On Mon, 15 Jun 2015 13:36:21 -0400, Randy Yates
<yates@digitalsignallabs.com> wrote:
>N0Spam@daqarta.com (Bob Masta) writes:
>
>> On Sun, 14 Jun 2015 12:23:02 +0000 (UTC),
>> spope33@speedymail.org (Steve Pope) wrote:
>>
>>>Bob Masta <NoSpam@daqarta.com> wrote:
>>>
>>>>On Sat, 13 Jun 2015 14:10:34 -0400, Randy Yates
>>>
>>>>>I wouldn't call it a vocoder. A vocoder uses subtractive synthesis,
>>>>>while this technique uses additive synthesis (or you could call it a
>>>>>type of additive synthesis). Also, a vocoder has two input channels,
>>>>>while this only has one.
>>>
>>>>Apparently there must be vocoders out there that I haven't
>>>>heard about, or you are using a different definition of
>>>>"subtractive synthesis" than conventional music synth usage.
>>>
>>>>The classical vocoder is a bank of filters applied to the
>>>>(single) input channel, each filter followed by some sort of
>>>>detector, which then controls the amplitude of an oscillator
>>>>(typically at the same frequency as the filter channel
>>>>center). That's pretty much exactly what the piano gadget
>>>>is doing.
>>>
>>>My belief is "vocoder" originally meant "voice encoder / decoder"
>>>(or something close to that) and there were at least two
>>>very common techniques, the phase vocoder being one, the others
>>>mostly based on linear prediction (which could be said to be
>>>"subtractive" but has nothing to do with "subtractive" synthesis
>>>in electronic music.
>>
>> Apologies to Randy... I should have checked Wikipedia
>> instead of relying on my ever-greying grey cells.
>
>Hi Bob,
>
>No problem. Been there, done that!
>
>> I was neglecting that the channel envelope followers in classical
>> vocoders controlled *filters*, not *oscillator* amplitude. So of
>> course the filter signal input is Randy's second input channel, and
>> the filtering would be the subtractive synthesis he mentioned.
>
>Yes, exactly! Now we're communicating.
>
>> [...]
>
>>>Anything similar to these voice coders also got called (by somebody..
>>>does anyone remember who?) "vocoders" in electronic music.
>>>But these could be broadly different from the original concepts.
>>>
>>>Steve
>>
>> Googling up on it now, it seems Moog and Carlos were using
>> the full vocoder concept Back In The Day. But I seem to
>> recall an "experimental" guy with a German or Eastern
>> European name. Hafta dig through my dead tree collection...
>
>When you say "Carlos" did you mean Wendy Carlos (the switched-on-Bach
>dude)?
The very same. In fact, I have that album from the actual
"dude" (Walter Carlos) days. Maybe it is a "collectible"
now?
>I was also trying to think of who Steve was referring to but came up
>blank.
>--
>Randy Yates
>Digital Signal Labs
>http://www.digitalsignallabs.com
----
Bob Masta
DAQARTA v7.60
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, Pitch Track, Pitch-to-MIDI
FREE Signal Generator, DaqMusiq generator
Science with your sound card!
Reply by robert bristow-johnson●June 15, 20152015-06-15
On 6/14/15 8:23 AM, Steve Pope wrote:
> Bob Masta<NoSpam@daqarta.com> wrote:
>
>> The classical vocoder is a bank of filters applied to the
>> (single) input channel, each filter followed by some sort of
>> detector, which then controls the amplitude of an oscillator
>> (typically at the same frequency as the filter channel
>> center). That's pretty much exactly what the piano gadget
>> is doing.
>
> My belief is "vocoder" originally meant "voice encoder / decoder"
> (or something close to that) and there were at least two
> very common techniques, the phase vocoder being one, the others
> mostly based on linear prediction (which could be said to be
> "subtractive" but has nothing to do with "subtractive" synthesis
> in electronic music.
>
> Anything similar to these voice coders also got called (by somebody..
> does anyone remember who?) "vocoders" in electronic music.
> But these could be broadly different from the original concepts.
thems would be called "channel vocoders" and are described by Bob above.
i might replace the word "classical" with "channel".
so there seems to me to be three different definitions or
implementations of what people call a "vocoder":
1. Phase vocoder (FFT, modify spectrum, iFFT)
2. LPC vocoder (autocorrelation, Levison-Durbin, re-jiggering
coefficients to Lattice or Ladder form)
3. Channel vocoder (filter bank for analysis hooked up to
another filter bank for synthesis)
--
r b-j rbj@audioimagination.com
"Imagination is more important than knowledge."
Reply by Randy Yates●June 15, 20152015-06-15
N0Spam@daqarta.com (Bob Masta) writes:
> On Sun, 14 Jun 2015 12:23:02 +0000 (UTC),
> spope33@speedymail.org (Steve Pope) wrote:
>
>>Bob Masta <NoSpam@daqarta.com> wrote:
>>
>>>On Sat, 13 Jun 2015 14:10:34 -0400, Randy Yates
>>
>>>>I wouldn't call it a vocoder. A vocoder uses subtractive synthesis,
>>>>while this technique uses additive synthesis (or you could call it a
>>>>type of additive synthesis). Also, a vocoder has two input channels,
>>>>while this only has one.
>>
>>>Apparently there must be vocoders out there that I haven't
>>>heard about, or you are using a different definition of
>>>"subtractive synthesis" than conventional music synth usage.
>>
>>>The classical vocoder is a bank of filters applied to the
>>>(single) input channel, each filter followed by some sort of
>>>detector, which then controls the amplitude of an oscillator
>>>(typically at the same frequency as the filter channel
>>>center). That's pretty much exactly what the piano gadget
>>>is doing.
>>
>>My belief is "vocoder" originally meant "voice encoder / decoder"
>>(or something close to that) and there were at least two
>>very common techniques, the phase vocoder being one, the others
>>mostly based on linear prediction (which could be said to be
>>"subtractive" but has nothing to do with "subtractive" synthesis
>>in electronic music.
>
> Apologies to Randy... I should have checked Wikipedia
> instead of relying on my ever-greying grey cells.
Hi Bob,
No problem. Been there, done that!
> I was neglecting that the channel envelope followers in classical
> vocoders controlled *filters*, not *oscillator* amplitude. So of
> course the filter signal input is Randy's second input channel, and
> the filtering would be the subtractive synthesis he mentioned.
Yes, exactly! Now we're communicating.
> [...]
>>Anything similar to these voice coders also got called (by somebody..
>>does anyone remember who?) "vocoders" in electronic music.
>>But these could be broadly different from the original concepts.
>>
>>Steve
>
> Googling up on it now, it seems Moog and Carlos were using
> the full vocoder concept Back In The Day. But I seem to
> recall an "experimental" guy with a German or Eastern
> European name. Hafta dig through my dead tree collection...
When you say "Carlos" did you mean Wendy Carlos (the switched-on-Bach
dude)?
I was also trying to think of who Steve was referring to but came up
blank.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com
Reply by Bob Masta●June 15, 20152015-06-15
On Sun, 14 Jun 2015 12:23:02 +0000 (UTC),
spope33@speedymail.org (Steve Pope) wrote:
>Bob Masta <NoSpam@daqarta.com> wrote:
>
>>On Sat, 13 Jun 2015 14:10:34 -0400, Randy Yates
>
>>>I wouldn't call it a vocoder. A vocoder uses subtractive synthesis,
>>>while this technique uses additive synthesis (or you could call it a
>>>type of additive synthesis). Also, a vocoder has two input channels,
>>>while this only has one.
>
>>Apparently there must be vocoders out there that I haven't
>>heard about, or you are using a different definition of
>>"subtractive synthesis" than conventional music synth usage.
>
>>The classical vocoder is a bank of filters applied to the
>>(single) input channel, each filter followed by some sort of
>>detector, which then controls the amplitude of an oscillator
>>(typically at the same frequency as the filter channel
>>center). That's pretty much exactly what the piano gadget
>>is doing.
>
>My belief is "vocoder" originally meant "voice encoder / decoder"
>(or something close to that) and there were at least two
>very common techniques, the phase vocoder being one, the others
>mostly based on linear prediction (which could be said to be
>"subtractive" but has nothing to do with "subtractive" synthesis
>in electronic music.
Apologies to Randy... I should have checked Wikipedia
instead of relying on my ever-greying grey cells. I was
neglecting that the channel envelope followers in classical
vocoders controlled *filters*, not *oscillator* amplitude.
So of course the filter signal input is Randy's second input
channel, and the filtering would be the subtractive
synthesis he mentioned.
So my proposed explanation for the piano gadget is really
just the front end of a vocoder, as far as the envelope
extractors, which drive the piano hammers.
In case anyone is interested, this "half-a-vocoder" is (or
was, when I was in the hearing research field) also the
basic front end of a cochlear implant driver. The incoming
sound is broken into several bands, whose envelopes drive
pulse generators attached to the implant electrodes. There
are various strategies for driving the electrodes, to try to
confine the generated fields to narrow regions of the
cochlea.
The problem is that the implant electrodes are like bands on
a long, thin rod that is snaked up the fluid-filled part of
the cochlea. Unfortunately, that's a distance from the
central core of the spiral where the hopefully-remaining
nerve fibers are. And the fluid the electrodes are in is
conductive (like seawater). The upshot is that in the best
of cases you can stimulate maybe 8 frequency regions, and
typically more like 5.
Note that the stimulus *location* is what controls the
perceived pitch, not the frequency of the stimulus pulses
(or not much, anyway), since the normal neuron firing rate
encodes the amplitude of the component at its given tuning
location along the cochlea.
So basically, the patient is hearing with a 5-band
representation of sound. This is *very* poor quality, even
for speech recognition. With practice, the patient may be
able to reach the "gold standard" of being able to
understand speech over a phone (no lip-reading). Music is
pretty much out of the question, though you would get timing
cues and maybe enough spectral cues to identify a piece you
already knew.
>Anything similar to these voice coders also got called (by somebody..
>does anyone remember who?) "vocoders" in electronic music.
>But these could be broadly different from the original concepts.
>
>Steve
Googling up on it now, it seems Moog and Carlos were using
the full vocoder concept Back In The Day. But I seem to
recall an "experimental" guy with a German or Eastern
European name. Hafta dig through my dead tree collection...
Best regards,
Bob Masta
DAQARTA v7.60
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, Pitch Track, Pitch-to-MIDI
FREE Signal Generator, DaqMusiq generator
Science with your sound card!
Reply by Bob Masta●June 14, 20152015-06-14
On Sat, 13 Jun 2015 14:10:34 -0400, Randy Yates
<yates@digitalsignallabs.com> wrote:
>N0Spam@daqarta.com (Bob Masta) writes:
>> [...]
>> This is pretty much a vocoder
>
>I wouldn't call it a vocoder. A vocoder uses subtractive synthesis,
>while this technique uses additive synthesis (or you could call it a
>type of additive synthesis). Also, a vocoder has two input channels,
>while this only has one.
Apparently there must be vocoders out there that I haven't
heard about, or you are using a different definition of
"subtractive synthesis" than conventional music synth usage.
The classical vocoder is a bank of filters applied to the
(single) input channel, each filter followed by some sort of
detector, which then controls the amplitude of an oscillator
(typically at the same frequency as the filter channel
center). That's pretty much exactly what the piano gadget
is doing.
Subtractive synthesis (in music synths, anyway) involves
filters that operate on a harmonic-rich waveform such as a
sawtooth. The sawtooth frequency (typically controlled by a
keyboard) sets the fundamental, and the filter
characteristics control the timbre by removing some of the
harmonics. Subtractive synthesis filters are almost always
dynamic, typically with cutoff and/or Q controlled by an
envelope generator that is triggered by the key press that
set the oscillator frequency.
Best regards,
Bob Masta
DAQARTA v7.60
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, Pitch Track, Pitch-to-MIDI
FREE Signal Generator, DaqMusiq generator
Science with your sound card!
Reply by Richard Dobson●June 14, 20152015-06-14
On 13/06/2015 19:10, Randy Yates wrote:
> N0Spam@daqarta.com (Bob Masta) writes:
>> [...]
>> This is pretty much a vocoder
>
> I wouldn't call it a vocoder. A vocoder uses subtractive synthesis,
> while this technique uses additive synthesis (or you could call it a
> type of additive synthesis). Also, a vocoder has two input channels,
> while this only has one.
>
It can be a useful hand-wavy comparison sometimes. The input data is the
dynamically changing quantised spectral envelope, which is then directly
quasi-synthesised. The phase vocoder (as distinct from the [analogue]
channel vocoder) operates in a broadly similar way - the analysis stage
(of course using the FFT, over many more points) is/can be followed by
[processing ad lib, then] a resynthesis oscillator bank. The piano
corresponds to a ~sort of~ constant-Q oscillator bank, where, obviously,
each oscillator is (approximately)an et12 piano note.
Just about everything in our world is a quasi, sort-of, hand-wavy
approximate thing; except where it isn't.
Richard Dobson
Reply by Randy Yates●June 13, 20152015-06-13
N0Spam@daqarta.com (Bob Masta) writes:
> [...]
> This is pretty much a vocoder
I wouldn't call it a vocoder. A vocoder uses subtractive synthesis,
while this technique uses additive synthesis (or you could call it a
type of additive synthesis). Also, a vocoder has two input channels,
while this only has one.
--
Randy Yates
Digital Signal Labs
http://www.digitalsignallabs.com
Reply by Bob Masta●June 13, 20152015-06-13
On Fri, 12 Jun 2015 20:38:30 -0500, Les Cargill
<lcargill99@comcast.com> wrote:
>Randy Yates wrote:
>> Randy Yates <yates@digitalsignallabs.com> writes:
>>
>>> Les Cargill <lcargill99@comcast.com> writes:
>>>
>>>> robert bristow-johnson wrote:
>>>>> On 6/6/15 1:34 PM, Les Cargill wrote:
>>>>>> Randy Yates wrote:
>>>>>>> http://createdigitalmusic.com/2009/10/the-speaking-piano-and-transforming-audio-to-midi/
>>>>>>>
>>>>>
>>>>> very interesting. thanks Randy.
>>>>>
>>>>>>
>>>>>> "...the original is almost certainly mixed in."
>>>>>
>>>>> who is saying that? i don't think that is the case.
>>>>
>>>>
>>>> The person who wrote the article at the link.
>>>>
>>>> " Edit: Listening again, the short answer to how you can hear so much
>>>> of the voice through the piano seems to be, you can’t; the original is
>>>> almost certainly mixed in. It’s nonetheless an interesting effect, and
>>>> I’d like to hear the piano on its own. "
>>>
>>> Les,
>>>
>>> Wasn't that just a comment from a reader?
>>
>> I see now it is not.
>>
>> This guy is just a writer.
>
>Correct.
>
>> I don't think he really knows what's going on
>> in the algorithm - just judging by ear (which could be wrong).
>>
>
>
>I have to say - my first reaction was that the voice was not
>actually being produced by the piano alone. I could easily be wrong,
>but the clicky stuff and the sonorous stuff don't sound "together",
>somehow.
>
>I don't know how you get from piano notes ( which are pretty
>complicated ) to formants. That's a protestation of ignorance,
>not a complaint :)
See my response of 6-7-15. The inventor talks about
"pixels", which I think is a big clue. Think about mapping
a spectrogram back to sound: The algorithm doesn't need to
know anything about formants, it just looks at each pixel
and creates a tone of that frequency and loudness, one
spectrogram column at a time. This is pretty much a
vocoder, with the difference that the output is mapped
logarithmically (to match the spacing of piano tones), and
that it drives mechanical actuators instead of oscillators.
And yes, piano notes are pretty complicated. But I think
that simply explains why the voice sounds "piano-y", with
lots of extra harmonics.
Best regards,
Bob Masta
DAQARTA v7.60
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, Pitch Track, Pitch-to-MIDI
FREE Signal Generator, DaqMusiq generator
Science with your sound card!
Reply by Les Cargill●June 12, 20152015-06-12
Richard Dobson wrote:
> On 12/06/2015 02:24, Les Cargill wrote:
>> robert bristow-johnson wrote:
>>> On 6/6/15 1:34 PM, Les Cargill wrote:
>>>> Randy Yates wrote:
>>>>> http://createdigitalmusic.com/2009/10/the-speaking-piano-and-transforming-audio-to-midi/
>>>>>
>>>>>
>>>>>
>>>
>>> very interesting. thanks Randy.
>>>
>>>>
>>>> "...the original is almost certainly mixed in."
>>>
>>> who is saying that? i don't think that is the case.
>>
>>
>> The person who wrote the article at the link.
>>
>> " Edit: Listening again, the short answer to how you can hear so much of
>> the voice through the piano seems to be, you can�t; the original is
>> almost certainly mixed in. It�s nonetheless an interesting effect, and
>> I�d like to hear the piano on its own. "
>>
>
>
> Using my musician's ear (and also because this example has been around
> for a while and pretty well known in computer music circles), I can
> assert that I most definitely do not hear any original mixed in - it's
> not. The piano is on its own. Sibilants are converted into clusters of
> (mostly) high notes, very familiar to anyone used to such things in
> various forms of contemporary music.
Which I am not - so my poor untrained brain made two things
out of it.
Cool. Goes back to what Randy was saying about the writer -
same deal here - I hear the voice separately from the piano.
Thanks for the post. Very interesting.
> The rest of the voice is
> represented "in the usual way" through the recognisability of the
> distinct formant regions of different vowels, and their voice-like
> manner of evolving over time. Our ears have been trained subliminally
> and biologically over millennia to be very sensitive to such things;
> much as certain combinations of shadows and shapes, in clouds, on the
> moon, in toast, may ineluctably suggest a human face.
>
> Richard Dobson
>
>
--
Les Cargill
Reply by Les Cargill●June 12, 20152015-06-12
Randy Yates wrote:
> Randy Yates <yates@digitalsignallabs.com> writes:
>
>> Les Cargill <lcargill99@comcast.com> writes:
>>
>>> robert bristow-johnson wrote:
>>>> On 6/6/15 1:34 PM, Les Cargill wrote:
>>>>> Randy Yates wrote:
>>>>>> http://createdigitalmusic.com/2009/10/the-speaking-piano-and-transforming-audio-to-midi/
>>>>>>
>>>>
>>>> very interesting. thanks Randy.
>>>>
>>>>>
>>>>> "...the original is almost certainly mixed in."
>>>>
>>>> who is saying that? i don't think that is the case.
>>>
>>>
>>> The person who wrote the article at the link.
>>>
>>> " Edit: Listening again, the short answer to how you can hear so much
>>> of the voice through the piano seems to be, you can’t; the original is
>>> almost certainly mixed in. It’s nonetheless an interesting effect, and
>>> I’d like to hear the piano on its own. "
>>
>> Les,
>>
>> Wasn't that just a comment from a reader?
>
> I see now it is not.
>
> This guy is just a writer.
Correct.
> I don't think he really knows what's going on
> in the algorithm - just judging by ear (which could be wrong).
>
I have to say - my first reaction was that the voice was not
actually being produced by the piano alone. I could easily be wrong,
but the clicky stuff and the sonorous stuff don't sound "together",
somehow.
I don't know how you get from piano notes ( which are pretty
complicated ) to formants. That's a protestation of ignorance,
not a complaint :)
It is, IMO, the limitations of that demo. I don't mean to
make a claim one way or the other.
--
Les Cargill