h a l f b a k e r ynon-lame halfbakery tagline
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Digital LP
All the hipster benefits of real vinyl on real old-school turntables, but with digital audio. | |
A compressed digital audio data stream is fed into a modem, producing an analogue signal which drives the cutting head on a vinyl mastering lathe. The LPs are pressed as usual onto vinyl.
For playback, a normal quality turntable and pickup are used. The pickup sends its analogue signal to a modem,
which provides the original compressed data stream to feed into a D-A converter. This drives the speakers via the usual amp &c.
Wiki "Modem" arcticle.
https://en.wikipedia.org/wiki/Modem
[scad mientist, Feb 28 2016]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
Could play those LPs using the currently available optical
turntables. [+] |
|
| |
You forget a coil heater to make the sound warmer. |
|
| |
// a coil heater to make the sound warmer // |
|
| |
Recently I heard a musician saying his tubes amp sounds great because the heat of the filaments is above normal and you know, warmer tubes produce warmer sound :-p |
|
| |
Needs a layer or two of error correction /
concealment. |
|
| |
Doesn't the needle emit some sound directly? That's not a
problem when the sound matches what's coming out of the
amp, but it might be an issue if it was screechy modem
sounds. |
|
| |
I don't think you'd have the
bandwidth to play anything in real time using this
method. Pretty sure you'd have to play the track first
to "download" the data into a buffer and play it from
that. I'm also guessing that a song would need several
LPs to get all that data in. |
|
| |
I could be wildly off on this since I'm pretty rusty
with
my digital audio knowledge, but I'm sure it's not a
one to one tradeoff, analog vibration to data on off
signals. For instance, a 32 bit waveform would
require you take one pulse, or one wiggle of the
analog record player needle and chop it into a bunch
of pieces that you
can put back together to draw that wave again. So to
draw one hump of that wave, you'd need the needle
to wiggle, that is, indicate on or off, 32 times,
compared to the analog playback just having the
needle move once. |
|
| |
But it's a cute idea as long as you don't mind it not
playing in real time, or playing very, very, very slowly
in real time so: [+] |
|
| |
I also wondered about bandwidth, but I assumed that compression could overcome that. I'm thinking I can live-stream digital audio down my telephone wire at a much higher quality than analog audio over the same wire. |
|
| |
You're sending a lot more information with that
digital audio. With the analog you're just wiggling
your power up and down to make the speakers
shake. |
|
| |
The way you digitize an analog waveform is to cut it
up in little slices like a loaf of bread. So if the top of
the wave is say 5db and the bottom is -4db. You have
the first slice taken at 1db, second at 2 db till you
get to the peak of the wave then you go back down.
The higher the resolution the more slices you're
taking and the smoother picture you get of the wave. |
|
| |
Each little slice is a piece of information you need to
send just to convey one little wiggle of the needle on
a record player. |
|
| |
But like I said, impractical but interesting inventions
are pretty much what the HP is all about. I'm very
proud of my mechanical internet which as I pointed
out, would do the job at 1 millionth the efficiency at
several million times the price. |
|
| |
[dr3] The reason TV and radio has converted to digital is that you can squeeze far more data into the same (analog) bandwidth. Mostly because of compression. |
|
| |
But the problem with vinyl is that the bandwidth changes from rim to hub... |
|
| |
[edit: I very nearly posted an idea just now for a vinyl record that played with constant linear velocity, changing the rpm to match the needle position. But then I googled and found this has already been done, and was a commercial failure.] |
|
| |
//The reason TV and radio has converted to digital is
that you can squeeze far more data into the same
(analog) bandwidth. Mostly because of
compression.// |
|
| |
Well, yes, but you're using the same method at one
level of sending information with both analog and
digital: electrical pulses, be they translated into
LEDs, lasers, radio waves etc. |
|
| |
With analog you're not fully utilizing how fast those
pulses can be, which is very very fast. |
|
| |
But you do agree that real time digital playback from
vinyl isn't a possibility, at least at standard 33 1/3rd
RPM, yes? |
|
| |
If somebody was really clever they'd be able to figure
this out: "How fast would a mechanical audio
turntable need to spin to get say, 1gps of data
output?" |
|
| |
And by the way, what portion of the people reading
this know that a turntable turns at 33 1/3rd
revolutions per minute? And how many can say how
fast a 7" record turns? Hint: The RPM of the smaller
record is what they're actually referred to. I'm
guessing that those born after 1990 wouldn't know
this. |
|
| |
//But you do agree that real time digital playback from vinyl isn't a possibility, yes?// |
|
| |
No, I don't see why it wouldn't be possible. Analog signals have bandwidth too - an FM broadcast occupies a certain amount of the EM spectrum. A digital radio broadcast takes up less of the spectrum. |
|
| |
So 33 1/3rd would give you the same bandwidth? Eh,
maybe, but that would surprise me. |
|
| |
I'm thinking you've got a limited amount of "wiggles"
on vinyl, a few hundred per inch, before the needle
starts bouncing out of the groove like a car hitting a
speed bump. Electrical pulses have no such
limitation. |
|
| |
An old, pre DVD laser disc, which is basically an
optical
vinyl player
with a laser instead of a needle
to get over the physical limitations of a wiggly slot
with a needle in it
turns at 1,800 RPM so I think you might be mistaken.
Of course those carry video information as well as an
audio track but 1,800 is a lot faster than 33 & 1/3rd. |
|
| |
Well, human hearing goes up to about 20kHz, so I assume a vinyl record has at least 20kHz of bandwidth at its smallest radius. Probably double that at the rim. |
|
| |
So I suppose that implies a bitrate of less than 20kbits/sec, which would be enough for low-quality audio reproduction. 'Telephone quality' is usually 8kbits/sec, but MP3s can be as high as 320kbits/sec, so you're right in that it wouldn't be as high quality as the analog signal alone. |
|
| |
I agree with [mixtela]. Here's why. A 56k modem can
transmit over a phone line that has MUCH poorer audio
quality than an LP. The phone line is limited to about
3.2KHz and in digital phone systems is sampled at 8 bit
8KHz. LPs have a much higher bandwidth. If we just say
that they are as good as a CD, that's 16 bit sampling at
44kHz. which is about 10 times. I'm sure an LP is less
noise than the 8 bit quantization causes, but it might not
actually be as good as 16 bit, so lets say 5x instead of
10x just to be safe. That means that a 280kps modem is
feasible, which is a pretty high quality mp3. I'm guessing
you could push it a lot farther than this, but I'm pretty
confident that 280kbps is achievable. |
|
| |
Actually if you presume the analog signal is a sine wave or a
summation of sine waves then you need only a few data
points to approximate it. |
|
| |
//I'm pretty confident that 280kbps is achievable.// |
|
| |
Are you sure enough to call this "solved with 100%
certainty"? We're talking real time, same resolution as
a regular record with no increase in speed. I'm still
dubious. |
|
| |
If you're not sure, I need to call somebody on this. I'm
pretty sure anybody who works at Bell Labs could
answer without any hesitation. I'd bet even the
receptionist or the parking lot attendant probably
knows. This is exactly what they deal in. |
|
| |
Who wants to contact them? If nobody else will, I'll
write them an email. I'm busy today but now this
question is bugging me. |
|
| |
Does somebody really know or do I write these guys? |
|
| |
Ok, I'll write them. I'll post what I get a definitive. |
|
| |
Go for it, I'm interested too. |
|
| |
I looked up some numbers to back up what I was saying earlier. For radio broadcasts, FM stations are spaced 500kHz apart. DAB radio in the UK is variable based on the desired quality, but the BBC broadcasts ten 128kbps stations in a 1.7MHz block. So the digital broadcasts are able to fit nearly three times as much audio into a given chunk of spectrum. |
|
| |
There's an equation (Shannon-Hartley) that lets you calculate the maximum bitrate for a noisy channel, but I usually just go with bitrate ~= bandwidth. 20kHz was a worst-case guess, and googling tells me that some LPs supposedly have bandwidths of 150kHz. If so, you could store a 128kbps stream in that. |
|
| |
[scad] What you're saying makes sense, but I don't know, weren't phone lines improved to enable higher modem speeds? Otherwise they could have started off at 56k instead of slowly speeding up. |
|
| |
Sent this. I'll let you know if they respond. |
|
| |
Can you direct me to someone in your company might
be
able to answer a general technical question? I was
part of a discussion about data storage and
specifically how very old data storage methods might
translate to modern data storage numbers. |
|
| |
There's a famous picture of Bill Gates holding a CD
atop a stack of paper several yards high representing
how much storage capacity this new method
represented compared to printed pages. |
|
| |
My question is in a similar vein. How much data might
theoretically be stored on an old vinyl record and
what method might be used, and what bandwith for
real time transmission might be possible with a
record turning at the old standard 33 1/3rd rpm? |
|
| |
This is of course just for fun but I thought somebody
in your company might enjoy taking a stab at this
thought experiment. |
|
| |
My husband collects. He just does. Mostly vinyl. We have
an entire room dedicated to vinyl. Actually a room and a
half. The rest of our flat has the walls lined with shelves of
CD's and DVD"s, except for the overstuffed bookcases which
surprisingly enough, hold books. |
|
| |
My husband has not played anything on his turntable since
we reunited 5 years ago, so maybe even 10 years or more.
Not once. We just look at them. We stare at them and
watch his collection grow. It gets dusty. He amuses himself
by dusting his collections once a month. |
|
| |
The moral of this story is; Don't let your husband collect
vinyl, if he's not going to actually play it. It just takes up
valuable, usable space, that I could use for other
meaningful purposes, like hanging tapestries on the wall
or something. |
|
| |
So a + for the idea, if it would motivate my husband to
buy, and listen, to vinyl. Not just stare at it. |
|
| |
Back when I was in the music biz in my youth, CDs
and cassettes
were how people purchased their music, but I do
have in my possession a vinyl pressing of my band
from Italy. There were also vinyl pressings available
in South Korea though I never got my hands on one. |
|
| |
For some reason I really treasure this goofy thing. I
grew up with vinyl so I'm glad I can say I was actually
part of the vinyl age sort of. |
|
| |
By the way, if you ever play an actual gold or
platinum record you'll find it's probably not actually
the band shown. I've never done it as you'd have to
pry the things out of their frames and they're glued
on so you'll probably wreck the thing but I'm told,
from people who probably got drunk and decided to
play their gold records that it's invariably something
completely un-related. If you're a heavy metal band
for instance they'll give you a Barny's greatest hits LP
with your label on it. |
|
| |
There's your useless trivia for the day. |
|
| |
The speed increase was primarily a result of them finding
ways to maximize the use of the available audio
bandwidth. Some line quality improvement did come into
play, but that was early on: the biggest improvement
being the legal action that forced the phone companies
to allow direct connection to the phone lines rather then
relying on coupling using audio through a standard
handset. For those interested in an good overview, the
Wikipedia Modem article is interesting [link]. After
reading that I did find that my previous analysis had a
bad assumption. 56kbps modems actually use more than
the 3.2kHz bandwidth of the telephone system, so that
rate could only be used for downlink from the digital
64kbps line at the phone company across the short
analog line to the house which is not limited to 3.2kHz. I
think this means that the ISP had to have a direct digital
link to the phone company to work. The fastest modems
that could work across the 3.2kHz limited phone system
were the 33.6kbps modems. Apparently this is very close
to the theoretical "Shannon limit". |
|
| |
So where I said that 280kbps is a pretty sure thing, that
should have been 168kbps (33.6 x 5), which still isn't too
bad for mp3s. Not to mention that I think I was being
very conservative with the bandwidth of an LP compared
to a phone line by only applying a 5x multiplier. |
|
| |
Can you break down for me what the actual signal in
a modem is? I guess it's conveying on/off signals by
varying frequency and amplitude but
how is it doing it? |
|
| |
So if I want to transmit:
0100100.01001010....etc where do I put
these on off signals in a carrier wave or whatever it is
that's traveling through a modem? Would you
have a certain frequency bleep for a zero and
another one for a 1? Basically my question is how
does analog data transmission work? |
|
| |
As I understand it, a modem turns data into bleeps of different frequencies. But each bleep can carry more than one bit of data. I just glanced at the wiki page (will read it later) and I wasn't aware that a 33.6k modem transmits as many as 10 bits per symbol - that implies there are 1024 different frequencies it uses. |
|
| |
So your example of 0100100.01001010 could be transmitted in less than two beeps (not accounting for all the error checking). |
|
| |
//there are 1024 different frequencies it uses. |
|
| |
So your example of 0100100.01001010 could be
transmitted in less than two beeps (not accounting for all
the error checking).// |
|
| |
Ok, now we're getting somewhere. (I think) Does amplitude
figure into it as well? That is, does a certain wave height
signify something different than a different wave height? |
|
| |
There's frequency shift keying, amplitude shift keying, phase shift keying, and many more combinations / more complicated forms of modulation that wikipedia will be much better at explaining than I am. |
|
| |
But to get to the bottom line, if you record a 56K or a 256k modem signal into a record and play it back, I know it'll sound the same to the ear, but how much of that data is actually preserved by a little needle wiggling in a plastic slot? |
|
| |
Now I'm wondering if anyone has ever constructed a data stream, which, when sent to a modem, produces a musical sound? |
|
| |
+. Nice idea in itself. Given for posting the link on portable
flagpole. |
|
| |
Why thank you most gracious sir. |
|
| |