Sylenth1 v3.050 Beta

[Elektronisch]

I played same chords, 2 stacks of jp8000 oscillators in Diva with multicore On and same with Sylenth1, applied same modulation, just with filters bypassed. In Fl it was ranging ~ 17 percent. Diva on on fast ~ 15 percent, draft 12-13 percent.

Would you mind sharing the two patches you used in the test please ? I'd like to try it here. Not saying you're wrong just curious what the numbers would be here.

Also are you using FL Studio's meter or Window's Task Manager ?

Still do not believe?

[chk071]

I don't think that is a fair comparison. The JP-8k oscillators are the most CPU saving of the oscillators in Diva, and, bypassing the filter is also a bit like cheating, because especially the filters are what's hogging the CPU in Diva.

I bypassed them in Sylenth1 not diva, you cant do that in diva.

Ah... which filter were you using in Diva? The most CPU intensive is the Moog ladder filter for me.

[Yorrrrrr]

Is this thing Retina compatible? It looks blurry in my MacBook Pro (2018).

This will be fixed in the next update!

Great!. It's really important.

[Elektronisch]

I don't think that is a fair comparison. The JP-8k oscillators are the most CPU saving of the oscillators in Diva, and, bypassing the filter is also a bit like cheating, because especially the filters are what's hogging the CPU in Diva.

I bypassed them in Sylenth1 not diva, you cant do that in diva.

Ah... which filter were you using in Diva? The most CPU intensive is the Moog ladder filter for me.

Cascade, why does it matter? the fact is Diva does take less CPU as you can see in Studio One 3.5 (with 2 stacks and multicore on)

[Lennard]

Still do not believe?

No. Diva uses 7 voices unison, your preset on sylenth1 uses 8 on all oscs. Also you're using 16 voices polyphony and 8 on Diva. And you're using Delay, Reverb and EQ on Sylenth1 and no effects on Diva.

With these settings the CPU measures pretty much the same on my computer. After making them equal I get 10% on Sylenth1 vs 13% on Diva. After turning off multicore on Diva it changes to 10% vs 18%:


Multicore processing may look nice on the CPU when you use 1 plugin, but there is very little advantage (if any at all) when using a lot of plugins, as the DAW will divide the plugin threads over all CPU cores anyway. And using lots of plugins is exactly what you'll be doing when making music.

Also, when switching from 'draft' mode to 'fast' mode in Diva, CPU use jumps up to 30%. In 'divine' mode it's even 60%.

[Lennard]

As for the oversampling, at the default 1x oversampling setting, Sylenth1 already uses 2x oversampling in the filters, 4x oversampling in the distortion effect and 2x oversampling in the chorus and phaser effects.

So at 4x oversampling, you're oversampling the distortion 16x for example. The option to oversample the entire synth was just added so that you can render your project to wav at insane quality. For realtime playback it is not really needed at all (although some presets will sound slightly sweeter at 2x oversampling already). It is recommended to set the realtime oversampling at 1x, or 2x if you prefer working in very high quality and are using a big CPU. And then you can set the rendering oversampling factor at 4, 8 or even 16x, so that it will export your track to wav in extremely high quality.

[Elektronisch]

When checking the presets in studio one i had turned off the effects Anyways, but glad you gave a bit more of explaination whats happening under the hood when the official update will come out ?

[Lennard]

Well the current beta hardly contains any problems at all, so you can already use it as official (beta will be removed from the filename, that's pretty much it). But there will be a new beta soon, v3.052 which will be compatible with Retina displays, will have some improvements for the AAX version and a fix for an offline installation problem. This will become the official release version shortly after.

[Echoes in the Attic]

Is this going to be one of those super long beta periods, or is the official release planned to come out soon?

[Lennard]

Although it is still named beta, it is the exact same version that will be moved to the official download spot. You can safely use it as no serious issues have been found. A new update will be ready soon, which will need only a short bit of testing before it can be moved to official release.

[Echoes in the Attic]

oh does that include the fixed blurriness mentioned in the thread?

Edit: Just installed the beta and I'm not a retina but the GUI looks fine when resized for me.

[mholloway]

Studio One's CPU meters are notoriously janky. I wouldn't treat anything you see there as proof of anything at all. Just clicking on the screen in Studio One makes the meters jump. This has all been discussed extensively in various old threads (usually ones in which hardcore S1 users get super righteous and shouty at those of us who express our concerns). So, yeah. If you wanna compare Diva to Sylenth in terms of CPU, go ahead, but I highly advise testing with something other than Studio One's internal CPU metering.

[chk071]

Well... i haven't really heard many complaints in the direction of Sylenth1 utilising more CPU than Diva, so, i doubt there's much call for action anyway.

[Echoes in the Attic]

Studio One's CPU meters are notoriously janky. I wouldn't treat anything you see there as proof of anything at all. Just clicking on the screen in Studio One makes the meters jump. This has all been discussed extensively in various old threads (usually ones in which hardcore S1 users get super righteous and shouty at those of us who express our concerns). So, yeah. If you wanna compare Diva to Sylenth in terms of CPU, go ahead, but I highly advise testing with something other than Studio One's internal CPU metering.

Best way to test performance is usually to put as many copies running simultaneously as possible until failure, so it tells you proper system usage without taking into account throttling and stuff. Easiest to do with multicore off and just stack em up on one track (with a rack or something) if you're comparing two things because the performance on a single core is all you need to know. Or have 4 tracks on a quadcore playing the same thing and keep duplicating those until failure. Compare number of plugin instances before crackles for the two plugins.

[Echoes in the Attic]

Swing is great to have by the way. I would have really liked this ten years ago though.