is this doable on photosounder?

Official support for: photosounder.com
Post Reply New Topic
RELATED
PRODUCTS

Post

hi

I don't own photosounder, and the trial version doesn't allow me to do a test.

but I want to know (from the creator or its users), if you can open a random complex picture in photosounder, create a sound of it, and later open that "image created" audio file in photosounder and retrieve something really similar to the original picture (I mean, perhapps a little distorted or so, but really close).

and what about colors?

thanks for your replies

Post

Well, Photosounder isn't meant to make something that looks good. Or rather, it can make something that looks good if you synthesise the image with a linear frequency scale, however as it is Photosounder is incapable of analysing a sound back in linear scale, but you can use just about any program that displays spectrograms to see this. This is demonstrated in this video ( http://www.youtube.com/watch?v=S64FROErFYA ) the image is synthesised into a sound in Photosounder, and the sound is displayed into an image that resembles the original in Spectrogram 16.

If you want to do something that looks right when you open the sound in Photosounder, which I don't recommend, you can synthesise it with a logarithmic frequency scale, but when you'll open the synthesised sound it will look blurry beyond recognition in the lower half of the image.

So I recommend you do it the way shown in the video, using another program to view it. If you want to make sure the intensities are rendered as closely as possible as the original I suggest you use ARSS for visualisation. It is the free open source precursor to Photosounder, and offers the possibility to analyse with a linear scale. See http://arss.sourceforge.net/examples.sh ... ansmission

Just out of curiosity, what do you want to use this for?

EDIT : As for colours, you'd either have to use three audio channels, or perhaps more simply just serialise the 3 colour channels, that is, play them one after the other.
Developer of Photosounder (a spectral editor/synth), SplineEQ and Spiral

Post

hi A_SN

thank you very much for your reply.

I'm an electronic artist, and I'm looking for the way to transmit an image as sound through air.

for example, recording that sound on an audio cd, cassette or other audio media (I prefer an analog one) and that being played by speakers. the sound being captured by a microphone attached to a computer that puts together the image (as close as possible). so you can hear and see the same information (a kind of synesthesia).

I'm researching for a serial binary audio data transmission (the way datassettes used to store binary files on audio cassettes) and also this kind of way.

the first can retrieve the exact same image, but it has a bigger chance to be affected by noise and such (parts of the image can be lost, even do that is not a big issue for me) and the sound is not very appealing or distinguishly.

the second can be less affected by noises, and the sound has a much more musical tone. but I can't retrieve colors (I think), and perhapps building a custom system is really difficult, I don't know.

what do you think?

thanks in advance

edit: also, can you please let me have the Spectrogram 16 software? I know it was the last version and it's freeware, but I can't find it anywhere =(

Post

jjplano wrote:hi A_SN

thank you very much for your reply.

I'm an electronic artist, and I'm looking for the way to transmit an image as sound through air.

for example, recording that sound on an audio cd, cassette or other audio media (I prefer an analog one) and that being played by speakers. the sound being captured by a microphone attached to a computer that puts together the image (as close as possible). so you can hear and see the same information (a kind of synesthesia).

I'm researching for a serial binary audio data transmission (the way datassettes used to store binary files on audio cassettes) and also this kind of way.

the first can retrieve the exact same image, but it has a bigger chance to be affected by noise and such (parts of the image can be lost, even do that is not a big issue for me) and the sound is not very appealing or distinguishly.

the second can be less affected by noises, and the sound has a much more musical tone. but I can't retrieve colors (I think), and perhapps building a custom system is really difficult, I don't know.

what do you think?

thanks in advance

edit: also, can you please let me have the Spectrogram 16 software? I know it was the last version and it's freeware, but I can't find it anywhere =(
For transmitting through air and recovering with a microphone I definitely recommend doing it the way it was done on the video I showed you. As for colours, well you'd pretty much have to develop your own program for that, as all spectrograms are meant to be in grey scale (even if gradients are mapped onto it). If you want to transmit colours in some way, you have a few choices. A first choice is to send the three channels one after the others in time. It's nice because each channel will be effected with dark bands (due to the non-flat frequency response of the whole transmission process) at the same heights. Stacking up the channels vertically would probably more desirable though, as you wouldn't need to synchronise the channels in time, the three channels would already be in sync, and so you could even have a program to display colours live as the sound is received. One thing to consider though is that while you might want to go for RGB, the different dark band thing will probably mess up the colours quite a bit (although it might give you a funky kind of effect). It might be better if you use the Lab layers (from the Lab color space) instead.

A few things to consider : basically as a rule of thumb you shouldn't try to transmit more pixels per second (I'm talking about square pixels, as in, height * width) than there are Hertz in your bandwidth for a grey image. So for example if you have a bandwidth of 10000 Hz, and that you want to transmit an image of dimensions 320x240 (which is 76800 pixels), you should make the image's sound last at least 7.68 seconds. Or to put it another way, you can divide the bandwidth by the height (10000 Hz / 240 px = 41.7 px/second) to tell you which pixel/second you can use at most in Photosounder. But the nice thing here is that you can make it last longer to have a better resistance to noise (e.g. if you make it last 9 times longer it'll be 3 times less affected by noise). However it'll still be affected by dark bands, but you can adjust that if you pad your image horizontally with white columns so that it tells you how to correct for them. Also it'd be extra nice if in the image, at the beginning of the transmission you wrote in the spectrum the characteristics of the image (i.e. its precise frequency range, its vertical height, its pixels/second setting and if you're going for colour something that tells you which copy of the image is which channel, for example write RGB vertically if the topmost layer is Red) so as to make the sound itself more self explanatory to anyone who'd receive it without being told its characteristics.

Spectrogram 16 isn't all that anyway, you might want to try Baudline or just about whatever spectrograph software you find.
Developer of Photosounder (a spectral editor/synth), SplineEQ and Spiral

Post

A_SN, thanks for your reply.

I was doing some tests with mic and speaker, and it worked ok.

I appreciatte all the technical advise you're giving to me, as I really have no deep knowledge of this kind of matter.

perhapps in the near future I would send you an email to clarify some of this points.

thanks again.

Post

jjplano wrote:A_SN, thanks for your reply.

I was doing some tests with mic and speaker, and it worked ok.

I appreciatte all the technical advise you're giving to me, as I really have no deep knowledge of this kind of matter.

perhapps in the near future I would send you an email to clarify some of this points.

thanks again.
Cool, hey could you email me some pictures of your results? I don't have the gear to do anything like that (I have zero gear) and it'd be curious to see how different experiments look.
Developer of Photosounder (a spectral editor/synth), SplineEQ and Spiral

Post Reply

Return to “Photosounder”