Comparing the Saws (of various Reason synths)

[eusti]

The kHs One sale thread got me thinking... Are all those synths really necessary?
Well, of course different layouts are leading to different approaches and that might spark one's creativity in different ways...
But I wanted to see what the differences are on the most basic level... So I did a small test with most of my Reason synths playing the same
(adjusted in pitch to allow for different tuning philosophies) sequences and matched them in volume using Selig's gain...

Here's what all the waves looked like all together...



Antidote:



kHs One:



Predator:



Ivoks:



Red70:



Tres:



MonoPoly:



SubTractor:



Malström:



Thor:



ReDominator:



ABL3:



Viking:



Revival:



Megasaur:



Oberon:



And here finally a link to the resulting audio file:
https://dl.dropboxusercontent.com/u/408 ... reSaws.mp3

Personal observations:
- I have too many synths! I like them all and haven't spent enough time with most of them to have really mastered most of them.
- Many of them sound similar to me. Interesting that the waveforms look so different in some of those cases.
- Although I leveled them all using Gain some of them sound less loud to me. Looking at the diagrams suggested that was because some of them are asymmetrical...

This test was not done to suggest that one synth is better than the other, but just to see what the basics are like and how they compare...
I tried to make all synths sound as close to each other as possible by adjusting the VCA envelope to fit. Filter was open with no resonance. If you feel I made a mistake in the setup please let me know and I'll look into that.

D.

[EnochLight]

Very cool. What I can see visually is there are definitive differences in some of these. What I can hear from your example audio file is differences in some, and similarities in others.

For me, all of these different synths mean one thing: access to a vast preset patch library with the ability to customize each through their own customized front-end. For me, that's priceless.

[eusti]

Very cool. What I can see visually is there are definitive differences in some of these. What I can hear from your example audio file is differences in some, and similarities in others.

For me, all of these different synths mean one thing: access to a vast preset patch library with the ability to customize each through their own customized front-end. For me, that's priceless.

Definitely a good point about the sound libraries... But I think as well that some synths just by the way they've been setup lend themselves more to same sounds than to others... And that's before we even talk about special features as crazy FX and so on...

D.

[DLDTech]

Megasaur looks funny because it has a high pass filter after the osc. This was also done on the original synth with a supersaw.

Great research, and interesting to see that some synths replicate analog oscillators, whilst others take a more mathematical approach.

[eusti]

Megasaur looks funny because it has a high pass filter after the osc. This was also done on the original synth with a supersaw.

Great research, and interesting to see that some synths replicate analog oscillators, whilst others take a more mathematical approach.

I tried to make sure that the HP was open like the LP. Please let me know if I made a mistake here! I like your synths and don't want to misrepresent them! Is there a way to get a saw with Snakebite? I looked at the manual and couldn't find one...

D.

[rcbuse]

Interesting analysis, but the shape of the waveform is a very poor indicator of how it sounds. Sure, if you see a perfect virgin pulse or saw, you have a good idea of the tiber it produces from experience, but the second you pass it though even a simple all-pass filter, you shift the phases of the harmonics all around. Then you have a waveform that sounds exactly like the original but looks nothing like it. If you really want to know how its going to sound, you have to look at the spectrum.

All these waves are identical to the ear


More info
http://www.till.com/articles/squares/index.html

[eusti]

Interesting analysis, but the shape of the waveform is a very poor indicator of how it sounds. Sure, if you see a perfect virgin pulse or saw, you have a good idea of the tiber it produces from experience, but the second you pass it though even a simple all-pass filter, you shift the phases of the harmonics all around. Then you have a waveform that sounds exactly like the original but looks nothing like it. If you really want to know how its going to sound, you have to look at the spectrum.

All these waves are identical to the ear


More info
http://www.till.com/articles/squares/index.html

Thanks, rcbuse. I only included the waveforms because I found them interesting to look at and as well was surprised to see that some of the ones that sounded very close to me looked quite differently... If you are interested I can add spectrogram views as well...

D.

[DLDTech]

Megasaur looks funny because it has a high pass filter after the osc. This was also done on the original synth with a supersaw.

Great research, and interesting to see that some synths replicate analog oscillators, whilst others take a more mathematical approach.

I tried to make sure that the HP was open like the LP. Please let me know if I made a mistake here! I like your synths and don't want to misrepresent them! Is there a way to get a saw with Snakebite? I looked at the manual and couldn't find one...

D.

Nope - you did it right. The HPF is to allow aliasing to occur at the top end (which adds 'sparkle') without getting aliasing at the bottom end (which sounds 'muddy'). It's part of the secret sauce of the supersaw.

Snakebite doesn't have a true saw as it merges from sine to triangle to square..

Be interesting to see the frequency responses of each wave.. Our ears are more tuned to frequencies rather than actual waveforms so the FFT may be more revealing.

[chk071]

Interesting analysis, but the shape of the waveform is a very poor indicator of how it sounds.

Agreed. The sawtooth in Sylenth1 is almost a perfect sawtooth wave too, yet, when comparing it to Dune 1, which also looks quite similar, they sound differently, and Thor, which also seems to be an almost perfect saw sounds different too.

[rcbuse]

Thanks, rcbuse. I only included the waveforms because I found them interesting to look at and as well was surprised to see that some of the ones that sounded very close to me looked quite differently... If you are interested I can add spectrogram views as well...

D.

Thanks, but I have too many spectrograms here! I've got an upcoming RE that makes extensive use of them.

[eusti]

Thanks, rcbuse. I only included the waveforms because I found them interesting to look at and as well was surprised to see that some of the ones that sounded very close to me looked quite differently... If you are interested I can add spectrogram views as well...

D.

Thanks, but I have too many spectrograms here! I've got an upcoming RE that makes extensive use of them.

Looking forward to your upcoming RE!

But too late for this thread now... Will add the new info to the first post in a moment!

D.

[eusti]

Megasaur looks funny because it has a high pass filter after the osc. This was also done on the original synth with a supersaw.

Great research, and interesting to see that some synths replicate analog oscillators, whilst others take a more mathematical approach.

I tried to make sure that the HP was open like the LP. Please let me know if I made a mistake here! I like your synths and don't want to misrepresent them! Is there a way to get a saw with Snakebite? I looked at the manual and couldn't find one...

D.

Nope - you did it right. The HPF is to allow aliasing to occur at the top end (which adds 'sparkle') without getting aliasing at the bottom end (which sounds 'muddy'). It's part of the secret sauce of the supersaw.

Snakebite doesn't have a true saw as it merges from sine to triangle to square..

Be interesting to see the frequency responses of each wave.. Our ears are more tuned to frequencies rather than actual waveforms so the FFT may be more revealing.

Thanks for the feedback!

D.

[Skullture]

Eusti, what key did you play this in? Just curious ^ ^ .

EDIT: nevermind, i found your audioclip nice test

[eusti]

Eusti, what key did you play this in? Just curious ^ ^ .

Notes are C & B. See attached image. The graphs are from the second note (B).

D.

[eusti]

Eusti, what key did you play this in? Just curious ^ ^ .

EDIT: nevermind, i found your audioclip nice test

Thanks!

D.

[avasopht]

Yes, you want to know what frequencies are present, which components are harmonics, which are tones from aliasing and which are inharmonic tones from distortions.

You can inquire further with noise components as well as modulation of frequency components to detect lfo, and detuned oscillators.

[eusti]

Yes, you want to know what frequencies are present which components are harmonics, tones from aliasing and inharmonic tones from distortions.

Uh? Not sure I understand... I updated the original post with spectrograph images...

D.

[XysteR]

Wow eusti nice work! I feel this study proves I was correct with my recent youtube argument. But like eXode said, it is fair to say that all synths overall character come from everything else in a synths signal flow. The differences between some of these sawtooths is really quite marginal, some look and sound pretty much the same as eachother. I think if it were possible to do the same test with a lot of different filters using the same sawtooth then we'd hear a big 'character' difference between filters. So, i'm feeling my understanding that a filter has more influence over a synths character than the OSC was on the right path - I'm not going mad after all lol

[eusti]

Wow eusti nice work! I feel this study proves I was correct with my recent youtube argument. But like eXode said, it is fair to say that all synths overall character come from everything else in a synths signal flow. The differences between some of these sawtooths is really quite marginal, some look and sound pretty much the same as eachother. I think if it were possible to do the same test with a lot of different filters using the same sawtooth then we'd hear a big 'character' difference between filters. So, i'm feeling my understanding that a filter has more influence over a synths character than the OSC was on the right path - I'm not going mad after all lol

Thanks, XysteR.

Not sure how to set up a test with filters... I seem to remember that there was a test of filters somewhat recently... Unfortunately not all tested synths allow for external audio to be filtered, but it would open up the test to external filters as well...
So, I guess one would need a standardized source material to be filtered. Then a decision should be made what kind of filter settings would be interesting and if there should be filter sweep and all that...
I'd be open to set something up like that, but a bit unsure what the parameters of interest would be...
Please chime in if there is any interest for that.

D.

[eusti]

By the way, I found the thread about the Reason filter comparison: http://www.reasontalk.com/viewtopic.php ... er#p211122

D.

[bpmorton]

Don't know if anybody pointed this out yet. Thor has at least 3 different sawtooth waves: Analog, Wavetable, Phase Mod, you can probably squeeze out another with some FM.

[eusti]

Don't know if anybody pointed this out yet. Thor has at least 3 different sawtooth waves: Analog, Wavetable, Phase Mod, you can probably squeeze out another with some FM.

Thanks. I went with the basic one here.

D.

[avasopht]

Yes, you want to know what frequencies are present which components are harmonics, tones from aliasing and inharmonic tones from distortions.

Uh? Not sure I understand... I updated the original post with spectrograph images...

D.

What I mean is, you can get a good idea of what's different visually by labeling / colouring the frequency components present. Simply seeing that significant components are not harmonics will tell you a great deal about why that sound is different, particularly if they could identify the specific artifact or signature sound.

[Bonkhead]

Last week I was just checking megasaurs wave in the spectrum, great job recreating the original supersaw

http://www.adamszabo.com/jp6k/

[orthodox]

I would subtract one mathematically accurate sawtooth spectrum from all samples, then the differences would be more clear to see.
I also have suspicions about lossy format of the audio file (mp3) and about low sampling rate (44100), these may affect the results.