Lame 3.88 beta released - Page 2

ricecake · 30th March 2001, 07:58

Quote:

Originally posted by Somebody
[B]2) it has higher frequencies recorded (this is why it makes it so good, you don't hear it, but you feel the crispness and richness of the music) (if you listen to a cd and a vinyl record, same album of course, and good equipment that does not make it crackle and shit, you will hear something missing from the cd, which the vinyl has)

The equipment has a _lot_ to do with the quality of the playback.

Quote:

3) that's it

Not true. It's much easier to scratch with vinyl :-)

Somebody · 30th March 2001, 19:32

Quote:

The equipment has a _lot_ to do with the quality of the playback.

when did I say equipment has nothing to do with quality?
I did say, good equipment that does NOT make it crackle and shit.
And yes, of course, it has a lot to do with quality.

Quote:

3) that's it

Not true. It's much easier to scratch with vinyl :-) [/QUOTE]
?

ricecake · 30th March 2001, 19:40

Quote:

Originally posted by Somebody
when did I say equipment has nothing to do with quality?
I did say, good equipment that does NOT make it crackle and shit.
And yes, of course, it has a lot to do with quality.

I'm not saying you didn't, just trying to point that fact out to the people who think that ALL vinyl is inferior to CDs.

Quote:

Not true. It's much easier to scratch with vinyl :-)

Quote:

?[/B]

Scratch as in the hip-hop DJ sense.

Somebody · 30th March 2001, 19:42

Sincere apologies,

I figured about the dj scratching thing.

But what I do not comprehend is where did you get the scratching from reason 3?

SNYder · 31st March 2001, 00:30

I actually knew about higher frequencys (out of human hearing ranges) affecting the feel of the music, but i had no idea vinyl could go higher then cd's. hmmm...

what is the frequency range for cd's? it must be more then 20,000, correct? i can't imagine them being limited to 20 to 20,000. And don't forget dvd audio wich is a big step up from cd audio. i wonder how that compares to vinyl.

why i asked in the first place is people make it sound as if digital can't even compare to analog. why is this? if it is recorded digitaly with a huge frequency rang and quality, why in gods name would it sound worse then analog. technicaly it is a approximation of an analog sound, but it is so percise an approximation, wouldn't it be imposible to tell the difference?

ricecake · 31st March 2001, 00:54

Quote:

Originally posted by Somebody
Sincere apologies

Sincerely accepted

Quote:

But what I do not comprehend is where did you get the scratching from reason 3?

You were talking about reasons why vinyl was better than CDs. When you said "3) That's it", I assumed you meant that that was all the reasons. Therefore, I pointed out the fact that it is much easier to scratch with vinyl, thereby making them better than CDs in that respect.

Somebody · 31st March 2001, 01:14

ricecake: ahh

snyder: cds can go from 20Hz to 20,000 Hz

Pretty much, in my opinion, you just can't compare analog and digital, you will always find pluses and minuses, being opposite mediums.

In addition, notice how more digital recording equipment is attempting to work, record, and sound like analog equipment. Notice how analog recording equipment is much more expensive than digital ones. Digital was supposed to replace and sound better than analog, but it didn't. So it's attempting to sound a lot like analog.

Sorry about DVD Audio, I really do not know that much about it at the moment (too lazy to inform myself, I suppose), but I think it will be great in surpassing the quality of cds.

wouldn't it be imposible to tell the difference?

To the average human ear, it would probably be impossible. For the well trained ear, well, that's another story. You can actually notice something missing from cds when you compare these to vinyl, sort of like a richness of sound.

Take for example, let's say you play some notes on a piano, in a professional recording environment, and let's say you had both analog and digital equipment. You play your music and such and record and mix and everything you have to do.

Once you play it back in digital, yes, it sounds nice and beautiful.

Once you play it in analog, it will sound as if that piano was playing in front of you, you will not hear the frequencies that are not audible but you may feel them, the richness and crispness. Keep in mind, this being with the best possible equipment you can have while listening.

Well, enough of my babbling, just found this site that may be of interest.

http://www.solstice.demon.co.uk/debate.htm

InvisableMan · 31st March 2001, 01:56

ROFLMAO

that shznit is hella funny

ReDVsion · 31st March 2001, 06:35

On the subject of LPs, when recording from record to computer, what's the best way of removing a WHOLE LOT of clicks & pops from a record, and what levels should I use to record?

Somebody · 31st March 2001, 12:46

this might be of interest, ReD
http://www.audiotools.co.uk/vinyl.htm

SNYder · 31st March 2001, 22:44

digital is still my rhelm of shiznitle bididdle. The microscopic differences of analog to digital won't concern me untill i have enough money to buy the 100 thousand dollar equipment I need. But I understand in full. Vey intresting stuff. I am going to look into dvd audio a bit further now. Thanks for the reply.

Peace

Somebody · 1st April 2001, 01:36

Somebody · 16th April 2001, 21:30

didn't want this to go into the void for some reason....

rm' · 16th April 2001, 21:51

Neither do I. I remember a long while ago there was a similar discussion about analog vs. digital (at least I'm assuming it's similar; I can't be bothered to read this whole thread). Anyway, we all got to talking about how you could store an analog signal with out any loss on a digital medium, so that you get all the benefits of digital (durability, portability, cost), with the superior sound of analog.

One proposal was that you translate the analog signal into an image of the wave. But you know what? I thought about this, and realized you'd get little or no compression. You'd end up with something even larger than a wave, and may even have to sacrifice signal integrity. Think about it... you need to know what frequency the signal is at a particular point to graph it, right? Well, that means you must sample the signal. Exactly like digital. You can't sample in infinetly small lengths, so you'd end up missing some of the signal...

But someone (God.. how I wish know who said this, because they are a genius) realized that you could attack the problem at the root. All waves can be expressed mathematically, so why don't you just translate the analog signal into it's base formula? Now, this would require a ton of processing power, and infinite amounts of programming know-how, but I do think it's possible (I know little about audio programming, but wouldn't the process be similar to how you display the signal of an mp3 file as an occiliscope readout?).

Imagine... if you could transform a 60 minute concerto into a few lines of numbers and mathematical operators (maybe I'm exaggerating... the end formula could possibly be hundreds of lines long). The resulting formula could be saved as a simple txt file, and transferred along copper lines in a flash. To be honest, I don't think you could actually listen to it real-time; processing demands would be too much. You'd probably have to translate the formula into a wave, then digitize it into a file format, or output to an analog recording machine.

Anyway, that's just a jumbled mess of thoughts... I leave it to you to interprete it.

Somebody · 17th April 2001, 00:00

Quote:

Anyway, that's just a jumbled mess of thoughts... I leave it to you to interprete it.

The way I interprete it, RM, is the following: digital audio.

:-D

SNYder · 17th April 2001, 01:17

RM. Your drunken babble is better then my normal speach

Anyway. Was all that about a way of compressing wavs without any quality lose? Cause I'm not sure what the point of all that was. sorry

rm' · 17th April 2001, 02:22

lol... yes, I'm thoroughly insane, and I guess you just proved me right; I'm the only one who has a clue what I'm talking about

I'm talking like 99% compression here. If you take the waveform of an analog signal, and can convert it to a mathematical formula (from which you can derive a waveform, i.e. graph it as an audio signal), you could turn any analog input into a bunch of numbers (the forumla). You could then store that data digitally, thus, turning a 100% analog signal (which can be derived from the mathematical formula) into digital information.

The advantage? First of all, it's much smaller. It'd only be a few lines of ASCII text at most. Plus, you'd get analog definition and character, but stored on a digital medium (a .txt file).

Any clearer, or am I still talking to the wall?

rm' · 17th April 2001, 03:04

My brain and mouth finally decided to work together, so now I know what I'm saying to try

Here's the human-readable format:

Graph y = sinq, and what do you get (That's supposed to be a theta, btw)? The simplest waveform- a sin wave. It just sounds like a pulsing "whoomph... woomph" (if the frequency is low enough), or a high pitch squeal. But sound nonetheless. "What's this supposed to mean to me?" you ask. It's just a little bit of noise after all... nothing to award a Grammy for. Well, have you ever seen a cosine wave? It's very erratic... much like natural music.

Theoretically, you could recreate any waveform using a mathematical formula. After all, what's a wave but a function of frequency vs. time? All functions (to the best of my knowledge) are derived from a formula. So, if you can find the forumula for any given signal, you're in business.

Here's what my scheme looks like on paper:

You have an analog source (say... a guitar), and you hook it up to your computer. You pipe the analog signal through your conversion software. Actually, it'd most likely have to be hardware, because with analog circuitry, you'd completely avoid worrying about DAC's (using one would actually defeat the whole purpose). So, this software, or machine, would take the inputted waveform, analyze it, and spit out a mathematical forumula. From there, it's as easy as saving it to plain and simple ASCII text, for rapid transfer around the globe. When it reaches the end user, he opens up the .txt with his decoder software. The software solves the formula, and recreates the waveform. From there, anything can happen. It can be converted into a digital signal to be saved as an mp3, or piped directly to speakers, for pure, nondestructible analog audio.

Just imagine how tiny you could make audio... one ASCII character takes 8 bits (one byte). A THOUSAND characters only makes 1kb! If you had a waveform formula that was 100 000 characters long, it'd only 100kb! Very similar to MIDI in data storage concept. You convert the audio to text, then recreate the original audio for playback. The difference between MIDI and this is that with MIDI, you are converting keystrokes; with this, you are converting the actual audio.

But like I said, this'd take immense brain power to see into action. I doubt that all the brains at MIT and Waterloo could accomplish the conversion algorithm.

Essentially, what I'm saying is to take a song. Take it's waveform. Find the mathematical formula of that specific waveform. Save that formula as a lowly text file. Solve the formula, and recreate the waveform. Bang! You hit the money, and played back the song.

The trick is to figure out how to get that formula from a waveform.

rm' · 17th April 2001, 03:06

I think this'd be an amazing application of chaos theory... try to find the order in the chaos of a musical waveform.

Tom · 17th April 2001, 03:18

RM,
I am not an expert in curve fitting, but every time I have seen/or performed this operation it has been on a data series. How you would fit a curve (which would be tied to a equation) to another curve is something I would have to give a little thought to. But as I see it now the only way would involve taking samples in some form or another and deriving a mathematical relationship based on those samples. The problem with this is that you are in essence creating a equation for a digitized wave. Because you are still sampling you will have the same faults that current digital audio has. The only way to avoid it would be to increase your sample rate so you are sure to accurately reproduce the data. As for what sample rate is sufficient to do this is beyond me. I suppose it would depend on the type of music which has an effect on the speed at which the waves decay. You would have to ensure it was small enough to get the "true" peaks of the waves, rather than a point on the rise and fall. This creates truncated peaks which I would assume is where the lack of richness occurs. As for fitting the whole wave to one equation I think it would be more practical to have an equation represent a certain length of the time, I think this would reduce the "encoding" time (but I could be wrong). Overall it is an interesting concept, however I think the equations would be much more complex than what you are anticipating.

Tom

rm' · 17th April 2001, 03:32

Like I said, I don't know what I'm talking about

Just a thought... couldn't you use a, like you were alluding to, a series of formulas? I don't know if it could be done, but can't you break down any analog waveform into sine waves? It'd be like additive theory for light, except here, you are adding together sine waves to get the musical waveform. I dunno if that'd work.

And yes, I fully realize that the formula you'd get would be insanely complex.

Tom · 17th April 2001, 03:52

Quote:

Originally posted by Radioactive Man
Like I said, I don't know what I'm talking about

Just a thought... couldn't you use a, like you were alluding to, a series of formulas? I don't know if it could be done, but can't you break down any analog waveform into sine waves? It'd be like additive theory for light, except here, you are adding together sine waves to get the musical waveform. I dunno if that'd work.

And yes, I fully realize that the formula you'd get would be insanely complex.

Yes I believe you should be able to represent this with a series if sin and cos functions. I was thinking that it might be easier if we were to switch domains (originally thinking of a Laplace transform since that is what I have dealt with for other stuff) but it appears as if a Fourier transform would be the best. Here is a page that get crazy with FFT and digital audio (a little math intensive though) http://www.intersrv.com/~dcross/fft.html

But I think it all goes back to the sample rate, your fit will only be as good as the points you feed it and since you are feeding it points you will not have an exact fit.

Tom

rm' · 17th April 2001, 04:28

Quote:

Originally posted by Tom
But I think it all goes back to the sample rate, your fit will only be as good as the points you feed it and since you are feeding it points you will not have an exact fit.

Tom

Now that I think of it, that's not really an issue of you are dealing with an analog circuit (hardware, as opposed to software). Since analog circuits deal in variable frequencies and current, they recieve a constant stream of data, i.e. infinite sampling points. The only problem I see in this would be the actual decision making; non-IC based decision making circuits are very inefficient and power comsumptive, so it would be hard to do any kind of number crunching with an analog circuit (think ENIAC, and you'll see what I mean).

Tom · 17th April 2001, 04:45

Yes but fitting an equation to an infinite number of points will need an infinite amount of processing power (at least by methods that I am currently aware of, and again am I not an expert in curve fitting so there maybe something out there but all the methods I can think of fit an equation to points).

rm' · 17th April 2001, 05:17

Ahhh! I knew it was too good to be true.

I dunno... I'll keep on thinking and researching about this... there's got to be a way.

31st March 2001, 06:35	#9
ReDVsion Major Dude Join Date: Aug 2000 Posts: 1,426	On the subject of LPs, when recording from record to computer, what's the best way of removing a WHOLE LOT of clicks & pops from a record, and what levels should I use to record? ?

16th April 2001, 21:51	#14
rm' Banned Join Date: Jul 2000 Posts: 11,361	Be forewarned: I am entering "Uninformed drunken babble mode", and have little idea Neither do I. I remember a long while ago there was a similar discussion about analog vs. digital (at least I'm assuming it's similar; I can't be bothered to read this whole thread). Anyway, we all got to talking about how you could store an analog signal with out any loss on a digital medium, so that you get all the benefits of digital (durability, portability, cost), with the superior sound of analog. One proposal was that you translate the analog signal into an image of the wave. But you know what? I thought about this, and realized you'd get little or no compression. You'd end up with something even larger than a wave, and may even have to sacrifice signal integrity. Think about it... you need to know what frequency the signal is at a particular point to graph it, right? Well, that means you must sample the signal. Exactly like digital. You can't sample in infinetly small lengths, so you'd end up missing some of the signal... But someone (God.. how I wish know who said this, because they are a genius) realized that you could attack the problem at the root. All waves can be expressed mathematically, so why don't you just translate the analog signal into it's base formula? Now, this would require a ton of processing power, and infinite amounts of programming know-how, but I do think it's possible (I know little about audio programming, but wouldn't the process be similar to how you display the signal of an mp3 file as an occiliscope readout?). Imagine... if you could transform a 60 minute concerto into a few lines of numbers and mathematical operators (maybe I'm exaggerating... the end formula could possibly be hundreds of lines long). The resulting formula could be saved as a simple txt file, and transferred along copper lines in a flash. To be honest, I don't think you could actually listen to it real-time; processing demands would be too much. You'd probably have to translate the formula into a wave, then digitize it into a file format, or output to an analog recording machine. Anyway, that's just a jumbled mess of thoughts... I leave it to you to interprete it.

17th April 2001, 03:04	#18
rm' Banned Join Date: Jul 2000 Posts: 11,361	Aha! My brain and mouth finally decided to work together, so now I know what I'm saying to try Here's the human-readable format: Graph y = sinq, and what do you get (That's supposed to be a theta, btw)? The simplest waveform- a sin wave. It just sounds like a pulsing "whoomph... woomph" (if the frequency is low enough), or a high pitch squeal. But sound nonetheless. "What's this supposed to mean to me?" you ask. It's just a little bit of noise after all... nothing to award a Grammy for. Well, have you ever seen a cosine wave? It's very erratic... much like natural music. Theoretically, you could recreate any waveform using a mathematical formula. After all, what's a wave but a function of frequency vs. time? All functions (to the best of my knowledge) are derived from a formula. So, if you can find the forumula for any given signal, you're in business. Here's what my scheme looks like on paper: You have an analog source (say... a guitar), and you hook it up to your computer. You pipe the analog signal through your conversion software. Actually, it'd most likely have to be hardware, because with analog circuitry, you'd completely avoid worrying about DAC's (using one would actually defeat the whole purpose). So, this software, or machine, would take the inputted waveform, analyze it, and spit out a mathematical forumula. From there, it's as easy as saving it to plain and simple ASCII text, for rapid transfer around the globe. When it reaches the end user, he opens up the .txt with his decoder software. The software solves the formula, and recreates the waveform. From there, anything can happen. It can be converted into a digital signal to be saved as an mp3, or piped directly to speakers, for pure, nondestructible analog audio. Just imagine how tiny you could make audio... one ASCII character takes 8 bits (one byte). A THOUSAND characters only makes 1kb! If you had a waveform formula that was 100 000 characters long, it'd only 100kb! Very similar to MIDI in data storage concept. You convert the audio to text, then recreate the original audio for playback. The difference between MIDI and this is that with MIDI, you are converting keystrokes; with this, you are converting the actual audio. But like I said, this'd take immense brain power to see into action. I doubt that all the brains at MIT and Waterloo could accomplish the conversion algorithm. Essentially, what I'm saying is to take a song. Take it's waveform. Find the mathematical formula of that specific waveform. Save that formula as a lowly text file. Solve the formula, and recreate the waveform. Bang! You hit the money, and played back the song. The trick is to figure out how to get that formula from a waveform.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Switch to Linear Mode Hybrid Mode Switch to Threaded Mode

30th March 2001, 19:42	#4
Somebody Major Dude Join Date: Oct 2000 Posts: 1,531	Sincere apologies, I figured about the dj scratching thing. But what I do not comprehend is where did you get the scratching from reason 3?

31st March 2001, 00:30	#5
SNYder Forum King Join Date: Nov 2000 Posts: 4,432	I actually knew about higher frequencys (out of human hearing ranges) affecting the feel of the music, but i had no idea vinyl could go higher then cd's. hmmm... what is the frequency range for cd's? it must be more then 20,000, correct? i can't imagine them being limited to 20 to 20,000. And don't forget dvd audio wich is a big step up from cd audio. i wonder how that compares to vinyl. why i asked in the first place is people make it sound as if digital can't even compare to analog. why is this? if it is recorded digitaly with a huge frequency rang and quality, why in gods name would it sound worse then analog. technicaly it is a approximation of an analog sound, but it is so percise an approximation, wouldn't it be imposible to tell the difference?

31st March 2001, 01:14	#7
Somebody Major Dude Join Date: Oct 2000 Posts: 1,531	ricecake: ahh snyder: cds can go from 20Hz to 20,000 Hz Pretty much, in my opinion, you just can't compare analog and digital, you will always find pluses and minuses, being opposite mediums. In addition, notice how more digital recording equipment is attempting to work, record, and sound like analog equipment. Notice how analog recording equipment is much more expensive than digital ones. Digital was supposed to replace and sound better than analog, but it didn't. So it's attempting to sound a lot like analog. Sorry about DVD Audio, I really do not know that much about it at the moment (too lazy to inform myself, I suppose), but I think it will be great in surpassing the quality of cds. wouldn't it be imposible to tell the difference? To the average human ear, it would probably be impossible. For the well trained ear, well, that's another story. You can actually notice something missing from cds when you compare these to vinyl, sort of like a richness of sound. Take for example, let's say you play some notes on a piano, in a professional recording environment, and let's say you had both analog and digital equipment. You play your music and such and record and mix and everything you have to do. Once you play it back in digital, yes, it sounds nice and beautiful. Once you play it in analog, it will sound as if that piano was playing in front of you, you will not hear the frequencies that are not audible but you may feel them, the richness and crispness. Keep in mind, this being with the best possible equipment you can have while listening. Well, enough of my babbling, just found this site that may be of interest. http://www.solstice.demon.co.uk/debate.htm

31st March 2001, 01:56	#8
InvisableMan Ninja Master! (Forum King) Join Date: Mar 2001 Location: Hotel California Posts: 4,333	ROFLMAO that shznit is hella funny

31st March 2001, 12:46	#10
Somebody Major Dude Join Date: Oct 2000 Posts: 1,531	this might be of interest, ReD http://www.audiotools.co.uk/vinyl.htm

31st March 2001, 22:44	#11
SNYder Forum King Join Date: Nov 2000 Posts: 4,432	digital is still my rhelm of shiznitle bididdle. The microscopic differences of analog to digital won't concern me untill i have enough money to buy the 100 thousand dollar equipment I need. But I understand in full. Vey intresting stuff. I am going to look into dvd audio a bit further now. Thanks for the reply. Peace

1st April 2001, 01:36	#12
Somebody Major Dude Join Date: Oct 2000 Posts: 1,531

16th April 2001, 21:30	#13
Somebody Major Dude Join Date: Oct 2000 Posts: 1,531	didn't want this to go into the void for some reason....

17th April 2001, 01:17	#16
SNYder Forum King Join Date: Nov 2000 Posts: 4,432	RM. Your drunken babble is better then my normal speach Anyway. Was all that about a way of compressing wavs without any quality lose? Cause I'm not sure what the point of all that was. sorry

17th April 2001, 02:22	#17
rm' Banned Join Date: Jul 2000 Posts: 11,361	lol... yes, I'm thoroughly insane, and I guess you just proved me right; I'm the only one who has a clue what I'm talking about I'm talking like 99% compression here. If you take the waveform of an analog signal, and can convert it to a mathematical formula (from which you can derive a waveform, i.e. graph it as an audio signal), you could turn any analog input into a bunch of numbers (the forumla). You could then store that data digitally, thus, turning a 100% analog signal (which can be derived from the mathematical formula) into digital information. The advantage? First of all, it's much smaller. It'd only be a few lines of ASCII text at most. Plus, you'd get analog definition and character, but stored on a digital medium (a .txt file). Any clearer, or am I still talking to the wall?

17th April 2001, 03:06	#19
rm' Banned Join Date: Jul 2000 Posts: 11,361	I think this'd be an amazing application of chaos theory... try to find the order in the chaos of a musical waveform.

17th April 2001, 03:18	#20
Tom Moderator Join Date: Apr 2000 Posts: 4,491	RM, I am not an expert in curve fitting, but every time I have seen/or performed this operation it has been on a data series. How you would fit a curve (which would be tied to a equation) to another curve is something I would have to give a little thought to. But as I see it now the only way would involve taking samples in some form or another and deriving a mathematical relationship based on those samples. The problem with this is that you are in essence creating a equation for a digitized wave. Because you are still sampling you will have the same faults that current digital audio has. The only way to avoid it would be to increase your sample rate so you are sure to accurately reproduce the data. As for what sample rate is sufficient to do this is beyond me. I suppose it would depend on the type of music which has an effect on the speed at which the waves decay. You would have to ensure it was small enough to get the "true" peaks of the waves, rather than a point on the rise and fall. This creates truncated peaks which I would assume is where the lack of richness occurs. As for fitting the whole wave to one equation I think it would be more practical to have an equation represent a certain length of the time, I think this would reduce the "encoding" time (but I could be wrong). Overall it is an interesting concept, however I think the equations would be much more complex than what you are anticipating. Tom

17th April 2001, 03:32	#21
rm' Banned Join Date: Jul 2000 Posts: 11,361	Like I said, I don't know what I'm talking about Just a thought... couldn't you use a, like you were alluding to, a series of formulas? I don't know if it could be done, but can't you break down any analog waveform into sine waves? It'd be like additive theory for light, except here, you are adding together sine waves to get the musical waveform. I dunno if that'd work. And yes, I fully realize that the formula you'd get would be insanely complex.

17th April 2001, 04:45	#24
Tom Moderator Join Date: Apr 2000 Posts: 4,491	Yes but fitting an equation to an infinite number of points will need an infinite amount of processing power (at least by methods that I am currently aware of, and again am I not an expert in curve fitting so there maybe something out there but all the methods I can think of fit an equation to points).

17th April 2001, 05:17	#25
rm' Banned Join Date: Jul 2000 Posts: 11,361	Ahhh! I knew it was too good to be true. I dunno... I'll keep on thinking and researching about this... there's got to be a way.