Haible Vokoder?

[ach_gott]

[Dego]

Size does matter!

[The Bad Producer]

Just saw these on ebay, cheap 1% caps:

http://cgi.ebay.co.uk/Vishay-10nF-50V-1-Ceramic-Caps-Capacitor-50x-/130346318385?cmd=ViewItem&pt=LH_DefaultDomain_0&hash=item1e593efa31

Useful? I hope so!

[TekniK]

[The Bad Producer]

Does that matter in this application?

[/mr]

I was curious about how different vocoder band distributions affect the sound... so I made a little comparison between different parametric EQ settings in a sound editing program just for fun.

The JH Living Vocoder has its 20 bands distributed somewhere around 3½ bands per octave, and here is a sound played through a 20-peak parametric eq with similarly placed peaks:
http://dl.dropbox.com/u/6263674/vocoder/vocoderbandmimicry_jh_living_vocoder.flac

Here is the same sound through an eq with 3 peaks per octave:
http://dl.dropbox.com/u/6263674/vocoder/vocoderbandmimicry_3_bands_per_octave.flac
And 4 peaks per octave:
http://dl.dropbox.com/u/6263674/vocoder/vocoderbandmimicry_4_bands_per_octave.flac

(Here is the original clean sound:
http://dl.dropbox.com/u/6263674/vocoder/vocoderbandmimicry_original.flac)

In the noise sound in the beginning, the 3-per-octave and 4-per-octave versions have a distinct "chord" quality to them (augmented chord and diminished chord, respectively), whereas the JH version is more diffuse in this respect. This was my primary aim with the experiment, since I expected this to make a difference. But which one is the best? I really don't know. For the best speech reproduction, I expect the 4-per-octave would be preferrable. For the most interesting vocoder sound, the 3-per-octave might be just as good or even better. For some applications I definitely think the more "non-harmonic" JH version is better, and in some cases it would be more interesting with the "chord" sound of the other ones.

Note that I used a parametric eq in software, not a vocoder where the phase behaviour is probably much messier. The amplitude responses, however, is set as close to the curve at JH's site with ~3dB between peaks and valleys, and all examples have 20 peaks.

(Btw, here's an invite to Dropbox, if you want the simplest way of sharing files: https://www.dropbox.com/referrals/NTYyNjM2NzQ5 )

[jhaible]

http://www.jhaible.de/vocoder/one_day_later.jpg

No idea if this whole thing is even going to work - just soldering and soldering ...

JH.

[2thick4uni]

[TekniK]

[drooo]

This project looks nuts. I think I have no choice but to actually build one!

[emdot_ambient]

[drooo]

I think there is a good reason that the Sennheiser Vocoder VSM201 was so expensive--we've only seen what the board looks like--the panel wiring is going to be very interesting to see too.
Juergen--are you bringing individual bands out for cross patching on your own personal vocoder?

[jhaible]

[axxim]

Greetings all,

I am new here and just registered after seeing that topic here. I am a vocoder fanatic and couldn't help as I landed in this forum. I see you have started a pretty but huge project here and even if I have read only the first and last posts to check if it is still active, I see you have done (or still are doing) a lot of work.

A was tempted to design and build my third vocoder some days (weeks) ago and are still weighthening the different design options. But let me first introduce myself a little:
I am an electrical engineer and due to my actual job I have a lot of experience in designing, layouting and let manufacture electronic boards. My first vocoder was based on elektors design about 20 years ago. Since it was not that quality I expected, I sold it to a musician who published several CD's using it. I was then captivated by building a digital one and was waiting for an affordable DSP evaluation system. At that time, Atari launched the Falcon-30, which had a DSP 56000, an audio matrix and a whole home computer behind it. I programmed and launched a vocoder, which as far as I know was the first available 16 band realtime software vocoder in the world. This was more than 17 years ago. Then my Atari ceased to work during a second project I was working on. Since then I had turned away from music (work and family). Months ago I bought a Korg Radias to reenter music and have a vocoder again.

Recently an Atari user community contacted me to freeware my vocoder and after searching a installable version of it and finally, when a friend found one on his old archives, I (as someone said at their forum) "blessed" it as freeware. This revival of my vocoder and some suggestions and requests to program or build one again, have lead me to where I am today: thinking to design and build my third vocoder.

Well it's late now and I just wanted to register and say hello tonight. I'll take a deeper look at the posts here and come later to comment and/or ask about this project and interchange ideas.

[JoeMorris]

Hows the prototype going Jurgen?

[jhaible]

[axxim]

Hi Jürgen

looking at that board... wow, a lot of work to do. Did you use a board flipper (or component holder) or do you have to turn it after each component to solder it?

I have a question about the analyzer envelopes: is the attack/decay setting global (equal) for all envelopes?

[jhaible]

[drooo]

Any idea what the power supply requirements are going to be for this one?

[jhaible]

[/mr]

[jhaible]

And then there are the LEDs.
I'm pretty confident that you don't need 80 Watts, but even if an 80 Watt transformer is too small, a 120 Watts type won't cost that much more.

JH.

[jhaible]

Powering it up!
Nothing was exploding.
Current measured on +12V and -12V was about 500mA each. This is without signal, and without channel LEDs.
Expect the current to stay below 1A (DC) with all LEDs turned on.
AC current (fuses after transformer) read 900mA each. Expect this to stay below 2A with LEDs and all.
So the 80W toroidal transformer (2 x 15V secondary) is slightly oversized, and exactly the right choice for this project.

JH.

[jhaible]

Ok, this is *not* how the final produt will sound. It's the first sounds coming out of the baby after powering it on, still unadjusted whatsoever, and not debugged. But it's vocoding, so there is no fundamental error to be expected, and I can start fine tuning the circuit. - Oh, and the overload is from the microphone of my mobile phone, which I used to make that crappy video.
But I'm glad tjis project has finally started to become alive.

http://www.youtube.com/watch?v=2gmmLl94NkM

JH.

[/mr]

Wonderful!

You should definitely put the URLs to this discussion thread and to your private Vocoder page in the YouTube description text, because this is going viral!