Hello all of you, First of all, I would like to thank the entire community for hours of both interesting commentary, helpful advice, and a product that is unrivaled (my opinion) to any other media player out there! Secondly, I hope that I could basically trouble you all for a little bit of help in answering a question. Recently, I have been playing around with OSS programming in Linux. It has gone rather well. My current problem revolves around the format of audio know as: AFMT_S16_LE Which is a little endian and is defaulted to 16 bit samples. I have been using a signed short to pull them out of a file after recording for analyzation. This has worked quite well indeed. I have svga graphing software I have written to handle and scale the values, it makes a pretty picture, but unfortunatly... I am having trouble conceptionalizing what it is supposed to mean. From the OSS site, the proposed meaning of a sample is the volume of the signal when it is measured. The thing is... How does one have negative volume? The threshold I imagine is zero, or is it that I am simply looking at it wrong and that some other value is the threshold that I should be concerned with? Is there a way I can relate this value to frequancy? Any help anyone could offer in this direction would be most helpful... Jon Larabee
On Sun, Feb 09, 2003 at 10:19:19PM +0000, Jon Larabee wrote:
Which is a little endian and is defaulted to 16 bit samples. I have been using a signed short to pull them out of a file after recording for analyzation. This has worked quite well indeed. I have svga graphing software I have written to handle and scale the values, it makes a pretty picture, but unfortunatly... I am having trouble conceptionalizing what it is supposed to mean. From the OSS site, the proposed meaning of a sample is the volume of the signal when it is measured. The thing is... How does one have negative volume? The threshold I imagine is zero, or is it that I
Well volume is a bad word for it; the best word is probably amplitude. It can be positive or negative because voltage can be positive or negative (i.e. higher or lower potential in the signal line than in the ground) and because pressure on your eardrum can be inward (above the pressure inside your ear) or outward (below internal pressure) as the vibration of the sound goes back and forth. Which direction is which isn't terribly important.
am simply looking at it wrong and that some other value is the threshold that I should be concerned with? Is there a way I can relate this value to frequancy? Any help anyone could offer in this direction would be most
This is a very good question. There is no direct way to get the frequency since the concept isn't even really well defined. Chopping up the sound into blocks and applying a fourier transform of some sort is the most common way I know of extracting frequency information. There are also various filters and wavelet type techniques I don't really know much about. Anyway, if you're seriously planning on analyzing sound like you say, you really need to get yourself a good text on the subject and read up... Good luck! Rich
Hey, Apperciate the help... A few things I was wondering: 1) What is the threshold for the 16 bit value? Is it supposed to be zero? 2) Is the frequancy that it is sampling at always going to be the same? If so could I do the following for better plotting? Amplitude (the sample scaled by some factor) * cos(2*pi*(frequancy of sampling rate * time) = y In which case I would assume that the frequancy it is sampling at would never change ie it being the sampling rate, and that the time can be determined by the position in the file (ie 22,000 hz/second rate is basically 22,000 samples / second) I don't know if this would work, and you sound like you know a heck of a lot more about it than I do ;) On Sun, 9 Feb 2003, D Richard Felker III wrote:
[Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html] On Sun, Feb 09, 2003 at 10:19:19PM +0000, Jon Larabee wrote:
Which is a little endian and is defaulted to 16 bit samples. I have been using a signed short to pull them out of a file after recording for analyzation. This has worked quite well indeed. I have svga graphing software I have written to handle and scale the values, it makes a pretty picture, but unfortunatly... I am having trouble conceptionalizing what it is supposed to mean. From the OSS site, the proposed meaning of a sample is the volume of the signal when it is measured. The thing is... How does one have negative volume? The threshold I imagine is zero, or is it that I
Well volume is a bad word for it; the best word is probably amplitude. It can be positive or negative because voltage can be positive or negative (i.e. higher or lower potential in the signal line than in the ground) and because pressure on your eardrum can be inward (above the pressure inside your ear) or outward (below internal pressure) as the vibration of the sound goes back and forth. Which direction is which isn't terribly important.
am simply looking at it wrong and that some other value is the threshold that I should be concerned with? Is there a way I can relate this value to frequancy? Any help anyone could offer in this direction would be most
This is a very good question. There is no direct way to get the frequency since the concept isn't even really well defined. Chopping up the sound into blocks and applying a fourier transform of some sort is the most common way I know of extracting frequency information. There are also various filters and wavelet type techniques I don't really know much about.
Anyway, if you're seriously planning on analyzing sound like you say, you really need to get yourself a good text on the subject and read up...
Good luck!
Rich
_______________________________________________ RTFM!!! http://www.MPlayerHQ.hu/DOCS Search: http://www.MPlayerHQ.hu/cgi-bin/htsearch http://mplayerhq.hu/mailman/listinfo/mplayer-users
Obiously my 2nd assertion is flawed. I can't do that because the sampling rate is the RATE of looking at the signal, not the frequacny of the signal, which obviously does not stay the same. Disregard that question. Your quite right... any good resources on fourier transformations anyone? ;) Jon On Sun, 9 Feb 2003, Jon Larabee wrote:
[Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html] Hey,
Apperciate the help... A few things I was wondering:
1) What is the threshold for the 16 bit value? Is it supposed to be zero?
2) Is the frequancy that it is sampling at always going to be the same? If so could I do the following for better plotting?
Amplitude (the sample scaled by some factor) * cos(2*pi*(frequancy of sampling rate * time) = y
In which case I would assume that the frequancy it is sampling at would never change ie it being the sampling rate, and that the time can be determined by the position in the file (ie 22,000 hz/second rate is basically 22,000 samples / second) I don't know if this would work, and you sound like you know a heck of a lot more about it than I do ;)
On Sun, 9 Feb 2003, D Richard Felker III wrote:
[Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html] On Sun, Feb 09, 2003 at 10:19:19PM +0000, Jon Larabee wrote:
Which is a little endian and is defaulted to 16 bit samples. I have been using a signed short to pull them out of a file after recording for analyzation. This has worked quite well indeed. I have svga graphing software I have written to handle and scale the values, it makes a pretty picture, but unfortunatly... I am having trouble conceptionalizing what it is supposed to mean. From the OSS site, the proposed meaning of a sample is the volume of the signal when it is measured. The thing is... How does one have negative volume? The threshold I imagine is zero, or is it that I
Well volume is a bad word for it; the best word is probably amplitude. It can be positive or negative because voltage can be positive or negative (i.e. higher or lower potential in the signal line than in the ground) and because pressure on your eardrum can be inward (above the pressure inside your ear) or outward (below internal pressure) as the vibration of the sound goes back and forth. Which direction is which isn't terribly important.
am simply looking at it wrong and that some other value is the threshold that I should be concerned with? Is there a way I can relate this value to frequancy? Any help anyone could offer in this direction would be most
This is a very good question. There is no direct way to get the frequency since the concept isn't even really well defined. Chopping up the sound into blocks and applying a fourier transform of some sort is the most common way I know of extracting frequency information. There are also various filters and wavelet type techniques I don't really know much about.
Anyway, if you're seriously planning on analyzing sound like you say, you really need to get yourself a good text on the subject and read up...
Good luck!
Rich
_______________________________________________ RTFM!!! http://www.MPlayerHQ.hu/DOCS Search: http://www.MPlayerHQ.hu/cgi-bin/htsearch http://mplayerhq.hu/mailman/listinfo/mplayer-users
_______________________________________________ RTFM!!! http://www.MPlayerHQ.hu/DOCS Search: http://www.MPlayerHQ.hu/cgi-bin/htsearch http://mplayerhq.hu/mailman/listinfo/mplayer-users
participants (2)
-
D Richard Felker III -
Jon Larabee