Page 1 of 1
New Analog Instrument Has a HUGE DC Offset
Posted: Fri Jan 04, 2008 1:02 pm
by e-theory
Has anyone else noticed that the Analog instrument suffers from a DC offset problem when using a good deal of patches? Initially I thought it was just when using square or PWM waves - but a lot of the factory patches show issues in various combinations - if you have the Analog presets pack installed from the website try loading the "Mallet Free" patch under the "Keys" directory, play a few notes and then load the "Mellow Pan Pipes" patch which is directly after it to see what I mean. You can see that the synth initially produces a massive positive DC offset that takes quite a while to settle.
This DC offset is in the form of a tendency for the instrument to create waves that are "shifted" too far ( unbalanced ) in the positive direction. Real analog synths would never exhibit such behaviour as simply hooking up an analogue cable between two analogue points creates a very low frequency high pass filter - therefore this behaviour is somewhat against the namesake of this synth....
This creates large problems when using the Analog instrument within a mix - you try and layer up multiple Analog instruments and it basically produces premature positive clipping that simply should not be there....
For the moment I am working around this by inserting a 20hz high pass filter after it to remove the offset, but am very disappointed that an otherwise truly beautiful sounding synth is plagued by such an easy to remedy problem.
Ableton - please DC balance all of the waveforms and filters for Analog immediately. At the moment without external filtering it is not particularly useful to me and I have had a number of issues using the PWM wave with unison turned on within my own patches. Thanks!
Re: New Analog Instrument Has a HUGE DC Offset
Posted: Fri Jan 04, 2008 3:01 pm
by Synthbuilder
e-theory wrote:Real analog synths would never exhibit such behaviour as simply hooking up an analogue cable between two analogue points creates a very low frequency high pass filter - therefore this behaviour is somewhat against the namesake of this synth....
I can't help you on the problems with Analog as I haven't got it myself - I didn't like the sound of the filters. However, I should say that real analogue synths do not feature DC blocking cables or connections. A cable will pass all signals equally up to a maximum frequency, so if anything a cable will act as a low pass filter with a very high cut-off point.
That said, most analogue synths will incorporate DC filtering between their various stages. This should get rid of any unwanted DC offsets caused by pulse width modulation and other causes. Another way, and the best way in my opinion, to reduce DC offset on PWM signals in a real VCO is to add the PWM CV to the audio signal at the output of the VCO. This could be done in virtual analogues too, but I have yet to hear it.
Sometimes I wish that virtual analogue software designers would hire a real analogue engineer to test the product. Hint, hint...

Posted: Sat Jan 05, 2008 1:28 am
by arachnaut
I've noticed that as well. Probably it is a result of the physical modeling algorithms.
Just put a Utility after it with only 'DC' set. That should fix things.
Re: New Analog Instrument Has a HUGE DC Offset
Posted: Sat Jan 05, 2008 2:40 am
by Machinate
first of all, I haven't noticed the dc offset in Analog, but I have to addres this bit:
e-theory wrote:
This DC offset is in the form of a tendency for the instrument to create waves that are "shifted" too far ( unbalanced ) in the positive direction. Real analog synths would never exhibit such behaviour
Unfortunately this is blatantly wrong. Most proper analogs suffer a bit from various types of DC offset, simply due to the often occuring non-linearity of the designs in filters and whatnot. Especially modulars, which, by design, rely a fair bit on chaos, are badly afflicted in my experience(and make for fun scope-watching when it all goes AWOL. Oh, and a DC offset can be negative, too.
e-theory wrote:For the moment I am working around this by inserting a 20hz high pass filter after it to remove the offset, but am very disappointed that an otherwise truly beautiful sounding synth is plagued by such an easy to remedy problem.
high-pass filtering is often a good idea (in a mix in general too, of course) but have you had a look at the Utility plugin for this yet? I believe it now has DC offset removal as a feature.
Posted: Sat Jan 05, 2008 4:11 am
by e-theory
Hello all, thankyou for all of your responses.
I think I should clarify a few points:
1.) Indeed I made a mistake in my comment about cable capacitance contributing to a cable being a high pass filter. It has been a while since I've soldered analogue devices with a soldering iron and simply confused the configurations of high pass and low pass filters - after a quick read here:
http://www.soundscapeav.com/cable_ar.html - all is clarified and I remembered that a low pass filter has a capacitor in parallel and high pass in series with the resistance - not the contrary like I falsely remembered.
2.) When I was talking about DC blocking with regards to Analogue circuits I was referring to the typical capacitor coupling used between stages of analogue signal paths - these create a low frequency high pass effect that removes that DC and centres the waveform - whilst I relinquish that not all designs had capacitive coupling - all of the classic designs I have encountered - i.e. TB303 filter etc. etc. all did and therefore when visualized via oscilloscope at their output were always nicely balanced ( with regard to their supply rails obviously ).
3.) Yes a DC offset can obviously be positive or negative - but the one that the Ableton Analog synth creates is always positive and that is the synth I was referring to.
4.) I don't spend x dollars on a synth just to have to put a DC blocking effect after it - that type of thing should always be built in. A synth that can literally damage speakers is not a good thing. The only other synths I have ever encountered with the same issue is Ableton Operator and Novation V-Station - curiously Tension and Electric have no such issue....
Also after further investigation I have found that it only occurs when using the square and PWM waves in Analog and the square wave in Operator - the saw waves are fine.
Cheers all!
Posted: Sat Jan 05, 2008 5:24 am
by dj superflat
i don't get what you're talking about. i use operator and analog lots, with multiple layered voices, never run into this. what am i missing? is this something specific to how you're using them?
Posted: Sat Jan 05, 2008 12:51 pm
by e-theory
Here is a simple example.
The patch took 2 seconds to make.
Basically as you can see there is a large NEGATIVE DC offset ( which again contradicts my previous posts about the offset always being positive....

) but nonetheless shows that the signal is very different after DC filtering indicating that the offset is indeed an issue.
Here is the Analog preset to play with too:
http://www.evolutionary-theory.com/temp ... Preset.adv

Posted: Sat Jan 05, 2008 1:38 pm
by hp#b
Hi,
I think this is the very low LFO1 rate (0,1 Hz) what you see as a DC offset.
This is not a bug this is a feature of your

patch.
HP
Posted: Sat Jan 05, 2008 2:50 pm
by e-theory
Thanks for the feedback - that is an interesting idea but I think there is something else going on here than the LFO.
I have never ever seen a synth that adds an LFO output as an actual signal to the output. Such a thing creates much more danger than good.
The reason I know it is not the LFO is that when I switch the patch to a sawtooth wave oscillator - the offset disappears - even with the LFO enabled - it is only ever present with the square wave - which really doesn't make sense - it should be compensated for internally by the synth itself.
Posted: Sat Jan 05, 2008 3:21 pm
by hp#b
Hi,
It seems that you set OSC1's Pulse Width Modulation (by LFO1) to 1.
I think this is the key.
HP
Posted: Sat Jan 05, 2008 3:34 pm
by e-theory
Actually that makes no difference - any value other than 0 creates exactly the same problem - however at a value of zero - i.e. a perfect square wave - the result is perfect - no dc offset - thank you for inadvertently making me check that....
Therefore you have helped me to conclude once and for all that the issue lies with the PWM feature of the square wave and that when modulated the PWM wave is not being correctly centered within the synth ( i.e. PWM modulation creates waves that must be shifted off center to be dc neutral ) - therefore creating an imbalanced wave and hence being the current bug with Analog.
Ableton please fix this problem.
Posted: Sat Jan 05, 2008 3:44 pm
by hp#b
Good if I was able to help you. Together we are strong

Austria and Australia
HP