VirSyn User Forum - Shaper Noise

YaBB Newbies

Admission of
ignorance is the
beginning of wisdom

Posts: 19
Gender: male

Shaper Noise
12/29/09 at 17:31:18

Happy New Year Harry and Everyone,

After a PC crash and rebuild earlier this year, I've happily rediscovered TERA3 when selecting the VSTis' to re-install.

I've been experimenting with waveshaping on TERA, and finding some very promising sounds except for one PROBLEM: noise artifacts.

On TERA 3.2.3 on PC running Windows XP SP3, Alesis iO26 Firewire running native ASIO @ 48K Hz

Routing: Starting with default "new program" patch
Osc1 (Sawtooth - 0 Transposition - Tracking Enabled) wired to WaveShaper then wired to Amplifier (no overdrive)

Waveshaper Settings:

Level = 0.75/Shape 7 -->> F3 and below sounds Ok*; G3 and above introduces prominent noise artifacts
Level = 0.25/Shape 7 -->> F3 and below sounds Ok*; G3 and above generates noticeable volume increase; noise not as pronounced as above.

Level = 0.25/Shape 6 --> F3 and below sound Ok*; G3 and above generates noticeable volume increase; noise artifacts are prominent.
Level = 0.75/Shape 6 --> F3 and below sounds Ok*; G3 and above generates a noticeable volume increase; noise artifacts are detectable.

Level = 0.75/Shape 3 --> Similar to Shape 7; F3 and below sounds Ok*; listen to the differences between F4, G4, and A4 however.

The other Shape #'s except for Shape 2 produce similar behavior.

* Below G3, there is some clicking that seems related to the pitch.

The problem occurs in the TERA standalone host and VST hosts. It is very pronounced with the Sawtooth wave. The Triangle wave does not produce the key related problems itemized above, but some of the general, pitch-related clicking noticed below F3 is noticeable with the Triangle. I don't hear any problems with the Parabolic or Sine waves, but wouldn't use these for waveshaping too much anyhow.

I really like the sounds below F3 with the Sawtooth, and I'm hoping that waveshaping on this waveform will be usable in 2010.

Can anyone recreate the problem, and if so the problem get fixed in a non-donglable TERA update?

Thank you very much.

Bob

Back to top

IP Logged

Schräge Musik

Senior Member

With all due
respect..........

Posts: 406
Gender: male

Re: Shaper Noise
Reply #1 - 12/29/09 at 18:07:48

Hi Bob. Happy new year to you too.

My studio is currently in bits while I install a new computer. Unfortunately I prefaced the operation with the phrase "this shouldn't take long" embarrassed

. Moving the old music computer from one side of the room to the other has persuaded it to throw a sulk of monumental proportions. The new comp is, of course a blank page and is going to take a while to get get loaded.

I'll have a go at recreating your problem (should be easy with the commendable amount of detail you've suplied) as soon as I get a machine working.

Back to top

I remember a time when....... Actually, no. I don't. That wasn't me.

Email | WWW

IP Logged

Bob Tormey YaBB Newbies Admission of ignorance is the beginning of wisdom Posts: 19 Gender:	Re: Shaper Noise Reply #2 - 01/15/10 at 01:57:54 Here's an MP3 of the first scenario. Notice notes 6-8 ... http://www.mediafire.com/?hkntztzkmli Here's the patch: http://www.mediafire.com/?lyqorvlem0m
Back to top	« Last Edit: 01/15/10 at 10:03:58 by Bob Tormey » IP Logged

Bob Tormey YaBB Newbies Admission of ignorance is the beginning of wisdom Posts: 19 Gender:	Re: Shaper Noise Reply #3 - 02/20/10 at 08:32:44 Here's a visual. You can see the noise creep in around 12K Hz. http://www.mediafire.com/file/qmmzyzzjcvo/ShaperNoise_Sawtooth.jpg
Back to top	IP Logged

Harry Gohs

YaBB Administrator

Posts: 740
Gender: male

Re: Shaper Noise
Reply #4 - 02/22/10 at 10:46:55

Hi Bob,

i have downloaded and analyzed your patch. The noise you hear is coming from aliasing
and is inevitable with this patch. The Shaper is a highly nonlinear distortion unit and is
therefore very prone to aliasing. We constructed the Shaper to be aliasing free with a sine
input signal, but the richer the input signal is the more aliasing you will have.
If you want to use a richer signal as an input there are only two ways to reduce the alisiang:

1. Increase the samplerate to 88.4 / 96.0 kHz

2. Lowpass the shaper input signal to reduce the amount of high partials in the signal.

Hope this is of help.

Harry

Back to top

IP Logged

Bob Tormey

YaBB Newbies

Admission of
ignorance is the
beginning of wisdom

Posts: 19
Gender: male

Re: Shaper Noise
Reply #5 - 02/23/10 at 04:58:01

Thank you for taking time to analyze the patches Harry.

Selecting a 192 Khz sample rate takes care of the problem! Wink

The Sawtooth through Tera's Shape 6 is a fantastically full, rich, and ironically organic sound until the aliasing kicks in. The first four or five partials are about equal in power which creates the tight but thick sound. The control of timbre makes for wonderfully expressive opportunities.

Looking for the next best equivalent, I tried using Tera's distortion effect module with low cutoff and high tone, and got something in the neighborhood (the fundamental still dominates), but the modulation opportunities aren't there while the aliasing is present similar to the shaper.

Not giving up, I stumbled upon THE SOLUTION!

Using Tera's "Custom Wave" Partial Editor, I drew only the first five partials of the Sawtooth. When I played that through Shape 6 ... Voila, no aliasing* on a sound that's awfully close to the native sawtooth through the shaper. I'm very happy, and will start working this scheme into some new sounds.

There aren't that many waveshaping synths out there (and perhaps now I know why). Thanks to your follow-up Harry, I'm expecting to have a lot of fun with a few partials and a "highly nonlinear distortion unit"!

Bob

* At 48 KHz sampling rate (more my speed), there's some "normal" aliasing artifacts with the truncated sawtooth at the very high end of the register, but nothing like the obtrusive buzzing and ticking when all of the partials are present.

Back to top

IP Logged