Peter Daniel said:
Why? I am not sure I understand exactly what are trying to
say here. I am not protesting against your claims that it makes
a difference to add a LED. I am only claiming that it cannot
affect only the LSBs of the data stream, so there must be
some other kind of optical phenomen at play here.
But you're right that I haven't tried the LED trick with CD players.
Probably will some day, but I am bit lazy and quite a bit of
a theoritician, so I'll probably try to convince my more tweakprone
friend to try it out first on his brand new SACD player before I
give it a try myself.
So what time is in Sweden now?
About four o'clock in the morning
How didi this happen??
Christer said:
Why? I am not sure I understand exactly what are trying to
say here. I am not protesting against your claims that it makes
a difference to add a LED. I am only claiming that it cannot
affect only the LSBs of the data stream, so there must be
some other kind of optical phenomen at play here.
I was trying to say that it's impossible to synchronise the LED this way. Everything occurs too fast.
LSB is "weaker" than other bits, that's why it benefits from added energy (by noise). That's what LED does.
The other bits are not affected neither by noise nor by daylight.
Sorry Peter, I still don't buy it. All the examples they bring up
for comparison are analog signals/systems, which is quite a
different thing. Doing it on the digital side requires that you
can control which bits are affected. BTW, even if possible you
would only benefit from this if using a 16-bit non-oversampling
DAC I guess. In all other cases you would get much better result
adding noise in later stages. I feel I am stepping out into
deeper water in the last sentences, so others may have a
different opinion here.
BTW, the clock was even 4:40 in the morning.
See, I am too tired to even know what time it is.
I'd better try to sleep on this issue, and I am sure there will
be others around tomorrow to provide more detailed and
refined opnions on this.
for comparison are analog signals/systems, which is quite a
different thing. Doing it on the digital side requires that you
can control which bits are affected. BTW, even if possible you
would only benefit from this if using a 16-bit non-oversampling
DAC I guess. In all other cases you would get much better result
adding noise in later stages. I feel I am stepping out into
deeper water in the last sentences, so others may have a
different opinion here.
BTW, the clock was even 4:40 in the morning.
See, I am too tired to even know what time it is.
I'd better try to sleep on this issue, and I am sure there will
be others around tomorrow to provide more detailed and
refined opnions on this.
Peter Daniel said:
OK, yes I also think it would be impossible or at least very
difficult to synchronize the LED. I was just trying to feed
you a last chance of hope for the theory.
I am sorry Peter, but I don't think you quite understand what
digital data is. All bits, including error correction bits etc. have
exactly the same "weight" in the digital data stream. It is only
in the conversion into analog that the bits have different
weights. On the CD, each bit has exaclty the same chance as
any other bit to be affected by the noise. A bit is either 1 or 0,
there are no intermediate values. A 1 in the MSB is exactly the
same thing as a 1 in the LSB, so they are affected in exactly the
same way. I can see no way you could affect the LSBs more
than any other bits, unless you manage to synchronise the
LED. However, if you can come up with convincing arguments
to the contrary I will reconsider my standpoint.
Peter Daniel said:
The LSB is a bit the same as every other bit. And no bit requires any more or less "energy" than any other bit. The LSB is only "weaker" in a purely mathematically symbolic sense. It simply denotes the bit whose change represents the smallest change possible. That's it. It's just a symbolic weighting factor.
Remember the "decimal place" system when you were in school? You had the 1s decimal place, the 10s decimal place, the 100s decimal place, etc? Well the 1s decimal place is akin to the LSB. You can call it the Least Significant Digit rather than the Least Significant Bit. It's purely symbolic. No digit is any different from any other digit except symbolically.
Also, the data doesn't actually exist on the disc in binary, linear PCM form to begin with. Contrary to popular belief, a pit doesn't designate a 1 and a land a 0 (or vice versa). In other words, if the data is 10110 you don't have a pit/land/pit/pit/land.
The data is encoded on the disc in what's called Eight-to-Fourteen Modulation (EFM) and exists as a series of pits/lands of 9 varying lengths, designated 3T, 4T, 5T, 6T, 7T, 8T, 9T 10T and 11T. The shortest pit/land length is between 0.833 and 0.972 um, and the longest is between 3.054 - 3.56 um.
The transition from a pit to a land or a land to a pit is determined by a logical 1 in the EFM data. The EFM data is encoded such that no less than two or no more than ten zeros follow any 1/0 or 0/1 transition.
Since 1/0 or 0/1 transitions are purely a function of the EFM data, and have nothing to do with LSBs, there's simply nothing on the actual disc that has any direct correlation to the actual LSB of the PCM data which is what is actually sent to the DAC.
se
My impressions of the LED claim
Pete,
I hope to try the experiment as soon as I can locate some blue LEDS. I have to say though, that from a scientific viewpoint, there is nothing I've come across to convince me that there should be anything audible, better yet, I'm not yet convinced that the LED should alter the decoded signal in any way.
Right now, I'm reviewing Chapter 8 - The Compact Disc, in Ken Pohlmann's "Principles of Digital Audio". This is highly recommended reading for anyone who wishes to learn about the steps in encoding, decoding, servo control, error correction, etc. From what I've read so far, it seems to me that if the blue LED affected the reflected laser beam's signal (via modulation) that it would happen more likely on 1's (pit edge) where there is ~25% reflection on long pits than on 0's (~95%). The LED interference would have to confuse the photodiode that the wrong value (1 or 0 was being read). The laser's output power is ~0.5mW. I'm not sure that the LEDs optical power is in this ballpark (I need to check). It may be too small in magnitude to cause a signal transition at the photodiode via constructive/destructive interference with the reflected signal from the CD. The way you show your LED in your diagram it looks like it may be 3-4x the distance from the photodiode as the reflected laser beam is to the photodiode?
Furthermore, if the data was altered, then there would be no way to alter only the LSB. All data would have to be altered (assuming it even could be). This is what makes the theory hard to believe: CDs which utilize EFM and CIRC are extremely robust at error correction, missing 1 error out of 10 - 100 billion. According to Pohlmann, when an error is present it is flagged (if detected of course) and then linearly interpolated (most of the time - depends on the CDP). I know of some which actually mute the output. Do you see what I'm getting at? Perhaps someone will argue that linearly interpolated (which is incorrect data) data sounds better on CDs. In this case, source data from the CD is missing during decode, not restored as the LED theorists believe. So, I do not know how optical dithering can be claimed.
Anyway, I've got to more to read. No one can tell you what you hear. If it sounds better to you with the LED than that is great, and in the end that's all that matters. The cause and effect of the claim are difficult for me to understand at this point. Just as important as being able to witness improved sound is for me to try to understand this phenomenon from an scientific standpoint.
-AudioEnthusiast
Pete,
I hope to try the experiment as soon as I can locate some blue LEDS. I have to say though, that from a scientific viewpoint, there is nothing I've come across to convince me that there should be anything audible, better yet, I'm not yet convinced that the LED should alter the decoded signal in any way.
Right now, I'm reviewing Chapter 8 - The Compact Disc, in Ken Pohlmann's "Principles of Digital Audio". This is highly recommended reading for anyone who wishes to learn about the steps in encoding, decoding, servo control, error correction, etc. From what I've read so far, it seems to me that if the blue LED affected the reflected laser beam's signal (via modulation) that it would happen more likely on 1's (pit edge) where there is ~25% reflection on long pits than on 0's (~95%). The LED interference would have to confuse the photodiode that the wrong value (1 or 0 was being read). The laser's output power is ~0.5mW. I'm not sure that the LEDs optical power is in this ballpark (I need to check). It may be too small in magnitude to cause a signal transition at the photodiode via constructive/destructive interference with the reflected signal from the CD. The way you show your LED in your diagram it looks like it may be 3-4x the distance from the photodiode as the reflected laser beam is to the photodiode?
Furthermore, if the data was altered, then there would be no way to alter only the LSB. All data would have to be altered (assuming it even could be). This is what makes the theory hard to believe: CDs which utilize EFM and CIRC are extremely robust at error correction, missing 1 error out of 10 - 100 billion. According to Pohlmann, when an error is present it is flagged (if detected of course) and then linearly interpolated (most of the time - depends on the CDP). I know of some which actually mute the output. Do you see what I'm getting at? Perhaps someone will argue that linearly interpolated (which is incorrect data) data sounds better on CDs. In this case, source data from the CD is missing during decode, not restored as the LED theorists believe. So, I do not know how optical dithering can be claimed.
Anyway, I've got to more to read. No one can tell you what you hear. If it sounds better to you with the LED than that is great, and in the end that's all that matters. The cause and effect of the claim are difficult for me to understand at this point. Just as important as being able to witness improved sound is for me to try to understand this phenomenon from an scientific standpoint.
-AudioEnthusiast
Steve's reply on CD format
From Steve's post:
Also, the data doesn't actually exist on the disc in binary, linear PCM form to begin with. Contrary to popular belief, a pit doesn't designate a 1 and a land a 0 (or vice versa). In other words, if the data is 10110 you don't have a pit/land/pit/pit/land.
>> Steve I think you should be more clear here on what you are calling data. I think you mean to say that the raw audio signal isn't encoded as pits and lands. This is true, however a rising edge on a pit is what defines a '1' and a land is a '0'. These '1's and '0's therefore refer to audio encoded with EFM and CIRC.
-AudioEnthusiast
From Steve's post:
Also, the data doesn't actually exist on the disc in binary, linear PCM form to begin with. Contrary to popular belief, a pit doesn't designate a 1 and a land a 0 (or vice versa). In other words, if the data is 10110 you don't have a pit/land/pit/pit/land.
>> Steve I think you should be more clear here on what you are calling data. I think you mean to say that the raw audio signal isn't encoded as pits and lands. This is true, however a rising edge on a pit is what defines a '1' and a land is a '0'. These '1's and '0's therefore refer to audio encoded with EFM and CIRC.
-AudioEnthusiast
Re: "shhhhhhhhhhhhhhhhhhhhhhhhhh"
My intent was to address the specific technical claims made by YBA in the sidebar.
If you'd care to offer up something on the other things you mentioned, I'd be happy to discuss them.
I've given it some thought and haven't been able to think of any way that the blue light would have any particular effect on jitter, BER, servos, etc.
First, the light's blue. Typically about 430 nm. The photodiodes used in the detection system have their peak sensitivities at or above the 780 nm wavelength of the laser diode so they have significantly less sensitivity down at 430 nm.
Second, the light output from the diode is constant so even if some of it manages to end up striking the photodiodes, I don't see how it would produce any noise to speak of.
Third, the laser system uses interferometry (the depth of the pits on the CD correspond to 1/4 wavelength of the 780 nm laser light accounting for the refraction index of the polycarbonate) to achieve a very high signal-to-noise ratio and a very high contrast level between pit/land transitions.
If you've any alternatives, let's hear 'em.
The only thing I've denied are the patently false technical claims made by YBA.
se
HarryHaller said:
My intent was to address the specific technical claims made by YBA in the sidebar.
If you'd care to offer up something on the other things you mentioned, I'd be happy to discuss them.
I've given it some thought and haven't been able to think of any way that the blue light would have any particular effect on jitter, BER, servos, etc.
First, the light's blue. Typically about 430 nm. The photodiodes used in the detection system have their peak sensitivities at or above the 780 nm wavelength of the laser diode so they have significantly less sensitivity down at 430 nm.
Second, the light output from the diode is constant so even if some of it manages to end up striking the photodiodes, I don't see how it would produce any noise to speak of.
Third, the laser system uses interferometry (the depth of the pits on the CD correspond to 1/4 wavelength of the 780 nm laser light accounting for the refraction index of the polycarbonate) to achieve a very high signal-to-noise ratio and a very high contrast level between pit/land transitions.
If you've any alternatives, let's hear 'em.
The only thing I've denied are the patently false technical claims made by YBA.
se
Truth vs. B.S.
You know, I don't mind Peter (you need a better picture)
or anyone else saying they like the sound of their CD
when bathed in pink, purple, or blue light. I think I'm
becoming more of a grumpy old "objectivist" as I get older,
but I've heard effects from some tweeks that I can't
explain (as well as noticing absolutely ZERO effect from
many, many more...). Let other people try it and eventually
it will sort itself out. I don't track the "high end" rags that
well, but I've not heard much about green markers lately.
HOWEVER, I do get VERY annoyed with things like this
clip from Stereophile, where some supposed guru designer
speaks ex cathedra about the "science" behind some
tweek and it's total unabashed B.S. No possibilty of being
true. And the scary thought is that if he really believe
what hes says, he has no concept how the technology
behind his product works (i.e., how audio bits are encoded
and recorded onto a CD)
Yikes, we are running headlong into a self-referenced
postmodern culture where there is no concept of objective
truth/reality and everything has truth/meaing only if it is
meaningful to me.
You know, I don't mind Peter (you need a better picture)
or anyone else saying they like the sound of their CD
when bathed in pink, purple, or blue light. I think I'm
becoming more of a grumpy old "objectivist" as I get older,
but I've heard effects from some tweeks that I can't
explain (as well as noticing absolutely ZERO effect from
many, many more...). Let other people try it and eventually
it will sort itself out. I don't track the "high end" rags that
well, but I've not heard much about green markers lately.
HOWEVER, I do get VERY annoyed with things like this
clip from Stereophile, where some supposed guru designer
speaks ex cathedra about the "science" behind some
tweek and it's total unabashed B.S. No possibilty of being
true. And the scary thought is that if he really believe
what hes says, he has no concept how the technology
behind his product works (i.e., how audio bits are encoded
and recorded onto a CD)
Yikes, we are running headlong into a self-referenced
postmodern culture where there is no concept of objective
truth/reality and everything has truth/meaing only if it is
meaningful to me.
My opinion on why some 'mid-fi' companies use the overhead diode..
I saw a few posts which stated that Magnavox (I believe) has or had a separate LED illuminate the CD. My old Fisher AD-822 has a red LED.
I think that the reason why these 'mid-fi' companies have employed these LEDs is solely for cosmetic purposes. In these products the light from the LED is reflected onto the label side of the disc, not the data side. In a lot of products, the center portion of the CD tray has an around 2 inches of clear plastic which permits the user to see the disc and the light from the LED. When the LED is ON, it is easy to see the disc spinning since the light hitting the disc label produces a strobe effect to the human eye - one source of user feedback that the motor is spinning and playing music. In this scenario, the light from the LED is obstructed from the photodiode by the CD itself.
-AudioEnthusiast
I saw a few posts which stated that Magnavox (I believe) has or had a separate LED illuminate the CD. My old Fisher AD-822 has a red LED.
I think that the reason why these 'mid-fi' companies have employed these LEDs is solely for cosmetic purposes. In these products the light from the LED is reflected onto the label side of the disc, not the data side. In a lot of products, the center portion of the CD tray has an around 2 inches of clear plastic which permits the user to see the disc and the light from the LED. When the LED is ON, it is easy to see the disc spinning since the light hitting the disc label produces a strobe effect to the human eye - one source of user feedback that the motor is spinning and playing music. In this scenario, the light from the LED is obstructed from the photodiode by the CD itself.
-AudioEnthusiast
Re: Steve's reply on CD format
That's incorrect. That would be the case if the EFM data were NRZ coded (Non Return to Zero). But instead, the EFM data is NRZI coded (Non Return to Zero Inverted). With NRZI coding, the 1s only represent pit/land transitions, while the pits and lands themselves alternately represent 0s.
For example, if the EFM data were:
0100010010001
The 1s are where the pit/land transitions occur. And let's say that we're heading into the first 1 from a land. Then we hit the first 1 which means we have a transition from a land to a pit. The pit represents the 0s until we get to the second 1, where we transition from a pit to a land. After the second 1, the land now represents the 0s until we get to the third one, after which the pit represents the 0s, and so on.
In other words, 1 represents neither a pit nor a land. Merely a transition from one to another, regardless of which one to which other.
se
AudioEnthusiast said:
That's incorrect. That would be the case if the EFM data were NRZ coded (Non Return to Zero). But instead, the EFM data is NRZI coded (Non Return to Zero Inverted). With NRZI coding, the 1s only represent pit/land transitions, while the pits and lands themselves alternately represent 0s.
For example, if the EFM data were:
0100010010001
The 1s are where the pit/land transitions occur. And let's say that we're heading into the first 1 from a land. Then we hit the first 1 which means we have a transition from a land to a pit. The pit represents the 0s until we get to the second 1, where we transition from a pit to a land. After the second 1, the land now represents the 0s until we get to the third one, after which the pit represents the 0s, and so on.
In other words, 1 represents neither a pit nor a land. Merely a transition from one to another, regardless of which one to which other.
se
Re: Re: Re: Steve's Reply on CD Format
No problem. And thank you for not taking offense simply because I disagreed.
The main point I was trying to make in my previous post was that there's simply nothing on the disc at an actual binary bit for bit level which means that there's nothing on the disc that equates in any way to the LSB of the actual PCM audio data.
se
AudioEnthusiast said:
No problem. And thank you for not taking offense simply because I disagreed.
The main point I was trying to make in my previous post was that there's simply nothing on the disc at an actual binary bit for bit level which means that there's nothing on the disc that equates in any way to the LSB of the actual PCM audio data.
se
This is how I see it
I found this on a usenet post on the same subject,
"Across a typical CD, less than a dozen uncorrected but concealed errors may be expected, and zero or one unconclealed (muted) errors. Concealed errors consist of interpolated values between correct levels, and it's not being very bold to suggest that they are TOTALLY inaudible, consisting as they do of sub-millisecond values of roughly correct amplitude occurring perhaps ten times in an hour of music. So we have for all intents and purposes a perfect data recovery system."
This is inline with my previous knowledge, and what I believe is commonly held as fact. Given the almost perfect accuracy of the recovered data, it seems to me that WHATEVER effect the blue LED has, it is in producing a recovered signal that is LESS ACCURATE than what is typically sent to the DAC. It is, however, apparently pleasing to some people. Maybe in the same way that tubes are preferable to some over SS. I have yet to try it.
I found this on a usenet post on the same subject,
"Across a typical CD, less than a dozen uncorrected but concealed errors may be expected, and zero or one unconclealed (muted) errors. Concealed errors consist of interpolated values between correct levels, and it's not being very bold to suggest that they are TOTALLY inaudible, consisting as they do of sub-millisecond values of roughly correct amplitude occurring perhaps ten times in an hour of music. So we have for all intents and purposes a perfect data recovery system."
This is inline with my previous knowledge, and what I believe is commonly held as fact. Given the almost perfect accuracy of the recovered data, it seems to me that WHATEVER effect the blue LED has, it is in producing a recovered signal that is LESS ACCURATE than what is typically sent to the DAC. It is, however, apparently pleasing to some people. Maybe in the same way that tubes are preferable to some over SS. I have yet to try it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.