SynthMaster Filter Improvements: Share Your Thoughts!

[kv331]

Hi folks,

Here's a recent update on the "analog" filters:

I made the "slope" parameter continuous for "analog" filters, so that you'll be able to set the slope between 0 dB/Oct - 24 dB/Oct

Here's a sound sample that demonstrates slope envelope:

http://www.kv331audio.com/synthmaster/s ... sample.mp3

There's only a single "additive" oscillator: 3 saws slightly detuned (using drift+detune). The cutoff frequency in the above sample is constant, the timbre comes from the change of the slope (via ADSR env2)

An ETA on the next update: Well, the R&D phase is really taking a lot of time, so I cant promise anything for 2.1 before I finish the filters + mod matrix. That'll be around the end of February.

[kv331]

BTW, I compared SynthMaster's CPU usage with Sylenth1, they're really close (using 4 osc + 1 24dB/oct analog filter, at "normal" engine setting. Sylenth1 is still slightly better at the CPU usage)

Another important that'll be in 2.1 is global settings for engine quality and engine buffer size. While rendering a project you'll be able to change the globals and render at better settings

[Aloysius]

Another important that'll be in 2.1 is global settings for engine quality and engine buffer size. While rendering a project you'll be able to change the globals and render at better settings

[kv331]

Hi everyone,

Dont be fooled by the silence, work is in "good" progress

I've been comparing the "digital" filters of SynthMaster with Access Virus TI, and so far here's what I came up with for the 2 pole LP filter. Please listen to the following mp3 and share your comments:

SynthMaster and Virus TI LP12 filters side by side

In the Virus filters there's a click at the beginning of the sweeps, I'll try to figure out how to implement that as well, probably by inserting some noise in the filter?

Tonight I'm going to be working on the 4 pole version and I hope that'll be identical to the Virus TI 4 pole as well.

I also decided to add more filter types:

Bandpass & Bandshelf (Peaking filter)
LP LP dual
LP HP dual
HP HP dual
BP BP dual

"Dual" means there are two filters running in one filter: The parallel/series parameter will also be continuous, it wont be a switch.

For the dual filters above: For "digital" ones the slopes will be 12/24 dB
For the analog dual filters: The slope of each filter will be continuously variable.

All of the above changes will take a couple of weeks but once I sort out the LP prototypes things will go faster. After I'm done with the filter implementation I'll update the "developer" skin and then hand it out to Jonathan for our new skin (which will contain a mod matrix to simplify the modulations)

[rectus_dominus]

Sounds like a massive update is coming...
Hope there'll be some more room for bug hunting

[FragileX]

Yeah fantastic news!! The mod matrix will be very welcome
Really looking forward to the next update...

Thanks for keeping us informed

[Cyforce]

If it goes so massive forward, how it looks, the synthmaster2 must really renamed till update 2.5 to synthmonster or synthmaster-with-magic-filters

[keel]

Damn sm low end sounds so faking awesome. 1hz-800hz, around that is super phat, just like in hardware. Then upper body comes in and ruins overal sound

That was just my opinion

Great work on filters.

[Aloysius]

Not sure if anyone mentioned it but SynthMaster is pictured ina short article on PG14 of SoundOnSound February 2010, Volume 25 - Issue 4! Apparently . . . you make a strong product . . .

[Cyforce]

Not sure if anyone mentioned it but SynthMaster is pictured ina short article on PG14 of SoundOnSound February 2010, Volume 25 - Issue 4! Apparently . . . you make a strong product . . .

really? cool!

Mag-reviews are for a synth extremly important. So bulent, after next update you should contacted all the big magazines

[Aloysius]

Just did a quick Google. You can see the short article here;

http://www.soundonsound.com/news?NewsID=12312

[zvenx]

I want to say it is much closer, but compared to the virus ti it still sounds a little bit too clinical/sterile to me, but this could be because I know which one played first because of the click you mentioned. would be interested in hearing a different comparison without you telling me which is which to see if I stand by the first sentence.
thanks, really appreciate all the work.

rsp

[kv331]

I want to say it is much closer, but compared to the virus ti it still sounds a little bit too clinical/sterile to me, but this could be because I know which one played first because of the click you mentioned. would be interested in hearing a different comparison without you telling me which is which to see if I stand by the first sentence.
thanks, really appreciate all the work.

rsp

That could also be related to the oscillators: The saw waveform in Virus is not the same as the saw in SynthMaster. Filterwise, I'm almost sure that the two filters are calculating the same biquad equations (They're both digital biquad filters by the way, no analog modeling at all). As I said there could only be one different between the two filters: Virus is probably adding a slight amount of noise in the filters. I'll add that to SM and compare again as well

[kv331]

really? cool!

Mag-reviews are for a synth extremly important. So bulent, after next update you should contacted all the big magazines

I am deliberately not doing that for now Once I improve the sound quality and release 2.1 I'll do that. Just want to get good reviews and 2.1 will be the "right" version to achieve that.

BTW this R&D phase is taking a lot of time for me, so I wont be able to release anything before the end of February. I'll try to finish the coding for the filters by the end of next week (end of February). Then I can release a development build with updated "development skin" and everyone starts playing with it. (Those who want to download it will need to PM me to get the download link)

BTW one extra goody I want to mention: I improved the the startup time of SM2. And maybe some little CPU optimizations for the oscillators.

Speaking of optimizations: last weekend I realized an interesting behavior on CPU cycles: Let's say you have a single basic osc patch, with 8 unison voices. Call this patch A. You also have another patch with an additive osc, all of the 8 waveforms are turned on for that additive osc. Compare the CPU comsumption of A & B and you'll see that A is a LOT worse than B. I'd honestly expect that A would be slighly worse but that's apparently not the case. In the short run I wont be able to do much for that but once I release 2.1 and then the Mac OSX AU & VST, I'll tackle this optimization issue.

[zvenx]

thanks for the open communication.
rsp