A really good article about vinyl, digital files and cd's

I think part of what makes vinyl a more enjoyable medium for many is the ambiance. The background noise. If you've ever been to a concert, you KNOW that it's not quiet, even when it's quiet. All those people, all that breathing, the humming, singing along, whatever. There is no silence in live music. And even a good record is still not totally silent. Unlike CD, where the music seems to come from nowhere and nothing, with a record, it's more like being there. Even if there never really was a "there."
 
"It is a fact that vinyl sounds different from CDs."

Not for my better recordings. I was setting the gain matching feature on my preamp last week by playing George Winston's Autumn album synchronized with CD and find little difference.

While the upper bass was a touch thinner on the LP, the two recordings sounded remarkable similar. That experience also helped me choose the best loading for the Dynavector.
 
I really find nothing to disagree with in the article. I think a lot of people confuse the recording sampling rate (44khz, 192khz or ?) with the kbits/s rate of compressed music. regardless of the kbits/sec, compression = loss, regardless or what they call the format. and the loudness wars as it were, is only half understood, its not just make the format louder (in the digital world, the relative volume of a section is just a number, very easy to do, very easy to apply to a range, heck I have free software from the web that lets me 'remaster' mp3s) but also to aid compression by making ALL facets of the recording, the same volume. "there is no dark side of the moon, really. as a matter of fact, its all dark" understand the implications behind that quote, and you understand the problem. Or to put it another way, doug gray taps an unmiked tambourine while at the same time Chris hicks rips off a blinding lick on a marshall stack. in a live performance, you get the subtlety of both without engineering the hand instrument to sound like a box of castanets just got dumped on your head

FWIW, when CDs came out, it was true at first the rendering was faulty, the mixing was faulty, but the players were pretty stout. now that we are in the future, the players are junk, unless you find a $2000 'audiophile' version somewhere, but at least they finally learned what to do with the bits...so for the 3-6 months the (any machine built after say 2005) player will last, you will like it.

FWIW2: I find that CD players, 1-12 disc built 1990-1999 are in fairly high demand as those who have re-discovered 2 channel goodness want CDs as well as vinyl. Kudos to the pioneer (and optimus) makes that take the 6 disc cartridge - people really love those. I blame this backlash on those who have <insert digital media player of choice> plugged into the good stereo with the good speakers. or in other words, if you want to hear just how shitty your laptop or ipod or ? sound in real life, plug it into my office sx890 and carvin monitors and take a listen. I keep a trash can close by so when you vomit....
 
I really find nothing to disagree with in the article. I think a lot of people confuse the recording sampling rate (44khz, 192khz or ?) with the kbits/s rate of compressed music. regardless of the kbits/sec, compression = loss, regardless or what they call the format.
Not quite. There is lossless compression and lossy compression. Lossless compression, used by audio file formats like FLAC and general-purpose compression like ZIP, uses various techniques to reduce the size of the file whilst not losing any data. For example, a lot of audio data contains repeated information.

KKKKKIIIIINNNNNDDDDDD OOOOFFFF LLLLLIIIIKKKKEEEE TTTTHHHHHHHIIIIISSSS​

You can represent the above by indicating the character followed by a count, like this:

K5I5N5D6 O4F4 L5I4K4E4 T4H7I5S4​

That takes less space, but represents exactly the same information. You can reliably always get back the original data, so it's losslessly compressed.

On the other hand, lossy compression actually removes data -- like the MP3 audio format -- whilst (hopefully) not being (too) noticeable. For example, my example could be lossy compressed to be:

KND OF LKE THS​

It's still essentially readable, but some data has been irretrievably lost. It is lossy compression.
 
So it actually is a "digital is better article", the summary at the end states as much. Now if he had stated something like "But don't say it's more clinically accurate" then I might believe there wasn't an agenda to the article. I see this as a heavily biased article, and I'll take vinyl over digital any day, the sound IS better.

Or, as Kees Immink says: "Some people like marmalade and some people like mustard. If people like to listen to vinyl, do so, enjoy life. But don't say that the sound is better."
 
Not quite. There is lossless compression and lossy compression. Lossless compression, used by audio file formats like FLAC and general-purpose compression like ZIP, uses various techniques to reduce the size of the file whilst not losing any data. For example, a lot of audio data contains repeated information.

KKKKKIIIIINNNNNDDDDDD OOOOFFFF LLLLLIIIIKKKKEEEE TTTTHHHHHHHIIIIISSSS​

You can represent the above by indicating the character followed by a count, like this:

K5I5N5D6 O4F4 L5I4K4E4 T4H7I5S4​

That takes less space, but represents exactly the same information. You can reliably always get back the original data, so it's losslessly compressed.

On the other hand, lossy compression actually removes data -- like the MP3 audio format -- whilst (hopefully) not being (too) noticeable. For example, my example could be lossy compressed to be:

KND OF LKE THS​

It's still essentially readable, but some data has been irretrievably lost. It is lossy compression.

DDDDDAAAAAVVVVVEEE

TTTHHHAANNNKKKSSS

FFFOOORRR

TTTHHHEEE

EEEXXXPPPLLLAAANNNAAAtion. ;-)”
 
Not quite. There is lossless compression and lossy compression. Lossless compression, used by audio file formats like FLAC and general-purpose compression like ZIP, uses various techniques to reduce the size of the file whilst not losing any data. For example, a lot of audio data contains repeated information.

KKKKKIIIIINNNNNDDDDDD OOOOFFFF LLLLLIIIIKKKKEEEE TTTTHHHHHHHIIIIISSSS​

You can represent the above by indicating the character followed by a count, like this:

K5I5N5D6 O4F4 L5I4K4E4 T4H7I5S4​

That takes less space, but represents exactly the same information. You can reliably always get back the original data, so it's losslessly compressed.

On the other hand, lossy compression actually removes data -- like the MP3 audio format -- whilst (hopefully) not being (too) noticeable. For example, my example could be lossy compressed to be:

KND OF LKE THS​

It's still essentially readable, but some data has been irretrievably lost. It is lossy compression.
yessir, that is true and it works well for repeatable data, be it data, audio, video, stills, etc.

however, unless your album consists of 1 khz tone for 40 minutes, it dont work exactly that way (and I only say this cuz I wrote quite a few algorithms for such data compression whilst at IBM) what we ALSO do is take (using your example) the relative font size of each character and instead of fonts from say 4pt to 26 pt, reduce them to font ranges such that 4-8pt is one, 9-14 is another, 16-20, 22-26, then group them, but when the font ranges change from say lower to next to higher etc, create a transistion artifact since the human ear CAN hear the tonal changes quite well. you hear the artifact as more as a slew - like an airplane turning towards you vs away, but hear the difference you will. the best example I have ever found is a pure analog recording of santana black magic woman, vs a digital one.

but if you are not sure you can hear it, go see it instead. in video, even 1080p, in dark passages, even on a 4k ultra tv, blacks and reds are VERY VERY VERY easy for the human eye to see variations in (that particular science is not in my pay grade) but very very very hard to compress and render and even the the lastest and greatest will have a digital cloud or pixelization effect. note, that this is irrespective of the actual number of datapoints in the signal.
 
I see this as a heavily biased article, and I'll take vinyl over digital any day, the sound IS better.
Then I assume you never listen to any records newer than 1981 or so? Because after then, digital audio was most likely involved somewhere in the production of your vinyl records. Heck, quite a few LPs from the mid-1980s proudly proclaimed "DIGITAL" on their cover.
 
yessir, that is true and it works well for repeatable data, be it data, audio, video, stills, etc.

however, unless your album consists of 1 khz tone for 40 minutes, it dont work exactly that way (and I only say this cuz I wrote quite a few algorithms for such data compression whilst at IBM) what we ALSO do is take (using your example) the relative font size of each character and instead of fonts from say 4pt to 26 pt, reduce them to font ranges such that 4-8pt is one, 9-14 is another, 16-20, 22-26, then group them, but when the font ranges change from say lower to next to higher etc, create a transistion artifact since the human ear CAN hear the tonal changes quite well. you hear the artifact as more as a slew - like an airplane turning towards you vs away, but hear the difference you will. the best example I have ever found is a pure analog recording of santana black magic woman, vs a digital one.
Quantization error?

In modern digital audio systems, dithering is done to convert quantization error to low-level noise.
but if you are not sure you can hear it, go see it instead. in video, even 1080p, in dark passages, even on a 4k ultra tv, blacks and reds are VERY VERY VERY easy for the human eye to see variations in (that particular science is not in my pay grade) but very very very hard to compress and render and even the the lastest and greatest will have a digital cloud or pixelization effect. note, that this is irrespective of the actual number of datapoints in the signal.
Digital video uses lossy compression; what you're seeing are lossy compression artifacts.
 
Quantization error?

In modern digital audio systems, dithering is done to convert quantization error to low-level noise.

Digital video uses lossy compression; what you're seeing are lossy compression artifacts.
Love ya dave.

first statement shows that it is a loss of something from the original. even if we engineer it to be very small, its still a very small error. aim small, miss small.

second sentence, in a netflix stream, yup. cue up a blue (or red) ray dvd and you can still see it (most streaming loss is due to frame compression anyways)

i dont think we are necessarily disagreeing
 
first statement shows that it is a loss of something from the original. even if we engineer it to be very small, its still a very small error. aim small, miss small.
It's a "loss" which for 16 bits and above is lower level than the typical background noise in typical room-temperature or above analog circuitry. In other words, there is equivalent or greater loss in a working analog system.

Therefore, using lossless audio compression, the input waveform is (measurably, if you try it) equal to the output waveform above the noise floor. The same is notionally true of any analog reproduction, though typical analog mechanisms are also subject to all manner of distortion.
second sentence, in a netflix stream, yup. cue up a blue (or red) ray dvd and you can still see it (most streaming loss is due to frame compression anyways)
A DVD is lossy MPEG-2 compression; BluRays are usually MPEG-4. On consumer devices, you normally won't ever see uncompressed video.
i dont think we are necessarily disagreeing
I think we're in violent agreement. :biggrin:
 
part of the article also mentioned that perversely, artists and listeners, prefer some sort of distortion and the preservation of the distortion is part of the intended product...
 
Then I assume you never listen to any records newer than 1981 or so? Because after then, digital audio was most likely involved somewhere in the production of your vinyl records. Heck, quite a few LPs from the mid-1980s proudly proclaimed "DIGITAL" on their cover.
Oh I do all the time, but you see, converting digital back to analogue removes/covers/smooths out the digital artifacts. It also works if you transfer CD to tape, try it some time.
 
Back
Top Bottom