u-he

Howard · Post by **Howard** » Fri Jun 26, 2009 9:01 am

Marbellous!

@hakey: Which part of "greatness" Yorkshire?
I'm originally from Hull (well, Patrington).

hakey · Post by **hakey** » Fri Jun 26, 2009 9:50 am

Howard wrote:@hakey: Which part of "greatness" Yorkshire?
I'm originally from Hull (well, Patrington).

Right over on t'other side

- near Settle at the western edge of the Dales.

(Another KVR member, a Lancastrian expat no-less, has 'the grimness of Yorkshire' as his location

)

billstei · Post by **billstei** » Fri Jun 26, 2009 3:12 pm

Howard wrote: Marbellous!

billstei · Post by **billstei** » Fri Jun 26, 2009 3:58 pm

hakey wrote:
billstei wrote:One of my highly guarded secret brainfarts is to put a post-fft harmonic processing module/panel into the middle space of the Italian section, that would do some of the stuff I was doing with Delta patches, for example "Increase Even Harmonics 50%" and other fiddle faddle of that sort.
Using Spectroblend within Zebra's OSC windows it's possible to draw harmonics/partials with both positive and negative amplitudes (that is above and below the x axis). Am I right in thinking that partials of opposite sign are of opposite phase? If so, might it not be an idea to allow some way to manipulate partials of both phase within the same window?

For example, reading up on the acoustics of brass instruments (poss. in one of the SOS 'Symth Secrets' articles) I'm sure that I read something along the lines of "in the case of 'x' the phase of every even partial is reversed".

If none of this makes sense just ignore.

Neither the audio source window(tabs) at the top, nor the harmonics window at the bottom are "editors" in the sense that you normally would use an editor, aka direct manipulation of the data points -- they are read-only views. Since Zebra already has an editor per se, it is not a priority for me to do that. In fact one could argue that these creations of mine are test-beds for features that might one day make their way into Zebra and make my program(s) obsolete -- let's hope so -- but if not then they can always be useful auxiliary tools.

Regarding negative values in the harmonics window, as I understand it, these are cosines, so that what we have available is a choice between a sine (at a given frequency) at 0 degrees, or a sine at 90 degrees phase shift (and not both simultaneously). What I have available from the fft analysis is a set of complex numbers, of sines at phases varying from 0 to 360... the question then is whether Zebra can be made to use that data directly? I don't think so. I have been meaning to look into whether there is a mathematic transform that can make better use of the data, but haven't gotten around to it

Hopefully Urs can offer advice, and/or his plans for future improvements to Zebra.

As a comparison, look at Sytrus from Image-Line, which allows the user (under the OSC tab of each Operator) to set both the magnitude of the harmonic with an upward stem, and the phase with a downward stem, simultaneously, with the phase varying from 0 to 360.

So, for now, I am taking the absolute value of the complex number, aka sqrt((real ^ 2) + (imaginary ^ 2)), and treating that as the magnitude of a sine at 0 degrees. And even corrupt data can sound

sometimes, so I would keep this "feature" even if I discovered other ways to use the data.

hakey · Post by **hakey** » Fri Jun 26, 2009 5:27 pm

Thanks for that comprehensive answer billstei. I've learnt a little bit there - I didn't know that fft analysis calculates sines of varying phase, but that makes sense when one thinks about it.

As regards the ability to do additional spectroblend editing (on top of fft sample to spectroblend conversion - that's what you're talking about here, right?), well I was just seizing upon your 'post-fft harmonic processing module comment' and obviously leapt to conclusions about how far you might like to go with it.

But it is your baby, you do as you see fit. I'm sure that it'll be fun.

billstei · Post by **billstei** » Sat Jun 27, 2009 2:12 am

hakey wrote:Thanks for that comprehensive answer billstei. I've learnt a little bit there - I didn't know that fft analysis calculates sines of varying phase, but that makes sense when one thinks about it.

As regards the ability to do additional spectroblend editing (on top of fft sample to spectroblend conversion - that's what you're talking about here, right?), well I was just seizing upon your 'post-fft harmonic processing module comment' and obviously leapt to conclusions about how far you might like to go with it.

But it is your baby, you do as you see fit. I'm sure that it'll be fun.

One way to accomplish what you are saying would be to have an offset parameter for all/each of 128 harmonics. But in the end what you see is entirely algorithmic, and is (more or less constantly) being updated in realtime.

billstei · Post by **billstei** » Sat Jun 27, 2009 4:28 am

@hakey - The more I think about the idea of an offset, the more I like it... what I envision would be a second graph line over the existing Zebra-ish stem/line Wave display, and this line would be editable directly. It's value is added (including negative values) to the resultant harmonic values. It really bothers me though that it interferes with the Italian part of the GUI

Urs · Post by **Urs** » Sat Jun 27, 2009 8:19 am

The idea of a phase offset is cool, but how do you get it into Zebra then?

billstei · Post by **billstei** » Sat Jun 27, 2009 9:03 am

Urs wrote:The idea of a phase offset is cool, but how do you get it into Zebra then?

Actually I meant an offset added to the magnitude of a given harmonic, not the phase of the harmonic.

Urs · Post by **Urs** » Sat Jun 27, 2009 9:05 am

billstei wrote:Actually I meant an offset added to the magnitude of a given harmonic, not the phase of the harmonic.

Oh... I guess I'm too confused with all those grapes around

billstei · Post by **billstei** » Sat Jun 27, 2009 9:12 am

Urs wrote:
billstei wrote:Actually I meant an offset added to the magnitude of a given harmonic, not the phase of the harmonic.
Oh... I guess I'm too confused with all those grapes around

Those are pre-fermented so please be careful.

billstei · Post by **billstei** » Thu Jul 02, 2009 7:02 pm

Real deal update:

Got a few more things working... The audio source waveforms are being drawn now, along with the first/last sample marker lines. It can write out an actual usable Zebra Script. In theory (I think) you could do everything that the original Ver 1.0 Octave script could do. Lots of stuff to work on yet, and I will be on vacation next week, but not to Italy. Speaking of Italy... where are the photographs of the Urs vacation? How am I supposed to make a decent background image of the Mediterranean without something to work with?

Urs · Post by **Urs** » Thu Jul 02, 2009 10:31 pm

billstei wrote:where are the photographs of the Urs vacation?

Oops, havn't even imported them into iPhoto yet

billstei · Post by **billstei** » Thu Jul 09, 2009 3:55 pm

Greetings from just south of the Great White North. Just before I left on vacation I fixed a fatal crash type bug that had been bugging the Blueberry Thing.

If I don't get a picture of the Mediterranean from Urs' vacation I may be forced to use a picture of the North Shore of Lake Superior from mine, and nobody wants to think about 39 F water in the middle of summer, but I'm a desperate non-artist, and I have a camera, so this might get ugly.

billstei · Post by **billstei** » Fri Jul 10, 2009 2:24 pm

Breathing all this fresh air has inspired me to consider using libsdl ( www.libsdl.org ) to solve the next big problem in Blueberry Thing (aside from a gazillion other "little" features), that of having playback of the audio sources/clips. I would be interested to know if OSX has libsdl available by default, or if it is a pain to install or use.

u-he

KVR Audio

Sine Sine Everywhere a Sine