[sdiy] Experiment with bandlimited ramps

[Tom Wiltshire]

On 6 Jan 2009, at 21:36, Seb Francis wrote:

> As an experiment I synthesised a full bandwidth ramp in Sound Forge  
> and noticed that this sounds like it has quite a lot more high  
> frequency content that even the 150 harmonic file Tom posted.
>
> http://burnit.co.uk/sdiy/stuff/FullBandWidthSaw.wav
> I guess what I'm hearing as extra high frequency content here is in  
> fact the aliasing distortion.

I don't know how Sound Forge makes ramp waves, but if it's a naive  
ramp, the aliasing will be a significant effect. The ramp wave only  
rolls off at roughly 6dB/oct, so you need to go stupidly high to get  
the harmonics down to an insignificant level (say, -90dB). Luckily,  
human hearing gives up well before that, so we can safely ignore a  
lot of this high frequency content. Hence my bandlimited experiment.

> I also made a recording of a MOTM Saw as a comparison to the  
> calculated waveform:
> http://burnit.co.uk/sdiy/stuff/MOTM_Saw.wav (MOTM)
> http://www.electricdruid.com/ramps/Ramp150.aiff (Tom's BLR)
>
> I matched the level and frequency precisely and while it sounds  
> pretty close, I have to say I do prefer the sound of the MOTM.  It  
> sounds to me, dare I say it, "warmer" or more "alive".

Chuckle chuckle! Be very careful here, young jedi!

I did a similar recording. In my case, I recorded a single ramp wave  
from my Pro-One (pair of CEM3340 ramps) and compared it with a pair  
of bandlimited ramps from my dsPIC DWGS oscillator (which only has 64  
harmonics incidentally).
Have a listen to that and see what you think:

http://www.electricdruid.com/ProOneShort.aiff

http://www.electricdruid.com/ProtowaveShort.aiff

There's some difference in the beat frequency between the two sets of  
oscillators, but in terms of actual sound quality, I was pleased with  
the digital version. Perhaps the fact there are two oscillators hides  
the digital sterility?! ;)
To be honest, I got all "engineer" about it, and decided that "good  
enough" probably is just that, even if there is some tiny/subtle/ 
measurable difference.

T.