Hello All!
Old: I was wondering if anyone here has already heard of Band pass filters here?
I think I have found one, and I was wondering if what I am "discovering" is really a discovery, or not?
If it is I will bring out my system, when it is fully operational. If not, I will just STFU. :x
EDIT 1: Now now now. I have found it. I have CREATED!!!!!
This is not yet the Band Pass Filter, but I have figured out the 'reacting' system, meaning the basis of all filters.
This is the system:
Let us call P the period of the signal, and n the delay (the delayer is just a repeater). I have not demonstrated anything, but I have found quite exactly, that if n=2P-1, then the output is stable and off. If not it keeps flickering.
How I found the formula?
If you want I will go into details, but when I noticed the effect noted above, I noted down the delay value and the period. Then in that particular case, I started just writing down, frame by frame, the sequence of values of the different branches. Then I noticed the different cases, when n diverged from the value for which I observed the phenomenon, and when I looked at the values, I saw some sort of pattern, and when the strechted or shrunk, I just kinda saw the relation I had to find, I had to find the right patern in between the value after, and before the delay, because that was what really was important. This is a vague description, and will devellop, if it is requested. No problem.
This is no proof, I don't know if the same thing happens for other values. I did not test yet for irregular patterns. I don't think it will work for irregular patterns, actually. But for a regular pattern, it works, for sure, whatever the value of P.
EDIT 2:
Quote from radicalapple »
I think this is exploiting the fact that redstone torches and circuits have a delay/lag when being used, right? So if it flicks too fast, the system can not react fast enough to pick up the flickering(or similar principles), right?
Yes, well actually, I still don't fully understand the system, but it's based on dephased signals really. You delay your signal, when the delay is right, the output is always off.
Anyway, I finished the Band Pass.
However, there is a residual noise, when the frequency is not right. It remains off, and suddenly, there can be one or two blips of 1 values. I'll show you right away the schema, and the real system, to have an idea of the scale. It's relatively small:
(above, i'ts a 7 clock, not 4, sorry)
Schema (I may have forgotten an inverser somewhere)
Uh.... how does it work. Good question. Two cases:
-The system is receiving the right frequency:
Then the nullifying system sends out a constant 0 signal
Then the RS NOR Latch receives a constant 0 signal, from one side.
From the Other Side, he receives a signal which blips from 0 to 1 from time to time.
So he stays in the position where that output delivers a 1(We will call it State A (opposed to STATE :cool.gif:). I then use the branch of the RS NOR latch which is ON in that state, invert it, so I have a constant NO arriving to an NOR command. The other input of the NOR is the the output of the nullifying system, which is also giving a constant NO. Therefore, the NOR gives out a yes. I invert that yes into a constant NO and branch the INPUT clock of the system and the NO signal I am receiving from the NOR to a NOR. It is on only when both are off. One is always off, so it delivers 1 when the input signal delivers 0. THAT is the output. And it delivers my signal. Good!
More complicated:
-The system does not have the right frequency.
The Nullefying System delivers a flickering signal.
The RS NOR Latch receives a relatively rapid flickering signal on one side, a long period one from the other.
When the Nullifying system flickers, it puts the system in state B. so I have 0 coming out of the branch I use.
I must then Have 1 coming into my first NOR Gate, so 0 in output, being inversed, it delivers 1 to the second latch which then delivers a 0 in output. No signal!
Now why does the RS NOR Latch have a second output that flickers? It's to allow an update of the system, really, to check from time to time that the signal is still not right. It's the only use of the 7 and 16 clocks, so it might not be useful to all. If it is not necessary to you to have an update system, no problemo, just chuck it out, it is not vital to the system. Actually, don't put it unless it is necessary, because it's that system that creates noise at the end, as I had told at the beginning of this text.
Uhhh, I hope this might help. My Idea now is to make a sort of oscillo
Yeah, consider my curiosity piqued. I have no idea how to reconcile what I know of band pass filters and the minecraft universe, redstone in particular. What would you be filtering?
Cool, for once I find something NOT know, apparently.
The input would be a periodic signal (type 5-clock, 7-clock, see what I mean) And only deliver this signal, if it is, well at specific frequency.
Maybe an example would be more clear?
A 5-band Pass, would be able to let a signal pass, only if the entry would have the right period, meaning the period of a 5-clock signal.
I already have a system that reacts to the period, shuting things on, or off, in fuction of the entry period, but it is purely experimental, for now. I still don't control. I'll be posting the solution tomorow, I gather.
Now now now. I have found it. I have CREATED!!!!!
This is not yet the Band Pass Filter, but I have figured out the 'reacting' system, meaning the basis of all filters.
This is the system:
Let us call P the period of the signal, and n the delay (the delayer is just a repeater). I have not demonstrated anything, but I have found quite exactly, that if n=2P-1, then the output is stable and off. If not it keeps flickering.
How I found the formula?
If you want I will go into details, but when I noticed the effect noted above, I noted down the delay value and the period. Then in that particular case, I started just writing down, frame by frame, the sequence of values of the different branches. Then I noticed the different cases, when n diverged from the value for which I observed the phenomenon, and when I looked at the values, I saw some sort of pattern, and when the strechted or shrunk, I just kinda saw the relation I had to find, I had to find the right patern in between the value after, and before the delay, because that was what really was important. This is a vague description, and will devellop, if it is requested. No problem.
This is no proof, I don't know if the same thing happens for other values. I did not test yet for irregular patterns. I don't think it will work for irregular patterns, actually. But for a regular pattern, it works, for sure, whatever the value of P.
Just have to ask something: does this merely let a signal pass if it matches it or if the entire patern matches it? I just ask because it seems like this would only work on a case by case basis. For example, using a fiveclock as your setting, and a 3 clock as your input, wouldn't the output be as if it was a 15 clock, as the inputs matched, desynced, and repeated?
I do not know if it would work as a 15 clock, actually I don't think it would be, but it definitly would be a clock. I am now working on getting rid of that clock effect when it is not the right input.
I think this is exploiting the fact that redstone torches and circuits have a delay/lag when being used, right? So if it flicks too fast, the system can not react fast enough to pick up the flickering(or similar principles), right?
If you want to make something more amazing, why don't you make an op-amp and capacitor? You can do a lot more with these than a band past filter.
I think this is exploiting the fact that redstone torches and circuits have a delay/lag when being used, right? So if it flicks too fast, the system can not react fast enough to pick up the flickering(or similar principles), right?
Yes, well actually, I still don't fully understand the system, but it's based on dephased signals really. You delay your signal, when the delay is right, the output is always off.
Anyway, I finished the Band Pass.
However, there is a residual noise, when the frequency is not right. It remains off, and suddenly, there can be one or two blips of 1 values. I'll show you right away the schema, and the real system, to have an idea of the scale. It's relatively small:
(above, i'ts a 7 clock, not 4, sorry)
Schema (I may have forgotten an inverser somewhere)
Uh.... how does it work. Good question. Two cases:
-The system is receiving the right frequency:
Then the nullifying system sends out a constant 0 signal
Then the RS NOR Latch receives a constant 0 signal, from one side.
From the Other Side, he receives a signal which blips from 0 to 1 from time to time.
So he stays in the position where that output delivers a 1(We will call it State A (opposed to STATE :cool.gif:). I then use the branch of the RS NOR latch which is ON in that state, invert it, so I have a constant NO arriving to an NOR command. The other input of the NOR is the the output of the nullifying system, which is also giving a constant NO. Therefore, the NOR gives out a yes. I invert that yes into a constant NO and branch the INPUT clock of the system and the NO signal I am receiving from the NOR to a NOR. It is on only when both are off. One is always off, so it delivers 1 when the input signal delivers 0. THAT is the output. And it delivers my signal. Good!
More complicated:
-The system does not have the right frequency.
The Nullefying System delivers a flickering signal.
The RS NOR Latch receives a relatively rapid flickering signal on one side, a long period one from the other.
When the Nullifying system flickers, it puts the system in state B. so I have 0 coming out of the branch I use.
I must then Have 1 coming into my first NOR Gate, so 0 in output, being inversed, it delivers 1 to the second latch which then delivers a 0 in output. No signal!
Now why does the RS NOR Latch have a second output that flickers? It's to allow an update of the system, really, to check from time to time that the signal is still not right. It's the only use of the 7 and 16 clocks, so it might not be useful to all. If it is not necessary to you to have an update system, no problemo, just chuck it out, it is not vital to the system. Actually, don't put it unless it is necessary, because it's that system that creates noise at the end, as I had told at the beginning of this text.
Uhhh, I hope this might help. My Idea now is to make a sort of oscilloscope. A sort of frequency in function of time, screen.
I have experimented with this kind of thing as a way to send > 1 bit of information over a redstone line.
The circuit can actually be simplified. You are effectively delaying by one period and inverting, thus only the correct frequency completely cancels itself. This also works if the delay is 1/2 the period and not inverted, effectively shifting the phase 180 degrees. This circuit uses half the number of repeaters and only one torch, like this:
in --+----+
| |
#>>>+#*-- out
- + dust # block > repeater * torch
By adding a convolution filter after the band-pass filter, the output can be smoothed so that it's always a steady on or off, never flickering. A convolution filter, in this case, is just a chain of repeaters with blocks between them, and a line of redstone dust running along the side so that it picks up the signal from every block. The total length of this delay line needs to be at least as long as the longest period of any signal that may be fed into the circuit.
in --+----+
| |
#>>>+#*>#>#>#>#
+---------- out
Practically, there are only a few different messages that can be sent like this i.e. five-clock, seven, nine, eleven. Beyond that, the circuit becomes too big and slow for my tastes.
I've also made a circuit that can decode an arbitrary bit sequence but it's pretty complicated. I'm currently looking at PWM as yet another possibility.
I have experimented with this kind of thing as a way to send > 1 bit of information over a redstone line.
The circuit can actually be simplified. You are effectively delaying by one period and inverting, thus only the correct frequency completely cancels itself. This also works if the delay is 1/2 the period and not inverted, effectively shifting the phase 180 degrees. This circuit uses half the number of repeaters and only one torch, like this:
in --+----+
| |
#>>>+#*-- out
- + dust # block > repeater * torch
By adding a convolution filter after the band-pass filter, the output can be smoothed so that it's always a steady on or off, never flickering. A convolution filter, in this case, is just a chain of repeaters with blocks between them, and a line of redstone dust running along the side so that it picks up the signal from every block. The total length of this delay line needs to be at least as long as the longest period of any signal that may be fed into the circuit.
in --+----+
| |
#>>>+#*>#>#>#>#
+---------- out
Practically, there are only a few different messages that can be sent like this i.e. five-clock, seven, nine, eleven. Beyond that, the circuit becomes too big and slow for my tastes.
I've also made a circuit that can decode an arbitrary bit sequence but it's pretty complicated. I'm currently looking at PWM as yet another possibility.
I'm told this is a PWM output, let me know if it's of use to you
Old:
I was wondering if anyone here has already heard of Band pass filters here?
I think I have found one, and I was wondering if what I am "discovering" is really a discovery, or not?
If it is I will bring out my system, when it is fully operational. If not, I will just STFU. :x
EDIT 1:
Now now now. I have found it. I have CREATED!!!!!
This is not yet the Band Pass Filter, but I have figured out the 'reacting' system, meaning the basis of all filters.
This is the system:
Let us call P the period of the signal, and n the delay (the delayer is just a repeater). I have not demonstrated anything, but I have found quite exactly, that if n=2P-1, then the output is stable and off. If not it keeps flickering.
How I found the formula?
If you want I will go into details, but when I noticed the effect noted above, I noted down the delay value and the period. Then in that particular case, I started just writing down, frame by frame, the sequence of values of the different branches. Then I noticed the different cases, when n diverged from the value for which I observed the phenomenon, and when I looked at the values, I saw some sort of pattern, and when the strechted or shrunk, I just kinda saw the relation I had to find, I had to find the right patern in between the value after, and before the delay, because that was what really was important. This is a vague description, and will devellop, if it is requested. No problem.
This is no proof, I don't know if the same thing happens for other values. I did not test yet for irregular patterns. I don't think it will work for irregular patterns, actually. But for a regular pattern, it works, for sure, whatever the value of P.
EDIT 2:
Yes, well actually, I still don't fully understand the system, but it's based on dephased signals really. You delay your signal, when the delay is right, the output is always off.
Anyway, I finished the Band Pass.
However, there is a residual noise, when the frequency is not right. It remains off, and suddenly, there can be one or two blips of 1 values. I'll show you right away the schema, and the real system, to have an idea of the scale. It's relatively small:
(above, i'ts a 7 clock, not 4, sorry)
Schema (I may have forgotten an inverser somewhere)
Uh.... how does it work. Good question. Two cases:
-The system is receiving the right frequency:
Then the nullifying system sends out a constant 0 signal
Then the RS NOR Latch receives a constant 0 signal, from one side.
From the Other Side, he receives a signal which blips from 0 to 1 from time to time.
So he stays in the position where that output delivers a 1(We will call it State A (opposed to STATE :cool.gif:). I then use the branch of the RS NOR latch which is ON in that state, invert it, so I have a constant NO arriving to an NOR command. The other input of the NOR is the the output of the nullifying system, which is also giving a constant NO. Therefore, the NOR gives out a yes. I invert that yes into a constant NO and branch the INPUT clock of the system and the NO signal I am receiving from the NOR to a NOR. It is on only when both are off. One is always off, so it delivers 1 when the input signal delivers 0. THAT is the output. And it delivers my signal. Good!
More complicated:
-The system does not have the right frequency.
The Nullefying System delivers a flickering signal.
The RS NOR Latch receives a relatively rapid flickering signal on one side, a long period one from the other.
When the Nullifying system flickers, it puts the system in state B. so I have 0 coming out of the branch I use.
I must then Have 1 coming into my first NOR Gate, so 0 in output, being inversed, it delivers 1 to the second latch which then delivers a 0 in output. No signal!
Now why does the RS NOR Latch have a second output that flickers? It's to allow an update of the system, really, to check from time to time that the signal is still not right. It's the only use of the 7 and 16 clocks, so it might not be useful to all. If it is not necessary to you to have an update system, no problemo, just chuck it out, it is not vital to the system. Actually, don't put it unless it is necessary, because it's that system that creates noise at the end, as I had told at the beginning of this text.
Uhhh, I hope this might help. My Idea now is to make a sort of oscillo
How would you even get a band pass filter in Minecraft?
Please share!
The input would be a periodic signal (type 5-clock, 7-clock, see what I mean) And only deliver this signal, if it is, well at specific frequency.
Maybe an example would be more clear?
A 5-band Pass, would be able to let a signal pass, only if the entry would have the right period, meaning the period of a 5-clock signal.
I already have a system that reacts to the period, shuting things on, or off, in fuction of the entry period, but it is purely experimental, for now. I still don't control. I'll be posting the solution tomorow, I gather.
Uh... what is a heavy dubstep bassline?
This is not yet the Band Pass Filter, but I have figured out the 'reacting' system, meaning the basis of all filters.
This is the system:
Let us call P the period of the signal, and n the delay (the delayer is just a repeater). I have not demonstrated anything, but I have found quite exactly, that if n=2P-1, then the output is stable and off. If not it keeps flickering.
How I found the formula?
If you want I will go into details, but when I noticed the effect noted above, I noted down the delay value and the period. Then in that particular case, I started just writing down, frame by frame, the sequence of values of the different branches. Then I noticed the different cases, when n diverged from the value for which I observed the phenomenon, and when I looked at the values, I saw some sort of pattern, and when the strechted or shrunk, I just kinda saw the relation I had to find, I had to find the right patern in between the value after, and before the delay, because that was what really was important. This is a vague description, and will devellop, if it is requested. No problem.
This is no proof, I don't know if the same thing happens for other values. I did not test yet for irregular patterns. I don't think it will work for irregular patterns, actually. But for a regular pattern, it works, for sure, whatever the value of P.
If you want to make something more amazing, why don't you make an op-amp and capacitor? You can do a lot more with these than a band past filter.
Yes, well actually, I still don't fully understand the system, but it's based on dephased signals really. You delay your signal, when the delay is right, the output is always off.
Anyway, I finished the Band Pass.
However, there is a residual noise, when the frequency is not right. It remains off, and suddenly, there can be one or two blips of 1 values. I'll show you right away the schema, and the real system, to have an idea of the scale. It's relatively small:
(above, i'ts a 7 clock, not 4, sorry)
Schema (I may have forgotten an inverser somewhere)
Uh.... how does it work. Good question. Two cases:
-The system is receiving the right frequency:
Then the nullifying system sends out a constant 0 signal
Then the RS NOR Latch receives a constant 0 signal, from one side.
From the Other Side, he receives a signal which blips from 0 to 1 from time to time.
So he stays in the position where that output delivers a 1(We will call it State A (opposed to STATE :cool.gif:). I then use the branch of the RS NOR latch which is ON in that state, invert it, so I have a constant NO arriving to an NOR command. The other input of the NOR is the the output of the nullifying system, which is also giving a constant NO. Therefore, the NOR gives out a yes. I invert that yes into a constant NO and branch the INPUT clock of the system and the NO signal I am receiving from the NOR to a NOR. It is on only when both are off. One is always off, so it delivers 1 when the input signal delivers 0. THAT is the output. And it delivers my signal. Good!
More complicated:
-The system does not have the right frequency.
The Nullefying System delivers a flickering signal.
The RS NOR Latch receives a relatively rapid flickering signal on one side, a long period one from the other.
When the Nullifying system flickers, it puts the system in state B. so I have 0 coming out of the branch I use.
I must then Have 1 coming into my first NOR Gate, so 0 in output, being inversed, it delivers 1 to the second latch which then delivers a 0 in output. No signal!
Now why does the RS NOR Latch have a second output that flickers? It's to allow an update of the system, really, to check from time to time that the signal is still not right. It's the only use of the 7 and 16 clocks, so it might not be useful to all. If it is not necessary to you to have an update system, no problemo, just chuck it out, it is not vital to the system. Actually, don't put it unless it is necessary, because it's that system that creates noise at the end, as I had told at the beginning of this text.
Uhhh, I hope this might help. My Idea now is to make a sort of oscilloscope. A sort of frequency in function of time, screen.
The circuit can actually be simplified. You are effectively delaying by one period and inverting, thus only the correct frequency completely cancels itself. This also works if the delay is 1/2 the period and not inverted, effectively shifting the phase 180 degrees. This circuit uses half the number of repeaters and only one torch, like this:
By adding a convolution filter after the band-pass filter, the output can be smoothed so that it's always a steady on or off, never flickering. A convolution filter, in this case, is just a chain of repeaters with blocks between them, and a line of redstone dust running along the side so that it picks up the signal from every block. The total length of this delay line needs to be at least as long as the longest period of any signal that may be fed into the circuit.
Practically, there are only a few different messages that can be sent like this i.e. five-clock, seven, nine, eleven. Beyond that, the circuit becomes too big and slow for my tastes.
I've also made a circuit that can decode an arbitrary bit sequence but it's pretty complicated. I'm currently looking at PWM as yet another possibility.
I'm told this is a PWM output, let me know if it's of use to you
My link