I set up a comparison between several different 24 dB/oct LPFs. They are:
1. CA Stock Filter
2. CA Ladder Filter
3. MRB Laboratory VCF
4. VM2500 Filtamp 1006
5. VM900 LPF 904 a
6. Nrgzr78 S100 Dual VCF
I recorded the session and uploaded the video to my YouTube channel.
The relative levels of the voice and VM output are quite unbalanced and the speech gets drowned out at times - sorry about that.
24 dB/oct Filter Shootout
24 dB/oct Filter Shootout
______________________
Dome Music Technologies
Dome Music Technologies
-
- Posts: 17
- Joined: Fri Jun 10, 2022 5:44 pm
Re: 24 dB/oct Filter Shootout
Thank you! I just happened to subscribe to your channel the other day! I've bookmarked the video to watch later so I'll let you know when I'm done working on it.
-
- Posts: 71
- Joined: Fri May 14, 2021 2:16 am
Re: 24 dB/oct Filter Shootout
So good. Would you be up for sharing the preset so we can switch out more filters in the test, please? I guess I could just recreate it from the screenshot but I'm lazy
Re: 24 dB/oct Filter Shootout
Here ya go! I'd be interested to see the results from other filters, too.GusGranite wrote: ↑Wed Jun 15, 2022 4:59 am So good. Would you be up for sharing the preset so we can switch out more filters in the test, please? I guess I could just recreate it from the screenshot but I'm lazy
Here at Dome Music Technologies, we like to eat our own dog food. You can replace the DC offsets from the ACE Constants & Multipliers module with the Cherry Audio DC Source.
______________________
Dome Music Technologies
Dome Music Technologies
Re: 24 dB/oct Filter Shootout
Since three of the filters in this shootout were coded by me, I thought I'd contribute a little clarification about their differences.
The CA Ladder, VM2500 Filtamp, and Mood 904 are all 4-pole, 24db/oct, Moog-style ladder filter designs with subtle differences between them. I don't know what's in the CA Filter or the S100 Dual Filter. The Laboratory VCF is quite different being a multi-mode state-variable design.
Ladder type filters are capable of a maximum Q*** (resonance) of 4 before they go into self oscillation. This makes for a very useful and musical sounding filter, but is hardly appropriate in many applications. For instance, if you try to cascade them to make speech sounds or simulate instrument resonances, they fail miserably. For those two applications, higher Qs (in the 10 to 50 range) and 2-pole designs are required. Also, if you want to bang the input and listen to the filter ring, you need even higher Q. The Lab VCF is capable of Q values up to 2000. It's actually clamped there and could probably go much higher, but I had to limit it somewhere. 2000 is plenty. The VM2500 1047 filter not featured in this shootout is similar and capable of Qs up to 500.
If you have the Lab Bundle, find the Talky preset which is programmed to say "Cherry Audio" and you'll quickly realize that that patch could never work with ladder filters. The VCFs' Qs are all set somewhere in the 20 range to simulate human vocal tract resonances. There's also a preset here viewtopic.php?t=1100 that shows percussive physical modeling using high-Q filters.
While this filter shootout was informative, in many ways it's an apples and oranges comparison. However, I would have loved to have heard some bwwwwaaah sweeps of all of them just to get a quick idea of the differences. Maybe somebody will do that.
Well, I didn't mean to write a manifesto, but I didn't have time to make it any shorter. I hope it was informative.
--mb
*** The Q of a resonant peak is equal to its frequency divided by its bandwidth (F/bw). Since both of these values are measured in Hz, Q has no unit and is a dimensionless figure. The Q literally stands for "quality", and you can see that the narrower (smaller) the bandwidth, the higher the Q. Additionally, unless compensated for, the gain of the filter at the peak is equal to the Q. Now you can see why I put an input level control and a meter on the Lab VCF in an attempt to save the planet from flying speakers!
The CA Ladder, VM2500 Filtamp, and Mood 904 are all 4-pole, 24db/oct, Moog-style ladder filter designs with subtle differences between them. I don't know what's in the CA Filter or the S100 Dual Filter. The Laboratory VCF is quite different being a multi-mode state-variable design.
Ladder type filters are capable of a maximum Q*** (resonance) of 4 before they go into self oscillation. This makes for a very useful and musical sounding filter, but is hardly appropriate in many applications. For instance, if you try to cascade them to make speech sounds or simulate instrument resonances, they fail miserably. For those two applications, higher Qs (in the 10 to 50 range) and 2-pole designs are required. Also, if you want to bang the input and listen to the filter ring, you need even higher Q. The Lab VCF is capable of Q values up to 2000. It's actually clamped there and could probably go much higher, but I had to limit it somewhere. 2000 is plenty. The VM2500 1047 filter not featured in this shootout is similar and capable of Qs up to 500.
If you have the Lab Bundle, find the Talky preset which is programmed to say "Cherry Audio" and you'll quickly realize that that patch could never work with ladder filters. The VCFs' Qs are all set somewhere in the 20 range to simulate human vocal tract resonances. There's also a preset here viewtopic.php?t=1100 that shows percussive physical modeling using high-Q filters.
While this filter shootout was informative, in many ways it's an apples and oranges comparison. However, I would have loved to have heard some bwwwwaaah sweeps of all of them just to get a quick idea of the differences. Maybe somebody will do that.
Well, I didn't mean to write a manifesto, but I didn't have time to make it any shorter. I hope it was informative.
--mb
*** The Q of a resonant peak is equal to its frequency divided by its bandwidth (F/bw). Since both of these values are measured in Hz, Q has no unit and is a dimensionless figure. The Q literally stands for "quality", and you can see that the narrower (smaller) the bandwidth, the higher the Q. Additionally, unless compensated for, the gain of the filter at the peak is equal to the Q. Now you can see why I put an input level control and a meter on the Lab VCF in an attempt to save the planet from flying speakers!
Re: 24 dB/oct Filter Shootout
I did originally include the 1047 (because I love it so much!), but with it being 'restricted' to 12dB/oct 2-pole response, it was vastly different to the others. Even moreso with the TB Filter, which is like nothing else out there!
That's bananas! Love it.If you have the Lab Bundle, find the Talky preset which is programmed to say "Cherry Audio"
When I get some spare time, I'll definitely set that up. I was thinking along the lines of:However, I would have loved to have heard some bwwwwaaah sweeps of all of them just to get a quick idea of the differences. Maybe somebody will do that.
1. zero resonance
2. just below self-oscillation
3. just above self-oscillation
4. max resonance
It was fascinating. Thanks for such a detailed response.Well, I didn't mean to write a manifesto, but I didn't have time to make it any shorter. I hope it was informative.
______________________
Dome Music Technologies
Dome Music Technologies
Re: 24 dB/oct Filter Shootout
______________________
Dome Music Technologies
Dome Music Technologies
-
- Posts: 71
- Joined: Fri May 14, 2021 2:16 am
Re: 24 dB/oct Filter Shootout
This is great!utdgrant wrote: ↑Sat Jun 18, 2022 1:37 pm Whaddaya know? I found some spare time!
Here is the "bwwwwaaah sweeps" video on YouTube.