Yeah, well the world has gotten really more exciting after reading your comment… like 0.00000000000000001 %.

 

Wavetable synthesizers often require more CPU power than other synthesizers because of the complexity of their sound generation and processing. Here are the main reasons:

1. Complex Waveforms and Interpolation
Wavetable synthesizers work with a multitude of waveforms, which are often interpolated in real time to create smooth transitions between individual waves. These real-time calculations are computationally intensive.

2. Polyphony and Number of Voices
The number of voices that can be played simultaneously (polyphony) has a significant impact on CPU utilization. Each additional voice increases the processing requirements, especially when multiple oscillators or unison modes are active.

3. Modulation and Effects
Wavetable synthesizers often offer extensive modulation options and effects such as FM (frequency modulation), which require additional computations. Filters, especially more complex types such as cytomic filters, also place a greater strain on the CPU.

4. Real-time calculation of unison voices
When unison modes are enabled, each voice is multiplied, requiring exponentially more processing power. For example, a chord with multiple oscillators and unison voices can quickly generate hundreds of voices, each of which must be calculated individually.

5. Long release times and complex envelopes
Long release times and non-linear envelopes also increase CPU load because they keep voices active for extended periods of time.
In summary, wavetable synthesizers require more processing power than simpler forms of synthesis such as subtractive or FM synthesis due to their versatility and ability to create complex soundscapes.

 

There is plenty to like about Serum 2, but fake sounding string lines that sound less realistic than what was available 22 years ago doesn't do it for me.

I really don't understand all the fan-boy-ism about Serum 2, it has some nice features but none of it is game changing.

 

It is what it is: a sophisticated virtual wavetable synthesizer. One of the best on the market. It's a modern musical instrument; you can create your own sounds with it or use presets for your music production. People who have absolutely no idea about the subject matter are always the first to criticize something they don't understand at all.

The people who know nothing about music are the ones always talking about it. Nat King Cole

 

Last edited: Apr 1, 2025 at 6:55 PM

Yeah there is no doubt that Serum is one of the best and most present sounding synths out there, but I find it can be too bright maybe for certain applications. It's unmatched for Bass and things but I like to use other softer synths for pads and less bright leads/plucks etc

 

Did you go under global tab and lock it to Good quality? Each preset can override the current setting, unless you hit the lock icon. Even the Serum v1 presets are set to High. Below that is S1 compatibility mode. With it enabled, Serum doesn't use any of the new dsp features; which are probably the most cpu demanding. Maybe that helps. You can enable "set to Ultra Quality when rendering", if you think you will forget when bouncing.

The Clips Player and Arp being able to send Midi Out makes me "wishlist" a copy job of the BandMate functions in EzKeys2. It already seems like one, with some Cthulhu thrown in.

 

Pianoteq is the outlier.

 

Nope, sorry. There's a problem with Serum 2 on PC/Cubase 14 Pro. I have a recent setup with 64GB of RAM. I can load tons of Serum 1, but Serum 2 completely crashes. I love it because I'm a pro, I see what it could produce, but it overloads my CPU, even with a 4096 sample buffer. It's frustrating. Steve my man, Help!

 

You better go back to v1!
Audient USB Audio - 44,1 kHz - 256 Samples | AMD Ryzen 7 5800X3D - 16 GB RAM | SERUM 10 % CPU

Spoiler: Appendix E: Optimizing Serum

Appendix E: Optimizing Serum

This appendix describes a series of easy-to-adopt approaches to optimizing your sound design results in
Serum.

Exploring CPU Optimization

Serum is designed to optimize audio processing and CPU performance. However, there are two key
approaches that you can incorporate to help your sounds achieve the very best performance and
maintain the highest-quality in sound design:
• Manage unison effectively
• Prioritize FX bus use over per-voice effects
This section describes these two techniques.

Managing Unison

Unison is a powerful tool for enhancing the depth and richness of a sound by layering multiple voices
slightly detuned or panned, creating a fuller, more powerful tonal presence. However, overuse of unison
can potentially lead to quality and performance issues.

Consider doing the following:
• Keep unison counts low
Using more than three to seven unisons per oscillator is often unnecessary and can potentially
negatively impact efficiency. This is because higher unison counts not only significantly increase
CPU usage, but can also introduce phasing issues, potentially degrading your sound quality.

• Use chorus instead of unison
Instead of stacking unisons to create a thick, chorused sound, use a dedicated FX bus with a
chorus effect. This approach offers two principal advantages. First, by apply the effect once instead
of processing it for every voice, this approach is considerably more CPU friendly.

Second, this provides greater flexibility by allowing easier tweaking and layering of effects without
duplicating processing effort.

Prioritizing FX Bus Use

Serum processes effects applied to an oscillator per voice (for each note played). While this is powerful
for creating rich, detailed sound designs, it can also dramatically increase CPU usage since the effects
are calculated for every active voice.

You can optimize performance by using an FX bus for shared processing whenever possible. This means
routing your sources to an FX bus and apply effects there.

This method offers the following advantages:
• Reduces CPU strain by processing the effects once for all voices combined
• Maintains cohesion, ensuring consistent and unified effects across all routed sources
Using an FX bus works best when applying a reverb, delay, EQ, or similar effect that doesn’t require
voice-level manipulation.

When to Use Per-Voice Effects
You should consider using effects at the oscillator level only if you need specific sound interactions
per voice, such as uniquely modulated or warped effects that cannot easily be achieved through bus
processing.

Source --> Serum 2 User Guide

 

Last edited: Apr 1, 2025 at 4:03 PM

I had to edit Serum2Prefs.json and set the following to get it to default to "Good" oversampling:

"Default Oversampling Level": 1,

 

Yes all of their instrument modelling is incredible if you consider how tiny their modelling files are.
Their modelling of the vibraphone is quite exceptional.