[[email protected]]

You have all been talking about it, but now you can download example
sound files I've been working with in photoshop using wav2bmp. I have
a screenshot of the program in action, but I want to make sure there
aren't any bugs before I release the software as a download.

I'm just discovering the limitations of using photoshop as an audio
tool. I have been able to begin blending two sound files made with a
vocoder in photoshop to create a fresh new mix. But you have to turn
up the volume a little, and I'm trying to find a way to minimize the
static. If you can help, or want to download the software let me know.
It is really novel and fun.

http://grimoire.genesismuds.com

[[email protected]]

I've just uploaded an example sound file without the static. Listen
and enjoy...

wrote:
You have all been talking about it, but now you can download example
sound files I've been working with in photoshop using wav2bmp. I have
a screenshot of the program in action, but I want to make sure there
aren't any bugs before I release the software as a download.

I'm just discovering the limitations of using photoshop as an audio
tool. I have been able to begin blending two sound files made with a
vocoder in photoshop to create a fresh new mix. But you have to turn
up the volume a little, and I'm trying to find a way to minimize the
static. If you can help, or want to download the software let me know.
It is really novel and fun.

http://grimoire.genesismuds.com

[Ethan Winer]

Corey,

I'm just discovering the limitations of using photoshop as an audio tool.


I bet you are! :-)

--Ethan

[David Carter]

Ethan Winer wrote:
Corey,

I'm just discovering the limitations of using photoshop as an audio tool.


I bet you are! :-)

--Ethan




No chance this stems from the fact that Photoshop isnt an audio editor I
suppose?

Carter

[[email protected]]

Yes that is an issue, but it is a great blackboard to begin designing
algorithms and experimenting with the raw bytes of sound. Right?

Version 2.0 is almost ready to be released, and my team has been
working out all of the buggs. And If you check out the site you can
listen to some cool effects I've been producing with Wav2Bmp.

http://grimoire.genesismuds.com

David Carter wrote:
Ethan Winer wrote:
Corey,

I'm just discovering the limitations of using photoshop as an audio tool.


I bet you are! :-)

--Ethan




No chance this stems from the fact that Photoshop isnt an audio editor I
suppose?

Carter

[martin griffith]

On Sat, 2 Sep 2006 11:11:25 -0400, in rec.audio.pro "Ethan Winer"
ethanw at ethanwiner dot com wrote:

Corey,

I'm just discovering the limitations of using photoshop as an audio tool.


I bet you are! :-)

--Ethan

Yeah, I find that the London Underground map is really useful in
Moscow


martin

[[email protected]]

wrote:
I'm just discovering the limitations of using photoshop as an
audio tool.

Any number of responses come to mind:

* It's a poor workman that blames his tools,

* Just becuase the only tool you have is a hammer,
that doesn't mean every problem is a nail,

* You'll put your eye out if you're not careful, and then
it won't be so funny any more.

* Whenever someone says, "Hey guys, watch this,"
be prepared for someone to get hurt.

Yes that is an issue, but it is a great blackboard to begin designing
algorithms and experimenting with the raw bytes of sound. Right?

Wrong.

It's no more correct than using an audio editor to manipulate
pictures.

Your premise is so incompletely and incorrectly formed
as to make it completely arbitrary the resulting representations.
For example, consider that sound, it its most fundamental
physical level, is a two-dimensional function of amplitude
vs time, whereas an image is at least a thre-dimensional
domain of position (two dimensions) and color.

Your arbitrary choice of mapping completely changes the
representation. Consider, say, 1 million consecutive
samples: do you map them to 1000x1000 pixels, or
5000X200 pixels, or 4000x250 pixels? Any choice you
make dramatically changes how ANY patterns in the
image appear.

You've further confused a number of well-understood
fundamental terms for example, you ask "what is the
frequency of a sample." A sample has no, indeed, it
cannot have a frequency. It has one very specific property
and one property alone: normalized amplitude.

Perhaps you are confusing your rather misguided efforts
with what you have seen elsewhere, referred to as "sono-
grams," which is a pseudo-3 dimensional mapping of
time (x axis), frequency (y axiz) and amplitude (z-axis,
often mapped, for representation, to grey scale). This is
produced by essentially windowing the data, taking a
spectrogram of that windowed dat. That gives you a
smpashot of the specturm of the sound over the interval
of that window. The window is then moved a little farther
along in time and the process is done again.

The problem with this process is that the very nature of
windowing is that you can end up with one particularly
unique visual representation that may have an infinite
number of audible representations that wholely map
to it. As a tool for CREATING or manipulating sound,
it makes such a tool useless, because the results of
manipulation the image results in a sound which is far
more determined by the fixed and hidden assumptions
in the conversion algorithm than in the contents of the
image.

This reminds me of the proclivity among, especially, the
younger, less experienced users of audio editing systems
who "grew up" using waveform-based editing systems,
and were forced into environments where they actually
had to (horrors!) use their ears as the primary editing tool.
They would cry "How can I edit the waveform if I can't see
the stuff I can't hear?" The answer is, "if you can't hear it,
why bother trying to edit it?"

Be all that as it may, good luck on your little project, though
I suspect you'll find that it will, at its very best, fall far short
of what you think it can do.

Version 2.0 is almost ready to be released, and my team has been
working out all of the buggs. And If you check out the site you can
listen to some cool effects I've been producing with Wav2Bmp.

I'm sure your team is working out all the bugs save one,
which they will never be able to fix, and that is in the original
concept. Any reasonably competent software team can
implement any design, no matter how ill-conceived the
design is.

[Don Pearce]

On 3 Sep 2006 09:25:45 -0700, wrote:

Version 2.0 is almost ready to be released, and my team has been
working out all of the buggs. And If you check out the site you can
listen to some cool effects I've been producing with Wav2Bmp.

I'm sure your team is working out all the bugs save one,
which they will never be able to fix, and that is in the original
concept. Any reasonably competent software team can
implement any design, no matter how ill-conceived the
design is.

Dick you have missed a very important point. You are dealing with a
man who seriously believes in fairies.

d

--
Pearce Consulting
http://www.pearce.uk.com

[[email protected]]

All good points that I've thought about, but I have version 2.0 of the
software developed and it works just fine. I'm doing all sorts of cool
stuff with it, and having a good time. In the next version maybe I
will allow the user to specify how the dimensions of the image relate
to time. But it is fairly arbitrary when doing the kind of work I am
doing in photoshop. Thanks, and if you want a copy of the software let
me know.



wrote:
wrote:
I'm just discovering the limitations of using photoshop as an
audio tool.

Any number of responses come to mind:

* It's a poor workman that blames his tools,

* Just becuase the only tool you have is a hammer,
that doesn't mean every problem is a nail,

* You'll put your eye out if you're not careful, and then
it won't be so funny any more.

* Whenever someone says, "Hey guys, watch this,"
be prepared for someone to get hurt.

Yes that is an issue, but it is a great blackboard to begin designing
algorithms and experimenting with the raw bytes of sound. Right?

Wrong.

It's no more correct than using an audio editor to manipulate
pictures.

Your premise is so incompletely and incorrectly formed
as to make it completely arbitrary the resulting representations.
For example, consider that sound, it its most fundamental
physical level, is a two-dimensional function of amplitude
vs time, whereas an image is at least a thre-dimensional
domain of position (two dimensions) and color.

Your arbitrary choice of mapping completely changes the
representation. Consider, say, 1 million consecutive
samples: do you map them to 1000x1000 pixels, or
5000X200 pixels, or 4000x250 pixels? Any choice you
make dramatically changes how ANY patterns in the
image appear.

You've further confused a number of well-understood
fundamental terms for example, you ask "what is the
frequency of a sample." A sample has no, indeed, it
cannot have a frequency. It has one very specific property
and one property alone: normalized amplitude.

Perhaps you are confusing your rather misguided efforts
with what you have seen elsewhere, referred to as "sono-
grams," which is a pseudo-3 dimensional mapping of
time (x axis), frequency (y axiz) and amplitude (z-axis,
often mapped, for representation, to grey scale). This is
produced by essentially windowing the data, taking a
spectrogram of that windowed dat. That gives you a
smpashot of the specturm of the sound over the interval
of that window. The window is then moved a little farther
along in time and the process is done again.

The problem with this process is that the very nature of
windowing is that you can end up with one particularly
unique visual representation that may have an infinite
number of audible representations that wholely map
to it. As a tool for CREATING or manipulating sound,
it makes such a tool useless, because the results of
manipulation the image results in a sound which is far
more determined by the fixed and hidden assumptions
in the conversion algorithm than in the contents of the
image.

This reminds me of the proclivity among, especially, the
younger, less experienced users of audio editing systems
who "grew up" using waveform-based editing systems,
and were forced into environments where they actually
had to (horrors!) use their ears as the primary editing tool.
They would cry "How can I edit the waveform if I can't see
the stuff I can't hear?" The answer is, "if you can't hear it,
why bother trying to edit it?"

Be all that as it may, good luck on your little project, though
I suspect you'll find that it will, at its very best, fall far short
of what you think it can do.

Version 2.0 is almost ready to be released, and my team has been
working out all of the buggs. And If you check out the site you can
listen to some cool effects I've been producing with Wav2Bmp.

I'm sure your team is working out all the bugs save one,
which they will never be able to fix, and that is in the original
concept. Any reasonably competent software team can
implement any design, no matter how ill-conceived the
design is.

[[email protected]]

Don Pearce wrote:

Dick you have missed a very important point. You are dealing with a
man who seriously believes in fairies.

No, I didn't miss the point. I was walking along
with my gun, saw a barrel of fish, ...

[[email protected]]

This is the next step of the program:

http://en.wikipedia.org/wiki/Spectrogram

wrote:
All good points that I've thought about, but I have version 2.0 of the
software developed and it works just fine. I'm doing all sorts of cool
stuff with it, and having a good time. In the next version maybe I
will allow the user to specify how the dimensions of the image relate
to time. But it is fairly arbitrary when doing the kind of work I am
doing in photoshop. Thanks, and if you want a copy of the software let
me know.



wrote:
wrote:
I'm just discovering the limitations of using photoshop as an
audio tool.

Any number of responses come to mind:

* It's a poor workman that blames his tools,

* Just becuase the only tool you have is a hammer,
that doesn't mean every problem is a nail,

* You'll put your eye out if you're not careful, and then
it won't be so funny any more.

* Whenever someone says, "Hey guys, watch this,"
be prepared for someone to get hurt.

Yes that is an issue, but it is a great blackboard to begin designing
algorithms and experimenting with the raw bytes of sound. Right?

Wrong.

It's no more correct than using an audio editor to manipulate
pictures.

Your premise is so incompletely and incorrectly formed
as to make it completely arbitrary the resulting representations.
For example, consider that sound, it its most fundamental
physical level, is a two-dimensional function of amplitude
vs time, whereas an image is at least a thre-dimensional
domain of position (two dimensions) and color.

Your arbitrary choice of mapping completely changes the
representation. Consider, say, 1 million consecutive
samples: do you map them to 1000x1000 pixels, or
5000X200 pixels, or 4000x250 pixels? Any choice you
make dramatically changes how ANY patterns in the
image appear.

You've further confused a number of well-understood
fundamental terms for example, you ask "what is the
frequency of a sample." A sample has no, indeed, it
cannot have a frequency. It has one very specific property
and one property alone: normalized amplitude.

Perhaps you are confusing your rather misguided efforts
with what you have seen elsewhere, referred to as "sono-
grams," which is a pseudo-3 dimensional mapping of
time (x axis), frequency (y axiz) and amplitude (z-axis,
often mapped, for representation, to grey scale). This is
produced by essentially windowing the data, taking a
spectrogram of that windowed dat. That gives you a
smpashot of the specturm of the sound over the interval
of that window. The window is then moved a little farther
along in time and the process is done again.

The problem with this process is that the very nature of
windowing is that you can end up with one particularly
unique visual representation that may have an infinite
number of audible representations that wholely map
to it. As a tool for CREATING or manipulating sound,
it makes such a tool useless, because the results of
manipulation the image results in a sound which is far
more determined by the fixed and hidden assumptions
in the conversion algorithm than in the contents of the
image.

This reminds me of the proclivity among, especially, the
younger, less experienced users of audio editing systems
who "grew up" using waveform-based editing systems,
and were forced into environments where they actually
had to (horrors!) use their ears as the primary editing tool.
They would cry "How can I edit the waveform if I can't see
the stuff I can't hear?" The answer is, "if you can't hear it,
why bother trying to edit it?"

Be all that as it may, good luck on your little project, though
I suspect you'll find that it will, at its very best, fall far short
of what you think it can do.

Version 2.0 is almost ready to be released, and my team has been
working out all of the buggs. And If you check out the site you can
listen to some cool effects I've been producing with Wav2Bmp.

I'm sure your team is working out all the bugs save one,
which they will never be able to fix, and that is in the original
concept. Any reasonably competent software team can
implement any design, no matter how ill-conceived the
design is.

[philicorda]

On Sun, 03 Sep 2006 11:46:52 -0700, CoreyWhite wrote:

This is the next step of the program:

http://en.wikipedia.org/wiki/Spectrogram

Do that and you have reinvented metasynth.

What you are doing already is kinda more interesting.

I suspect you are coming up against the same challenges that I did, namely
that there is not a lot you can do to the raw audio data in the graphics
editor without making really nasty noises. (Though you may have a
different idea of what a nasty noise is

The idea of relating the image size to time sounds like it will lead
somewhere. Perhaps relating it to pitch would be useful.
Ie, drawing a 25 pixel wide vertical line in a picture 100 pixels wide
should give you a 25% pulse wave a roughly 440hz. (At 44100 samples per
second.)

If the user could change the width of the displayed picture while editing
it, they could select different pitches and draw them in. To do this would
require writing your own graphics editor, which is why I suggested it in
an earlier post. (You want to be able to change the width of the displayed
picture without interpolating or altering the data at all.)

It would also be handy if the graphical editor would let you hear the
changes in real time, without having to save and convert to wav again.

Not as hard as it looks as bitmap editors are quite straightforward. It
looks to me as though most of the useful operations would be a tiny subset
of what photoshop does, and work in a way somewhat different to how a
conventional graphics editor operates.

[[email protected]]

Connect to the grimoire to listen to my new sound sample. I have
perfected my technique and developed a sound file that is absolutely
perfect!

http://grimoire.genesismuds.com/perfect.wav


wrote:
This is the next step of the program:

http://en.wikipedia.org/wiki/Spectrogram

wrote:
All good points that I've thought about, but I have version 2.0 of the
software developed and it works just fine. I'm doing all sorts of cool
stuff with it, and having a good time. In the next version maybe I
will allow the user to specify how the dimensions of the image relate
to time. But it is fairly arbitrary when doing the kind of work I am
doing in photoshop. Thanks, and if you want a copy of the software let
me know.



wrote:
wrote:
I'm just discovering the limitations of using photoshop as an
audio tool.

Any number of responses come to mind:

* It's a poor workman that blames his tools,

* Just becuase the only tool you have is a hammer,
that doesn't mean every problem is a nail,

* You'll put your eye out if you're not careful, and then
it won't be so funny any more.

* Whenever someone says, "Hey guys, watch this,"
be prepared for someone to get hurt.

Yes that is an issue, but it is a great blackboard to begin designing
algorithms and experimenting with the raw bytes of sound. Right?

Wrong.

It's no more correct than using an audio editor to manipulate
pictures.

Your premise is so incompletely and incorrectly formed
as to make it completely arbitrary the resulting representations.
For example, consider that sound, it its most fundamental
physical level, is a two-dimensional function of amplitude
vs time, whereas an image is at least a thre-dimensional
domain of position (two dimensions) and color.

Your arbitrary choice of mapping completely changes the
representation. Consider, say, 1 million consecutive
samples: do you map them to 1000x1000 pixels, or
5000X200 pixels, or 4000x250 pixels? Any choice you
make dramatically changes how ANY patterns in the
image appear.

You've further confused a number of well-understood
fundamental terms for example, you ask "what is the
frequency of a sample." A sample has no, indeed, it
cannot have a frequency. It has one very specific property
and one property alone: normalized amplitude.

Perhaps you are confusing your rather misguided efforts
with what you have seen elsewhere, referred to as "sono-
grams," which is a pseudo-3 dimensional mapping of
time (x axis), frequency (y axiz) and amplitude (z-axis,
often mapped, for representation, to grey scale). This is
produced by essentially windowing the data, taking a
spectrogram of that windowed dat. That gives you a
smpashot of the specturm of the sound over the interval
of that window. The window is then moved a little farther
along in time and the process is done again.

The problem with this process is that the very nature of
windowing is that you can end up with one particularly
unique visual representation that may have an infinite
number of audible representations that wholely map
to it. As a tool for CREATING or manipulating sound,
it makes such a tool useless, because the results of
manipulation the image results in a sound which is far
more determined by the fixed and hidden assumptions
in the conversion algorithm than in the contents of the
image.

This reminds me of the proclivity among, especially, the
younger, less experienced users of audio editing systems
who "grew up" using waveform-based editing systems,
and were forced into environments where they actually
had to (horrors!) use their ears as the primary editing tool.
They would cry "How can I edit the waveform if I can't see
the stuff I can't hear?" The answer is, "if you can't hear it,
why bother trying to edit it?"

Be all that as it may, good luck on your little project, though
I suspect you'll find that it will, at its very best, fall far short
of what you think it can do.

Version 2.0 is almost ready to be released, and my team has been
working out all of the buggs. And If you check out the site you can
listen to some cool effects I've been producing with Wav2Bmp.

I'm sure your team is working out all the bugs save one,
which they will never be able to fix, and that is in the original
concept. Any reasonably competent software team can
implement any design, no matter how ill-conceived the
design is.

[Roy W. Rising]

wrote:
Connect to the grimoire to listen to my new sound sample. I have
perfected my technique and developed a sound file that is absolutely
perfect!

Perfectly what? It sounds like a primative "robot voice" from the 1950s!

--
~ Roy
"If you notice the sound, it's wrong!"

[Ron Capik]

"Roy W. Rising" wrote:

wrote:
Connect to the grimoire to listen to my new sound sample. I have
perfected my technique and developed a sound file that is absolutely
perfect!

Perfectly what? It sounds like a primative "robot voice" from the 1950s!

--
~ Roy
"If you notice the sound, it's wrong!"

....and all the "t"s are missing.


Later...

Ron Capik
--

[[email protected]]

Corey White...

Can you repost the link to your software ?

Although the Wikipedia page is fascinating and the stealing away of
images in a waveform if so Hermetic that it would probably have Egyptian
mystics wetting their pants...I can't help but think there's a more
elegant way.

Metasynth is basically producing an SSTV like signal (I think), but in
the audio range...the bottom line of pixels is the lower frequency
, the next line up is the next frequency range up the spectral waveform
image in your sound software. All the way up to the higher frequencies.

Although interesting...my heart always jumps when I see images of cats
in a spectral readout! It's NOT how an image sounds...right? Because
reproducing the image depends on how spectral readouts work...and that
you have to read your image by each line of pixels.

More interesting would be to look at an image, a photo as what it
actually IS which is a record of wavelengths of light. If you feed
your, say, digital camera image into ENVI (for example)...a high end
satellite spectroscopy software...you can actually categorise the
frequencies of light in the image (this is used to identify minerals
in sat images). ENVI does this by comparing the RGB channels, which
are, of course at KNOWN frequencies.

Now shift those frequencies into the AUDIO range and...well I hav'nt
tried it. What Philcordia is saying is interesting...there's a problem
translating between the different dimensions of sound and image.

I wonder how a Miss Piggy image would sound ?

http://www.angelfire.com/emo/on/artwork.html

DJ Barney

philicorda wrote:
On Sun, 03 Sep 2006 11:46:52 -0700, CoreyWhite wrote:


This is the next step of the program:

http://en.wikipedia.org/wiki/Spectrogram


Do that and you have reinvented metasynth.

What you are doing already is kinda more interesting.

I suspect you are coming up against the same challenges that I did, namely
that there is not a lot you can do to the raw audio data in the graphics
editor without making really nasty noises. (Though you may have a
different idea of what a nasty noise is

The idea of relating the image size to time sounds like it will lead
somewhere. Perhaps relating it to pitch would be useful.
Ie, drawing a 25 pixel wide vertical line in a picture 100 pixels wide
should give you a 25% pulse wave a roughly 440hz. (At 44100 samples per
second.)

If the user could change the width of the displayed picture while editing
it, they could select different pitches and draw them in. To do this would
require writing your own graphics editor, which is why I suggested it in
an earlier post. (You want to be able to change the width of the displayed
picture without interpolating or altering the data at all.)

It would also be handy if the graphical editor would let you hear the
changes in real time, without having to save and convert to wav again.

Not as hard as it looks as bitmap editors are quite straightforward. It
looks to me as though most of the useful operations would be a tiny subset
of what photoshop does, and work in a way somewhat different to how a
conventional graphics editor operates.