Ok, so my previous info for CPU/oversampling modes was based off of Task Manager's Performance Monitor (CPU tab), and showed CPU for everything running on this PC as a total. I also stated that I had 2 Edge tabs open and Kaspersky Internet Security 16 running for that.
So, to be more accurate than before, these are CPU/oversampling mode averages I observed from the Resource Monitor (CPU tab, Cubase process by itself/with Spark loaded):
Cubase Pro 8.5 by itself (no Spark): CPU: .90 - 1.10%
Spark Oversampling Modes observed averages (Default preset):
NO: 11 - 12%
2X: 11.8 - 12.5%
4X: 13.2 - 14%
8X: 16.8 - 17.3%
I didn't see any real difference in turning off Kaspersky AV or Wi-Fi, averages were near the same either way.
@ Barmaleus: I didn't have time to do screenshots and linking them today, so maybe in the near future. I had just a bit of time for this post & info. As for OC, I forgot to try turning it off, so it will have to be another time as well, it is as easy as a mouse-click to do. And, I may give Reaper a try too, I am happy with Cubase, but always been curious to see how it works. It IS known for being the DAW that is least demanding on a system, yes?
Mercuriall released Spark
-
- KVRist
- 222 posts since 24 Dec, 2016
@ShawnMH
I understand.
Spark Oversampling Modes observed averages (Default preset):
NO: 11 - 12%
2X: 11.8 - 12.5%
4X: 13.2 - 14%
8X: 16.8 - 17.3%
These values are what I would expect to see.
It seems to me the Cubase meter is showing the load for one CPU, not an average across all the CPUs. The results from the Task Manager's Performance Monitor (CPU tab) you posted are what we see in our tests as well.
We currently work on optimizing Spark, but definitely don't expect a 50% less CPU consumption =) It's not the lightest plugin indeed, as it models all electronic components found within the amp. This is the price for not using EQ matching and wave-shaping technologies.
As for the Gate concerns you have when using Fernandes Revolver - does it have a very low noise floor?
Re. Reaper - yup, it's very light on system resources indeed. Of course, no need to switch DAWs, but it is interesting to see the performance there.
I understand.
Spark Oversampling Modes observed averages (Default preset):
NO: 11 - 12%
2X: 11.8 - 12.5%
4X: 13.2 - 14%
8X: 16.8 - 17.3%
These values are what I would expect to see.
It seems to me the Cubase meter is showing the load for one CPU, not an average across all the CPUs. The results from the Task Manager's Performance Monitor (CPU tab) you posted are what we see in our tests as well.
We currently work on optimizing Spark, but definitely don't expect a 50% less CPU consumption =) It's not the lightest plugin indeed, as it models all electronic components found within the amp. This is the price for not using EQ matching and wave-shaping technologies.
As for the Gate concerns you have when using Fernandes Revolver - does it have a very low noise floor?
Re. Reaper - yup, it's very light on system resources indeed. Of course, no need to switch DAWs, but it is interesting to see the performance there.
Last edited by Barmaleus on Wed Dec 28, 2016 8:30 pm, edited 1 time in total.
-
- KVRist
- 215 posts since 11 Oct, 2012
A 20-25% improvement especially on the No oversampling setting + no crackling during tweaking, would make it usable for me too I guess.
Thanks
Thanks
-
- KVRist
- 177 posts since 3 Jun, 2016
The Fernandes Revolver (w/Sustainer pickups) is very low noise without the Sustainer engaged, a bit of low level noise when engaged, and fairly high output (EMG). The only time it can be really noisy is when the battery is depleted, and it's very easy to tell when that happens. Using Bias FX for quite a bit today, the built-in noise gate and pedal module gate both worked fine with the Revolver in all situations.Barmaleus wrote:@ShawnMH
We currently work on optimizing Spark, but definitely don't expect a 50% less CPU consumption =) It's not the lightest plugin indeed, as it models all electronic components found within the amp. This is the price for not using EQ matching and wave-shaping technologies.
As for the Gate concerns you have when using Fernandes Revolver - does it have a very low noise floor?
Re. Reaper - yup, it's very light on system resources indeed. Of course, no need to switch DAWs, but it is interesting to see the performance there.
I fiddled with input levels in Spark, and on the Zoom UAC-2 interface - measuring with Cubase and bx_meter.
In the midst of that, the gate worked only once in a while (with the same exact or near same range of parameters for when it wasn't working?!). More often, the middle and neck pickups had the gate working (when it did), and the bridge/noisegate combo came to life maybe twice.
Either time that it did work, the threshold was cranked for bridge, 3/4 or higher for neck and mid. All testing was with presets "8 ANRISEONS SOUND Metal" and "9 ANRISEONS SOUND 800 R". The Revolver simulates all 3 pickup/positions/sounds from the bridge pickup - by way of an EMG HZ, as the neck pickup is nothing BUT the Sustainer. Anyhow, I changed the batteries and it made no difference, and the old one was still quite juiced. A couple times, switching presets and back made it work.
As for the CPU juice Spark needs, I have no worries, it is usable for me no matter what as is. The noise gate is not any big worry to me, was originally just an observation, I generally try to make a preset with very low (or no) noise - to avoid having to be using a gate. With sound like this one has, I can see why it takes a bit more CPU. There is clearly more than the average amp simulation here - I KNOW this is state of the art Marshall sound in the box, and word WILL spread.
-
- KVRer
- 4 posts since 26 Dec, 2014
No CPU problem with my i7 4790k and Reaper (both VST and host 64bit), it gets 8% maximum, around 15% in stereo, 8x oversampling.
The plugin is outstanding IMHO, I own BIAS, Thermionik and S-Gear and for Marshall tones nothing can beat this in terms of sound and feel.
I whish Mercuriall team ports the effects, OD and cab sections to the U530, that would be fantastic.
The plugin is outstanding IMHO, I own BIAS, Thermionik and S-Gear and for Marshall tones nothing can beat this in terms of sound and feel.
I whish Mercuriall team ports the effects, OD and cab sections to the U530, that would be fantastic.
-
- KVRist
- 73 posts since 22 May, 2014
Still hopeful Mercuriall will come out with an update that fixes the CPU issue some of us are experiencing. Even though the CPU has been running at 60% or more per instance on my system I was able to demo the plugin long enough to put it on a couple guitar tracks and compare a current mix with BIAS, S-Gear, and Amplitube 4.
The good news: SPARK sounded the best. Thick and chunky on the palm mutes and sounded like a recorded amp, not a digital recreation like my other sims.
The bad news: I can't buy a plugin that doesn't work right on my system. I'll just wait and see, I guess.
The good news: SPARK sounded the best. Thick and chunky on the palm mutes and sounded like a recorded amp, not a digital recreation like my other sims.
The bad news: I can't buy a plugin that doesn't work right on my system. I'll just wait and see, I guess.
-
- KVRist
- 222 posts since 24 Dec, 2016
@Baxterbrown
1. Could you remind me what are your computer specifications?
2. Does the 60% CPU usage come from the DAW metering or the Task Manager?
3. You wrote earlier that there were no issues in Reaper, only in Studio One - is this correct?
4. Are you running a 32 or 64 bit plugin?
1. Could you remind me what are your computer specifications?
2. Does the 60% CPU usage come from the DAW metering or the Task Manager?
3. You wrote earlier that there were no issues in Reaper, only in Studio One - is this correct?
4. Are you running a 32 or 64 bit plugin?
-
- KVRist
- 215 posts since 11 Oct, 2012
Hi Slava,
Ca we hope for an update of Spark soon?
Thanks
Ca we hope for an update of Spark soon?
Thanks
-
- KVRist
- 222 posts since 24 Dec, 2016
@andy4trance
Hey!
Need to check on the status. I know the work on further optimization is ongoing.
Hey!
Need to check on the status. I know the work on further optimization is ongoing.
-
- KVRAF
- Topic Starter
- 4468 posts since 15 Nov, 2006 from Hell
Finally tried the demo yesterday.
The CPU usage is way too high, the thing crackles when dials are moved (to be fair, all of these are supposed to be fixed in next update), but my biggest issue with it is it's got the same weird usability all the other Mercuriall plugins have. Maybe it's just the novel modeling method, but i feel like the amp always "lags behind" any changes made to the GUI. For example, turning the OD pedal on and off always results in momentary silence followed by a gradual volume increase, and turning the dials, while doesn't result in silence, still feels like the changes are applied with a small delay. It's not latency either, as I was playing the thing live. It reminded me of Nebula, where there can be radically different plugins that all share usability traits common to all Nebula plugins.
That said, the sound was actually great, and it seemed to react to different input levels way better than Mercuriall's previous incarnations of JCM800 (probably as a result of built-in OD pedals forcing the devs to consider this particular aspect more closely). I totally loved the different Marshalls and how they were different yet unmistakably Marshall-like. However, i didn't feel like i needed Spark, as the sounds, while great, didn't really blow everything else i have out of the water. If i was looking for a good Marshall sim, i would totally pick this up, but as it stands, i have Thermionik, S-Gear and Amplitube 4, all of which have great Marshall (or Marshall-like) amps which are very hard to beat. I certainly don't need it at a $99/$120 price point.
So, in the end, i'll have to pass on this one. I would heartily recommend others to give this one a try though, especially if you're not satisfied with Marshall offerings from other developers!
The CPU usage is way too high, the thing crackles when dials are moved (to be fair, all of these are supposed to be fixed in next update), but my biggest issue with it is it's got the same weird usability all the other Mercuriall plugins have. Maybe it's just the novel modeling method, but i feel like the amp always "lags behind" any changes made to the GUI. For example, turning the OD pedal on and off always results in momentary silence followed by a gradual volume increase, and turning the dials, while doesn't result in silence, still feels like the changes are applied with a small delay. It's not latency either, as I was playing the thing live. It reminded me of Nebula, where there can be radically different plugins that all share usability traits common to all Nebula plugins.
That said, the sound was actually great, and it seemed to react to different input levels way better than Mercuriall's previous incarnations of JCM800 (probably as a result of built-in OD pedals forcing the devs to consider this particular aspect more closely). I totally loved the different Marshalls and how they were different yet unmistakably Marshall-like. However, i didn't feel like i needed Spark, as the sounds, while great, didn't really blow everything else i have out of the water. If i was looking for a good Marshall sim, i would totally pick this up, but as it stands, i have Thermionik, S-Gear and Amplitube 4, all of which have great Marshall (or Marshall-like) amps which are very hard to beat. I certainly don't need it at a $99/$120 price point.
So, in the end, i'll have to pass on this one. I would heartily recommend others to give this one a try though, especially if you're not satisfied with Marshall offerings from other developers!
I don't know what to write here that won't be censored, as I can only speak in profanity.
-
- KVRist
- 73 posts since 22 May, 2014
@Barmaleus, my replies to your questions are inline below. Thanks.
1. Could you remind me what are your computer specifications?
My specs: PC, 8 GB RAM, Phenom II x4 processor, 64 bit, Studio One Version 3.3
Side note - I also tried the plugin on my laptop, 8GB RAM, i7 4 core processor, same Studio One version, and it performed even worse - spike up to 100% and average running of 80% cpu meter
2. Does the 60% CPU usage come from the DAW metering or the Task Manager?
I was using the metering from my DAW, which as I understand it, shows only the core that is most taxed. That said, I use plenty of other heavy CPU plugins/instruments (BFD3, UHE Diva, and Acoustica Audio) and none of them taxed the system this hard or erratically.
3. You wrote earlier that there were no issues in Reaper, only in Studio One - is this correct?
That is correct.
4. Are you running a 32 or 64 bit plugin?
64 bit
1. Could you remind me what are your computer specifications?
My specs: PC, 8 GB RAM, Phenom II x4 processor, 64 bit, Studio One Version 3.3
Side note - I also tried the plugin on my laptop, 8GB RAM, i7 4 core processor, same Studio One version, and it performed even worse - spike up to 100% and average running of 80% cpu meter
2. Does the 60% CPU usage come from the DAW metering or the Task Manager?
I was using the metering from my DAW, which as I understand it, shows only the core that is most taxed. That said, I use plenty of other heavy CPU plugins/instruments (BFD3, UHE Diva, and Acoustica Audio) and none of them taxed the system this hard or erratically.
3. You wrote earlier that there were no issues in Reaper, only in Studio One - is this correct?
That is correct.
4. Are you running a 32 or 64 bit plugin?
64 bit
-
- KVRist
- 222 posts since 24 Dec, 2016
@Burillo
Thanks a lot for the detailed feedback, appreciated!
1. Which dials were crackling for you when moved?
There is no easy immediate fix. Each increment of the dials' position is studied by neural networks. For each increment there is a separate model of differential equations. When the "surfaces" of two neighboring models' solutions are not identical, the crackling may appear.
Basically, neural networks need to be trained again (and this physically takes a lot of time), but we need to know the exact plugin state to reproduce it. If you can make a GUI-less screenshot, that'd be amazing.
2. Have not heard about lagging so far. Need to check.
3. Yes, CPU usage will be above average, that's a trade-off for modeling the entire thing without any wave-shapers or EQ-matching.
@Baxterbrown
1.Could you tell me the exact models of the AMD and Intel processors you have? Unless it's a secret =)
2. To clarify - the correct way to check the CPU usage is by looking at the Task Manager. You should not be worrying about the DAW meters jumping around, especially in Studio One. Reaper's meters look much more robust to me.
3. To clarify - are you experiencing bad plugin performance (cracks and pops and similar) or just worried about high CPU usage? This part is not clear to me yet, sorry.
4. See #3 above in my reply to Burillo =)
Thanks a lot for the detailed feedback, appreciated!
1. Which dials were crackling for you when moved?
There is no easy immediate fix. Each increment of the dials' position is studied by neural networks. For each increment there is a separate model of differential equations. When the "surfaces" of two neighboring models' solutions are not identical, the crackling may appear.
Basically, neural networks need to be trained again (and this physically takes a lot of time), but we need to know the exact plugin state to reproduce it. If you can make a GUI-less screenshot, that'd be amazing.
2. Have not heard about lagging so far. Need to check.
3. Yes, CPU usage will be above average, that's a trade-off for modeling the entire thing without any wave-shapers or EQ-matching.
@Baxterbrown
1.Could you tell me the exact models of the AMD and Intel processors you have? Unless it's a secret =)
2. To clarify - the correct way to check the CPU usage is by looking at the Task Manager. You should not be worrying about the DAW meters jumping around, especially in Studio One. Reaper's meters look much more robust to me.
3. To clarify - are you experiencing bad plugin performance (cracks and pops and similar) or just worried about high CPU usage? This part is not clear to me yet, sorry.
4. See #3 above in my reply to Burillo =)
-
- KVRist
- 215 posts since 11 Oct, 2012
[quote="Barmaleus"]@Burillo
Thanks a lot for the detailed feedback, appreciated!
1. Which dials were crackling for you when moved?
There is no easy immediate fix. Each increment of the dials' position is studied by neural networks. For each increment there is a separate model of differential equations. When the "surfaces" of two neighboring models' solutions are not identical, the crackling may appear.
Basically, neural networks need to be trained again (and this physically takes a lot of time), but we need to know the exact plugin state to reproduce it. If you can make a GUI-less screenshot, that'd be amazing.
2. Have not heard about lagging so far. Need to check.
3. Yes, CPU usage will be above average, that's a trade-off for modeling the entire thing without any wave-shapers or EQ-matching.[/quote]
It's good to know there are devs trying very different paths in amp modeling, and indeed the method has a sensible impact on what we hear. Congrats for that!
Still, if the demand for your product will be limited to ones having only high end computers, the success of the tool can be just as much.
I still think that the crackles are due to CPU "abuse" during tweaking, but not to the actual algos or model overlap. If it was the case, everybody would have reported the noises during tweak. What I hear though is that crackles are correlated with weaker CPU and/or less optimized DAWs.
Like others around, still hope for further optimizations of the model.
Thanks a lot for the detailed feedback, appreciated!
1. Which dials were crackling for you when moved?
There is no easy immediate fix. Each increment of the dials' position is studied by neural networks. For each increment there is a separate model of differential equations. When the "surfaces" of two neighboring models' solutions are not identical, the crackling may appear.
Basically, neural networks need to be trained again (and this physically takes a lot of time), but we need to know the exact plugin state to reproduce it. If you can make a GUI-less screenshot, that'd be amazing.
2. Have not heard about lagging so far. Need to check.
3. Yes, CPU usage will be above average, that's a trade-off for modeling the entire thing without any wave-shapers or EQ-matching.[/quote]
It's good to know there are devs trying very different paths in amp modeling, and indeed the method has a sensible impact on what we hear. Congrats for that!
Still, if the demand for your product will be limited to ones having only high end computers, the success of the tool can be just as much.
I still think that the crackles are due to CPU "abuse" during tweaking, but not to the actual algos or model overlap. If it was the case, everybody would have reported the noises during tweak. What I hear though is that crackles are correlated with weaker CPU and/or less optimized DAWs.
Like others around, still hope for further optimizations of the model.
-
- KVRist
- 73 posts since 22 May, 2014
@Barmaleus - Answers inline below:
1.Could you tell me the exact models of the AMD and Intel processors you have? Unless it's a secret =)
Yes, normally this is top-secret, but for you I'll make an exception
HP Desktop PC Pavilion p6751c (QP688AAR#ABA) Phenom II X4 830 (2.8 GHz) 8GB 1 TB HDD ATI Radeon HD 4200 Windows 10 Home Premium 64-Bit
2. To clarify - the correct way to check the CPU usage is by looking at the Task Manager. You should not be worrying about the DAW meters jumping around, especially in Studio One. Reaper's meters look much more robust to me.
I understand that S1's CPU meter is "unique," but I also know that one instance of Blue Cat's Destructor runs at 7% CPU, and BIAS amp runs at 6%, and TH3 runs at 5% on average on my computer through the S1 CPU meter. One instance of SPARK on my desktop runs at 60% on a blank project with no other plugins. To me that's unusual. But if you're saying that with SPARK's current programming I should expect CPU usage to be 10x my other plugins and that's normal behavior, then I guess I have to accept that.
3. To clarify - are you experiencing bad plugin performance (cracks and pops and similar) or just worried about high CPU usage? This part is not clear to me yet, sorry.
Both, but it depends on how many plugins I'm running. On a blank project with other plugins I can run a couple instances of SPARK with no click/pops. But if I incorporate more plugins/tracks, then it will overload and start clicking/popping.
Thanks for looking into all this. I only care because SPARK sounds so good.
1.Could you tell me the exact models of the AMD and Intel processors you have? Unless it's a secret =)
Yes, normally this is top-secret, but for you I'll make an exception
HP Desktop PC Pavilion p6751c (QP688AAR#ABA) Phenom II X4 830 (2.8 GHz) 8GB 1 TB HDD ATI Radeon HD 4200 Windows 10 Home Premium 64-Bit2. To clarify - the correct way to check the CPU usage is by looking at the Task Manager. You should not be worrying about the DAW meters jumping around, especially in Studio One. Reaper's meters look much more robust to me.
I understand that S1's CPU meter is "unique," but I also know that one instance of Blue Cat's Destructor runs at 7% CPU, and BIAS amp runs at 6%, and TH3 runs at 5% on average on my computer through the S1 CPU meter. One instance of SPARK on my desktop runs at 60% on a blank project with no other plugins. To me that's unusual. But if you're saying that with SPARK's current programming I should expect CPU usage to be 10x my other plugins and that's normal behavior, then I guess I have to accept that.
3. To clarify - are you experiencing bad plugin performance (cracks and pops and similar) or just worried about high CPU usage? This part is not clear to me yet, sorry.
Both, but it depends on how many plugins I'm running. On a blank project with other plugins I can run a couple instances of SPARK with no click/pops. But if I incorporate more plugins/tracks, then it will overload and start clicking/popping.
Thanks for looking into all this. I only care because SPARK sounds so good.