Mercuriall released Spark

VST, AU, AAX, CLAP, etc. Plugin Virtual Effects Discussion
Post Reply New Topic
RELATED
PRODUCTS
Spark

Post

Baxterbrown wrote:
andy4trance wrote:There must be some sort of compatibility problem, it shouldnt push my CPU to 80%(even at no oversampling) while all other guitar amps stay below 20%.
I'm on windows 7, with focusrite audio interface.
Having the same issue here. What DAW(s) are people using that are having this problem? I use Studio One 3.3 and the CPU hit averages around 60%. On my laptop with Studio One it was at 80% or more. I tested Spark in Reaper and it worked just fine, 6 to 8%. That's why I'm wondering if this issue is just Studio One related, or if it's other DAWs as well. FWIW, I also own U530 and it works just fine. Thanks.
Are you using Studio One's Performance Meter to determine CPU? If yes, then DON'T. It's confusing as hell to new users, but my own testing, and confirmation from Presonus shows that Studio One's Performance meter will show you the maximum CPU load of the highest used core in Windows. So what does that mean?

1. Let's say you've got a quad core i7 with hyperthreading
2. Windows/Studio One sees 8 cores
3. You launch Studio One, you load a heavy VST plugin

Result: The plugin gets loaded to 1 of those 8 cores and uses up (let's say) 60% of that core. The other 7 cores are at 0%. Studio One will report 60% CPU usage in the Performance meter.

4. Now, launch Windows Task Manager -> go to the Performance tab

Result: CPU usage is MUCH lower than what Studio One reports.

If you're going to compare Reaper to Studio One, don't use the performance meter in either, look at what Windows is reporting. Reaper is still more efficient than S1, but S1's Performance Meter logic makes it look much worse than it really is.

Another thing you can do is see how many instances of Spark you can run in Studio One. If it eats up 60% CPU, that would mean you can only run 1 right? I think you'll be pleasantly surprised to find that you'll be able to run at least 8x as many. Why? Because as long as you put each Spark on a different track, they'll get loaded onto a different core. 60% per core, means you can run at least 8 instances on 8 separate tracks. If you loaded 2 instances on a single track though, you'd max out that core since each effect bin is limited to running on one core.

I confirmed all this with U-He's Repro. Studio One was showing 26% CPU usage on some patches (with the beta) per instance, but I was able to load way more than 4 tracks. How would that be possible? Because it's showing you the performance of 1 core at a time, and always the highest core, not your total CPU use.

Post

Funkybot's Evil Twin wrote:
Are you using Studio One's Performance Meter to determine CPU? If yes, then DON'T. It's confusing as hell to new users, but my own testing, and confirmation from Presonus shows that Studio One's Performance meter will show you the maximum CPU load of the highest used core in Windows. So what does that mean?
Thanks for the detailed reply. That's good information to remember. I'm familiar with Studio One's unusual CPU reporting, but wouldn't it all be relative? What I mean is, if Amplitube 4 runs at 20% CPU on my Studio One meter and Spark runs at 80%, then Spark is using 4x the processing. If that's the case, then maybe it's not a bug, but Spark just uses a lot more CPU than my other sims? Additionally though, the CPU meter jumps all over the place with Spark, which it doesn't do for any of my other plugins (including U530), which leads me to believe it's more than just slower performance on my machines.

Post

Baxterbrown wrote:
Funkybot's Evil Twin wrote:
Are you using Studio One's Performance Meter to determine CPU? If yes, then DON'T. It's confusing as hell to new users, but my own testing, and confirmation from Presonus shows that Studio One's Performance meter will show you the maximum CPU load of the highest used core in Windows. So what does that mean?
Thanks for the detailed reply. That's good information to remember. I'm familiar with Studio One's unusual CPU reporting, but wouldn't it all be relative? What I mean is, if Amplitube 4 runs at 20% CPU on my Studio One meter and Spark runs at 80%, then Spark is using 4x the processing. If that's the case, then maybe it's not a bug, but Spark just uses a lot more CPU than my other sims? Additionally though, the CPU meter jumps all over the place with Spark, which it doesn't do for any of my other plugins (including U530), which leads me to believe it's more than just slower performance on my machines.
On comparing AT4 CPU vs. Spark - yes, they're relative (just be careful that there's nothing else in your project when making that comparison). So yeah, it's possible that Mercurial's Spark with high oversampling is 4x as intensive as AT4. I haven't tried Spark yet so I can't confirm.

Regarding CPU jumping, is that showing up in Windows Performance monitor? I think being overly accurate is a "feature" of Studio One's performance monitor. I've seen spiking in S1's monitor that doesn't register on Windows CPU monitor.

FYI - if you're a member of the Presonus forum, I created a feature request for an enhanced performance monitor here: http://answers.presonus.com/12855/enhan ... e-accurate

Post

For me it doesn't seem a DAW problem but a Spark problem. This time it's not Studio One. Getting the same poor performance in Reaper.

Post

Funkybot's Evil Twin wrote: Regarding CPU jumping, is that showing up in Windows Performance monitor? I think being overly accurate is a "feature" of Studio One's performance monitor. I've seen spiking in S1's monitor that doesn't register on Windows CPU monitor.

FYI - if you're a member of the Presonus forum, I created a feature request for an enhanced performance monitor here: http://answers.presonus.com/12855/enhan ... e-accurate
The windows CPU monitor shows Studio One running at about 19% CPU while the Studio One CPU monitor jumps between 60 and 80% with only Spark running. Like you said, I can actually run multiple instances of Spark on separate tracks without issue. The other interesting thing though, is that it doesn't matter if I use any oversampling or not with Spark. It runs high and jumps on the CPU meters regardless. I don't know what's going on, but it is actually usable (no popping or lags or anything) - just weird visual behavior in the Studio One meter. Still don't know if it's a Studio One issue, Mercuriall issue, or both.

I also voted for your feature request on the Presonus forum. Hopefully it'll make the list for 3.4 or 3.5. Fingers crossed!

Post

andy4trance wrote:For me it doesn't seem a DAW problem but a Spark problem. This time it's not Studio One. Getting the same poor performance in Reaper.
Interesting. Glad to know I'm not the only one. Hopefully they'll come out with a quick update to fix things. It really does sound great.

Post

Definitely Mercuriall should clarify things. Hopefully it will be an easy fix.

Post

In reference to my previous post and observations: I was using Windows Performance Monitor (for CPU usage), and Cubase VST Performance monitor. WPM was holding steady within 2 or 3 % for each (corresponding) readout/setting in Spark. The VST monitor was running higher than for all my other amp sims, with more fluctuation and periodic spikes.

Again, has anybody noticed their noise gate not working? Working?? I haven't heard it do anything, no matter what the settings are.

Still loving this, now experimenting with cabs/mics.

Post

Twrogstudio wrote:No problem here in Studio One 3.3.
ditto
Say 'NO' to Clap

Post

ShawnMH wrote:I absolutely love this . . .
Couldn't agree more, really like it myself.

As to the Noise Gate, it works fine here, does what it's supposed to do.
Last edited by jinotsuh on Sat Dec 24, 2016 8:05 am, edited 1 time in total.
Say 'NO' to Clap

Post

You all know I'm a cpu ho.... :hihi:

If more than one person is reporting it on different hosts, then there is a problem. I'm guessing this is the demo? (kinda pricey to just buy but sounds pretty good in the examples) I'll give the real-world test maybe.

Post

Ok, unless you go to 4/8x OS, it seems to be about in-line with s-gear, sans front pedal (which I have to provide) When you OS that high, good luck with anything (other than mixing ;) ) :hihi:

Specs:

4770k, not o'c'ed
44.1k/128 buffer
FL studio (12.4x)
RME babyface pro*

That is a micro try so tomorrow I'll give it more work, but it doesn't seem to be seriously over the top CPU wise so far.

It's too early to tell sound wise, but I was really impressed for that short period of time :o Tomorrow is a better day to play with it and record with it to see.

Some notes:

-It's tough when the fade out is so quick. I get it, CP is a good thing, but it's a little rough when trying to fiddle the dials.
-I get some noise when moving the dials. Not sure if it's the plug of FL (FL isn't very good for that)
-MASSIVE volume issues when trying to check models. That's going to be a tough one.
-It's a bit steep price wise considering the past, then again, seems to sound really good overall.
-resale is important. I never know when I'm going to ditch the plugs for other things :lol:



*I've noticed better overall cpu performance switching to this when it came out. I get better latency and better sound w/o needing to go so low in buffer.

Post

Looks like I need 5 messages to make a long post...

Post

Hi guys/gals!

I am Slava from Mercuriall, dealing with public relations. Will try to appear here more often.

We need more structured/detailed information about the problem you experience with Spark. Afterwards, we'll analyze it to understand what exactly is wrong and if it is the Spark's issue.

Kindly add these details (those who experience issues; if you care, you can post details even if Spark runs smoothly):

1. Your OS
2. CPU type and speed
3. Overall RAM size and Free RAM size before you open the DAW
4. CPU load before you open the DAW
5. Your sound interface name and buffer settings/sampling frequency you run it at
6. Are you on HDD or SSD
7. What programs do you run in the background? Maybe we could find some common program that behaves badly with Spark.
Like antivirus (which one), cloud storage software, messaging apps, etc.
8. CPU load and Memory performance when Spark is in use. Please check it from Task Manager (not sure how it is called on Mac), not the DAW performance display. Screenshots are appreciated
9. Spark oversampling settings and what is turned on (gate/pedals/cab/FX)
10. Exact description of the problem (please don't go with "same as others")
11. Other comments/observations you may have.
12. Anything else I forgot to ask =)

Thanks a lot! I am sure we'll identify the issue with your kind help!

Post

Hello Slava,

in my case, I hear lots of crakles, and they go nuts when i try to tweak any parameter on the GUI.
Details:
1.Windows 7
2. Dual Core E8500, 3.2 Ghz
3. 8Gb
4.1-5%
5. Focusrite 2i2
6. 44.1k Sample rate, 3 ms buffer length (the RTL value I don't know for sure, Cubase is telling 13.8 output latency)
7. HDD
8. Aside Windows stuff and Video Card Drivers, nothing. You can see in the attach the CPU usage, it peaks at 40-45%, for one instance with no oversamling. As a note, u530 at 4X sampling is a breeze compared to Spark with no oversampling (U530 never takes CPU above 10-15%)
9. As above. Crackles go nuts when oversampling;
10. Spark is a CPU hog for my setup, it gives these regular wavy CPU usage rates, that at the lowest settings, no oversampling, no FX take the CPU to 40-50% at the peak. It may me more computationally intensive than U530 but soundwise I hear no such difference in quality as the CPU is sweating over it.

Thanks
You do not have the required permissions to view the files attached to this post.

Post Reply

Return to “Effects”