Scamp sounds sweet but it's a bit cpu hungry.
-
- KVRAF
- 2935 posts since 14 Dec, 2003 from Edinburgh
Am I the only one who doesn't think this is too bad a CPU usage on such a slow machine if they even tell you it uses quite a lot?
-
- KVRAF
- 37518 posts since 14 Sep, 2002 from In teh net
They did but it took a while and several revisions to get there. Maybe over time AQ will do the same.ttoz wrote:but yeah discodsp managed to heavily optimize discovery, etc etc..i don't believe it can't be optimized...sorry.
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Hey everyone. Thanks for pointing this thread out to me bleebsen, I had missed it so far.
A few points:
1: SCAMP (and its little brother Naive LPF) were designed for sound quality first, so they are definitely on the CPU heavy side compared to most everything else. But this is also how it gets a more convincing analog sound than most everything else.
2: If you're only processing a mono or stereo track, make sure you're not running SCAMP in 6 channel mode. In some hosts SCAMP sometimes runs on 6 channels even though your host may only have 2 channels to process (Wavelab, Ableton, some others...) If you're experiencing what seems to be way too much CPU load, then try bypassing channels 3 to 6. If this helps, then your host was connecting to SCAMP as 6 channel. (SCAMP will indicate if a channel is disconnected when you hit the corresponding bypass button.) You can limit the maximum number of channels by editing the SCAMP.ini file (which is automatically created in your vstplugins directory the first time you run SCAMP) and setting NUMCHANNELS 2. Same thing with Naive LPF, though it lacks the bypass buttons so you'll have to tweak the .ini file to check. If you're not using SCAMP on multichannel signals anyway then this won't limit anything.
2a: Try SCAMP's QUALITY switch set to Low. That should roughly half the required CPU power (and decrease the frequency response by an octave but it'll still sound pretty good, particularly in the lower cutoff frequencies.)
3. I've found some further optimizations during Poly-Ana's development which will be rolled into SCAMP v 1.1. Upgrades within the same major version number are free, so current SCAMP owners can expect a fairly significant speed boost sometime in the coming months. Can't promise a date yet as I'm deep in Poly-Ana at the moment, but will get back to speeding up SCAMP as soon as Poly's happy.
4: Yes there is some funky stuff happening with resonance at the high range of cutoff frequencies. Basically what's happening internally is that resonance has to be reduced to zero at the top end of the cutoff range. What you're hearing is the loudness of the entire signal getting a bit of boost as the resonance is faded out at the top end of the cutoff (resonance takes energy from the rest of the signal). Note that real world filters did/do this too. I could tweak this so it's a bit less noticable, but I also know some are using it to effect so I won't be introducing any major changes to the behavior/sound in version 1.x of SCAMP. For example, I find there's some good amp-sim type sounds possible by exploiting this artifact. So I'd recommend if that's bothering you to just avoid the top end of the cutoff range (cutoff within this filter goes WAY beyond your project's sample rate) and/or reduce the amount of resonance you're using. With resonance at zero you won't hear this at all.
Hope this is some help. And thanks SCAMP Evangelists for spreading the word about its unique sound!
A few points:
1: SCAMP (and its little brother Naive LPF) were designed for sound quality first, so they are definitely on the CPU heavy side compared to most everything else. But this is also how it gets a more convincing analog sound than most everything else.
2: If you're only processing a mono or stereo track, make sure you're not running SCAMP in 6 channel mode. In some hosts SCAMP sometimes runs on 6 channels even though your host may only have 2 channels to process (Wavelab, Ableton, some others...) If you're experiencing what seems to be way too much CPU load, then try bypassing channels 3 to 6. If this helps, then your host was connecting to SCAMP as 6 channel. (SCAMP will indicate if a channel is disconnected when you hit the corresponding bypass button.) You can limit the maximum number of channels by editing the SCAMP.ini file (which is automatically created in your vstplugins directory the first time you run SCAMP) and setting NUMCHANNELS 2. Same thing with Naive LPF, though it lacks the bypass buttons so you'll have to tweak the .ini file to check. If you're not using SCAMP on multichannel signals anyway then this won't limit anything.
2a: Try SCAMP's QUALITY switch set to Low. That should roughly half the required CPU power (and decrease the frequency response by an octave but it'll still sound pretty good, particularly in the lower cutoff frequencies.)
3. I've found some further optimizations during Poly-Ana's development which will be rolled into SCAMP v 1.1. Upgrades within the same major version number are free, so current SCAMP owners can expect a fairly significant speed boost sometime in the coming months. Can't promise a date yet as I'm deep in Poly-Ana at the moment, but will get back to speeding up SCAMP as soon as Poly's happy.
4: Yes there is some funky stuff happening with resonance at the high range of cutoff frequencies. Basically what's happening internally is that resonance has to be reduced to zero at the top end of the cutoff range. What you're hearing is the loudness of the entire signal getting a bit of boost as the resonance is faded out at the top end of the cutoff (resonance takes energy from the rest of the signal). Note that real world filters did/do this too. I could tweak this so it's a bit less noticable, but I also know some are using it to effect so I won't be introducing any major changes to the behavior/sound in version 1.x of SCAMP. For example, I find there's some good amp-sim type sounds possible by exploiting this artifact. So I'd recommend if that's bothering you to just avoid the top end of the cutoff range (cutoff within this filter goes WAY beyond your project's sample rate) and/or reduce the amount of resonance you're using. With resonance at zero you won't hear this at all.
Hope this is some help. And thanks SCAMP Evangelists for spreading the word about its unique sound!
-
- KVRAF
- 3617 posts since 26 Sep, 2003 from Bradford - The Armpit of Britain
Thanks for the info AQ - bypassing channels 3-6 in ableton Live on my system lowered the cpu useage considerably - down to around 11% HQ & 7-8% Economy.
Still on the high side for me personally, but a damn site better
Still on the high side for me personally, but a damn site better
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
There you go! So you were actually running the equivalent of 3 stereo SCAMPs, or 6 monos. (If you never work on surround tracks then you might as well limit it to NUMCHANNELS 2 in the .ini file. Will save you having to bypass those channels every time you load the plug-in)
Sorry about this issue but the various hosts all have different ways of negotiating channel connections. Unfortunately it's a very fuzzy part of the spec.
Sorry about this issue but the various hosts all have different ways of negotiating channel connections. Unfortunately it's a very fuzzy part of the spec.
-
- KVRAF
- 3617 posts since 26 Sep, 2003 from Bradford - The Armpit of Britain
My fault really, I should have read up more on the 3 stereo channels thing.
It may be a good idea for you to highlight that on your site somehow, so users of the affected hosts are immediately aware.
Many casual browsers/potential purchasers may just stick the demo on a track, notice the comparatively large cpu useage & immediately delete the demo with no real further exploration (like I did
), which is a disservice to yourself & the plug in.
It may be a good idea for you to highlight that on your site somehow, so users of the affected hosts are immediately aware.
Many casual browsers/potential purchasers may just stick the demo on a track, notice the comparatively large cpu useage & immediately delete the demo with no real further exploration (like I did
), which is a disservice to yourself & the plug in.-
- KVRian
- 690 posts since 31 May, 2002 from chez moi
ttoz, have you also tried Volcano and Frantic filter? What did you think of them? I love the 2d handle for cutoff/resonance on the volcano gui.
s
s
-
- Banned
- Topic Starter
- 22457 posts since 5 Sep, 2001
[DELETED]
-
- Banned
- Topic Starter
- 22457 posts since 5 Sep, 2001
[DELETED]
-
- Banned
- Topic Starter
- 22457 posts since 5 Sep, 2001
[DELETED]
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Can you not patch it as a send instead of an insert to get a wet/dry mix?ttoz wrote:WOW! What an improvement changing the ini file! Running my dothan at full speed (1.8ghz), I am at 13 scamps in HI quality mode, stably, adding another will make the system overload.
So let's say it works out to roughly 7.5% cpu per instance, a HELL of a lot better than close to 30%!
Admiral, what we need is an onboard wet/dry slider.
If you promise upcoming future optimizations, as I still would like it to be lower, and wet/dry slider, I will buy now.
I do promise upcoming optimizations, but it may be another few months before I can get to a ver 1.1 of SCAMP. But it will come. Promise.
As far as a wet dry mix knob... hmmm... I won't promise it yet but will seriously consider it. Can't think of any big reasons not to but still, reserving the right to change my mind on that one.
Let me know if you can't work around it somehow in your host. Glad your CPU issues are OK now! SCAMP was a bit of an experiment in attempting multi-channel support across all the hosts, but enough users are getting caught by this that I think the next rev will have to report clearly at startup on how many channels it's processing. Maybe those bypass buttons should become lights, to indicate the process is running on that channel.
And what's T2? Tracktion? Can you describe the click issue some more? How can I replicate it?
Thanks!
-
- KVRist
- 71 posts since 26 Feb, 2005
What ini file did you change? Just curious. I need to optimize my PC.ttoz wrote:WOW! What an improvement changing the ini file! Running my dothan at full speed (1.8ghz), I am at 13 scamps in HI quality mode, stably, adding another will make the system overload.
-
- KVRist
- 135 posts since 31 Mar, 2005
AQ....if Poly-Ana sounds as wonderful as Scamp, expect a LOT more love from the community.
I can't wait for its release!
I can't wait for its release!
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Not a "PC optimization", we're just talking about the .ini files for SCAMP and NaiveLPF. If you're not using SCAMP or Naive for more than stereo processing you might as well change NUMCHANNELS 6 to NUMCHANNELS 2 in their .ini file.frank001 wrote:What ini file did you change? Just curious. I need to optimize my PC.ttoz wrote:WOW! What an improvement changing the ini file! Running my dothan at full speed (1.8ghz), I am at 13 scamps in HI quality mode, stably, adding another will make the system overload.
S'all.
-
- KVRAF
- 3615 posts since 28 Jan, 2006 from Phoenix, AZ
admiral quality GUI annoys the hell out of me, that's all I know. I already have FilterBank2. You know what though, filterbank2 is more expensive. I'm not complaining, just making a point. I love filterbank2.