewire/gui etc) and that's absolutely all my computer could do. No other effects. No instruments whatsoever. Just 7 audio tracks with one instance of this each.

If you have the cpu power, i can not recommend this enough. It really DOES add some kind of "mojo"/warmth, in a pleasant way on almost anything it goes on. For me I will stick to radiator which is similar but uses a tenth the cpu if that. But I just wanted to give this plugin it's props.. if i had the *choice*, I would actually choose this over radiator sound wise by a nose. When i eventually get another machine I can see myself using this alot, and will donate then of course. A question, does the donate box go away if one donates?

First, thanks for the recommendation.
Here are some CPU % utilization numbers for SGA1566 MkII and SGA1566 as measured on 2.6 GHz quad i7 ivy bridge macbook pro running windows 7 64 bit.
Both plug-ins were the only plug-in running at the time. Signal was stereo white noise fed through the line in port on the macbook pro. Processing was stereo realtime. All controls other than OVERSAMPLING and CPU all were left at default with the exception of EQ POSITION on SGA1566 MkII which was set to POST to match the default EQ POSITION for SGA1566.

Numbers in the two tables are percent CPU utilization of one "thread" as a function of the CPU Mode and Oversampling

SGA1566 MkII

Code: Select all

CPU  |   Oversampling
mode | 1x  |  2x |  4x | 8x
-----|-----|-----|-----|-------
LOW  | 2.6 | 3.6 | 5.6 | 10
MID  | 13  | 26  | 50  | 97
HIGH | 20  | 34  | 63  | >100

the MID CPU mode uses about 20-25% less CPU than the HIGH CPU mode (i.e. low CPU mode ~ 0.75 * HIGH)

SGA1566

Code: Select all

CPU  |  Oversampling
mode | 1x | 2x | 4x
-----|----|----|-------
LOW  | 7  | 8  | 10
HIGH | 27 | 42 | 74

SGA1566 MkII HIGH CPU mode uses about 15-25% less CPU than the HIGH CPU mode with SGA1566.

EvilDragon · Post by **EvilDragon** » Sun Feb 01, 2015 6:45 pm

SGA wrote:
chokehold wrote:Deleted the plugin. Won't bother you anymore, sorry to have taken up your time.
I am really sorry you had bad experience with SGA1566.

It doesn't matter if hundreds of beta testers didn't notice the issue. If even one person encounters the issue when the plugin is out in the wild, you should do the extra mile and spend some effort to try and fix it for that person. Even more so if there's more than one person encountering issues (as it has been reported in this thread, it seems). This was very badly handled from your side IMHO, and everything reads from your response to chokehold.

You should improve your potential customer relations. As it stands, seeing how you responded to chokehold, I'm not even interested in trying out your plugin, no matter how good it might be. You showed unwillingness to go that extra mile, which is oh so important when customer support is concerned.

chokehold wrote:I never suggested the bad experience I had was with the plugin.

b4serenity · Post by **b4serenity** » Sun Feb 01, 2015 7:55 pm

nice work on the plug-in...keep 'em coming.

ObsoleteAcc99 · Post by **ObsoleteAcc99** » Mon Feb 02, 2015 1:00 pm

[DELETED]

audiosabre · Post by **audiosabre** » Mon Feb 02, 2015 3:00 pm

I'd like to make this thread a little more balanced

Maybe I'm spoiled, or too quick to judge, but I really don't like it. It breaks up in an unpleasant way, and the CPU is outrageous. In a project, CPU jumped from 12% CPU to +70%.

Replaced it with SoundToys Radiator (which is pretty much the same control set/idea), and boom, only 7-8% CPU, less boomy bottom end, pleasing distortion breakup, etc.

I guess you can't please everyone. Thanks anyway

SGA · Post by **SGA** » Mon Feb 02, 2015 3:11 pm

TheoM wrote: Hi, you are using windows even though on mac hardware in that example.. and windows throttles A LOT less than mac in high performance mode (or can be disabled completely with throttlestop).

Also, you are talking real actual cpu usage figures.. And that is correct.. the actual cpu usage on my 2.2 might be 35% real cpu usage, but that does not mean logic can run 3 on a core. Logic's meter is very accurate to what *logic* can do and when IT will crap out, and it never even closely correlates to the actual cpu usage. That's because DAW dsp usage is a different thing and relies on many factors and the limit is determined also by many factors./

That said, I will test the commercial mark 2 and see how I go! Cheers

Forgot to say in my post that the sampling frequency was 48 kHz with a 128 samples buffer (~4.7 ms latency).
The reason for giving you numbers from Windows is because I was logged into Windows. While doing testing I set windows to the High Perf mode. The High Perf mode gives more consistent results than the Balanced mode. The numbers in the table are the "actual" CPU numbers as Reaper reports them in the Perf Meter (RT CPU). I checked that number against Process Hacker2. Process Hacker consistently reported a lower number on the most heavily loaded thread with Reaper. For SGA1566 MkII 4x oversampling in HIGH CPU mode Reaper reports ~63% RT CPU whereas Process Hacker shows 40-50% one thread load. I could also see in the Process Hacker at least three more threads belonging to Reaper each using a few percent.

As far as I recall, Mac (with 10.9.5) numbers were pretty close. It could be that Mac's core audio drivers are much better than ASIO4ALL I used under Windows. CPU frequency was roughly the same under Mac and Windows; you can monitor it with some Intel app on Mac and Intel Gadget on Windows.

SGA · Post by **SGA** » Mon Feb 02, 2015 3:18 pm

audiosabre wrote: I guess you can't please everyone. Thanks anyway

Sadly, you cannot please everyone. You are welcome anyway.

Tp3 · Post by **Tp3** » Tue Feb 03, 2015 6:43 pm

audiosabre wrote:Maybe I'm spoiled, or too quick to judge, but I really don't like it.

That's fine.

audiosabre wrote:It breaks up in an unpleasant way

Sorry, but that simply isn't true (in my experience).

Just properly gain stage it, and it WILL sing beautifully. simple as that (exactly as with with Slates tape, for instance). as with most processor of this type.

audiosabre wrote:and the CPU is outrageous.

Unfortunately, I suspect that it is true, and I suspect that older CPUs suffer more.

On my 9550, CPU is indeed outrageous (40-50%) but on the other hand, on my internet (not optimized) i3 (i3, yes ?...) it is certainly bearable (I can only assume that on i5s and newer i7s, CPU is not really an issue)

audiosabre wrote:less boomy bottom end, pleasing distortion breakup, etc.

Don't let misuse "color" your opinion. try to gain stage it with plugs like Perception, and differences would be night and day.

Ivan_C · Post by **Ivan_C** » Wed Feb 04, 2015 2:54 pm

Sounds great !

I just wanted to report there are some denormalization issues to address. When I stop the playing, the CPU jauge increases a little on my computer

electro · Post by **electro** » Wed Feb 04, 2015 5:14 pm

has anyone compared SGA with Nebula preamps?

Tp3 · Post by **Tp3** » Wed Feb 04, 2015 6:58 pm

electro wrote:has anyone compared SGA with Nebula preamps?

yes, I have (to stradipad and API preamp)

Very hard to tell, but apart from Henry Olonga's program, I preferred SGA. it is "fuller" and more "in your face".

YMMV, though

Shattered Glass Audio releases SGA1566 - free, circuit simulation powered preamp plug-in