Does vinyl really sound better?

richardhomer

I am prepared to believe (and indeed have heard) that a higher sampling rate sounds better than CD - but I don't believe this is anything to do with high frequency extention - the argument which is posited by the analogue brigade.

The limit of reproducing high frequencies (assuming the source is capable of capturing information beyond 20Khz) is set by the band width of the amplification and speakers. And the most analogue part of the whole process - us!

The reason that higher sampling rates sound more 'real' is because the original signal is simply 'captured' more accurately - the values between sampling points are less of a 'guess'.

Cacofonix

monquixote said:

Cacofonix said:

monquixote said:

Danny1969 said:

Once that resolution gets lower than the noise floor then it effectively doesn't exist it just becomes entropy. If you watch Monty's vid from about ten minutes in you will see that noise floor and bit depth are equivalent so there is no resolution on vinyl beyond the noise floor. If it was greater then the dynamic range would be higher.

I'm not sure that's the case.  That's like saying a page that's been scribbled on is unreadable.  It might not be, and an expert can extract audio data from white noise: there was even a job ad for a forensic noise analyst for precisely that purpose.  If it can be done by digital means, it can certainly be done by human audio processing systems.

Secondly, there is a surmise that you can't hear super-sonic noise.  That is also not correct, there is equipment that can be purchased which is designed to detect whether you are being subjected to sonic attack.  Even though it may be above the frequencies considered within hearing range, that does not mean it does not affect you.  Subsonics are known to do so, so why not supersonics?

I think you might have been watching too much CSI

Below the noise threshold there can be no information. All that you will find is white noise, pure entropy and no information whatsoever. If you could you would be breaking the laws of thermodynamics. 

If you are interested to understand why this is then read up on the Shannon Hartley Theorem.

http://en.wikipedia.org/wiki/Shannon–Hartley_theorem

As to CSI, no, although it was a very funny joke indeed.  There was an advert in SOS for an audio engineer to work in the force to isolate fraudulent recordings, and the primary focus was on the underlying EM fields which vary day to day, and which can 'date' a recording with a high degree of accuracy. Below the noise floor.

LastMantra

Vinyl smells better.

ICBM

LastMantra said:

Vinyl smells better.

It usually smells of weed, in my experience...

Cirrus

richardhomer said:

I am prepared to believe (and indeed have heard) that a higher sampling rate sounds better than CD - but I don't believe this is anything to do with high frequency extention - the argument which is posited by the analogue brigade.

The limit of reproducing high frequencies (assuming the source is capable of capturing information beyond 20Khz) is set by the band width of the amplification and speakers. And the most analogue part of the whole process - us!

The reason that higher sampling rates sound more 'real' is because the original signal is simply 'captured' more accurately - the values between sampling points are less of a 'guess'.

All sounds are made up of sine waves, and 44.1k is enough to accurately capture ANY sound up to 22.05kHz. Nothing is lost. 

 So why the difference? It's because at 44.1kHz sampling rate, you need a very sharp low pass filter to make sure no frequencies above 22.05kHz actually reach the converter - if they do, they're reflected down into the audio range in a manner not unlike a ring modulator ie totally non harmonic and bad sounding. The filter needs to be analogue which means it's not a perfect filter. To be basically silent by 22.05kz it needs to start acting on frequencies much lower - I don't have the numbers offhand but I'd guess it'd be 3db down by 20k then as steep as possible. 

This means a couple of things. Firstly, it's slightly attenuating frequencies below 20k in the audio range, and secondly it's introducing frequency dependent phase shifts and resonances which result in less natural high end. It also means the entire signal is running through a shitty cheap little analogue filter which may or may not have unintended side effects in other areas. At higher sampling rates, the filter doesn't need to be as steep which results in more natural high end. 

The jury is out on all the ultrasonic stuff but it may well affect our perception of lower frequencies, so that's another possible benefit. 

A side effect of the above is that cheaper converters, with poor filters, often sound better at higher sample rates while top end pro stuff has much less of a sonic hit at 44.1 because the filters are simply of a higher quality. Then there's other quality-dependent effects like clock jitter and the slew rate of the op amps which can also limit fidelity.

 Bit depth is much more important IMO. At 16 bit, I can hear some detail, the really quiet stuff like cymbal tails and room reverb, sounds flat and less real. 24 bit and higher sounds better.

LastMantra

ICBM said:

LastMantra said:

Vinyl smells better.

It usually smells of weed, in my experience...

An LP cover if good for skinning-up on, CD's are too small.

monquixote

Cacofonix said:

I see.  So there is no such thing as a stereogram, then?  That's the visual equivalent.

A stereogram has information encoded in it to be found.

White noise by definition does not. 

imalone

monquixote said:

Cacofonix said:

monquixote said:

Danny1969 said:

Once that resolution gets lower than the noise floor then it effectively doesn't exist it just becomes entropy. If you watch Monty's vid from about ten minutes in you will see that noise floor and bit depth are equivalent so there is no resolution on vinyl beyond the noise floor. If it was greater then the dynamic range would be higher.

I'm not sure that's the case.  That's like saying a page that's been scribbled on is unreadable.  It might not be, and an expert can extract audio data from white noise: there was even a job ad for a forensic noise analyst for precisely that purpose.  If it can be done by digital means, it can certainly be done by human audio processing systems.

Secondly, there is a surmise that you can't hear super-sonic noise.  That is also not correct, there is equipment that can be purchased which is designed to detect whether you are being subjected to sonic attack.  Even though it may be above the frequencies considered within hearing range, that does not mean it does not affect you.  Subsonics are known to do so, so why not supersonics?

I think you might have been watching too much CSI

Below the noise threshold there can be no information. All that you will find is white noise, pure entropy and no information whatsoever. If you could you would be breaking the laws of thermodynamics. 

If you are interested to understand why this is then read up on the Shannon Hartley Theorem.

http://en.wikipedia.org/wiki/Shannon–Hartley_theorem

You can though extract a signal beneath the sample noise floor by using a sufficiently large sampling time. Not that it has any bearing. (It has to be a constant signal within the system bandwidth anyway and it's not something your ears can do.)

Anyway, you can't hear supersonic noise by definition (otherwise it would be sonic noise). The limits for hearing are the points at which the hearing threshold drops beneath the pain threshold, i.e. you can sense it if it's intense enough, but mainly through discomfort. As you might imagine this isn't very relevant to audio reproduction outside of Guantanamo bay.

monquixote

imalone said:

monquixote said:

Cacofonix said:

monquixote said:

Danny1969 said:

Once that resolution gets lower than the noise floor then it effectively doesn't exist it just becomes entropy. If you watch Monty's vid from about ten minutes in you will see that noise floor and bit depth are equivalent so there is no resolution on vinyl beyond the noise floor. If it was greater then the dynamic range would be higher.

I'm not sure that's the case.  That's like saying a page that's been scribbled on is unreadable.  It might not be, and an expert can extract audio data from white noise: there was even a job ad for a forensic noise analyst for precisely that purpose.  If it can be done by digital means, it can certainly be done by human audio processing systems.

Secondly, there is a surmise that you can't hear super-sonic noise.  That is also not correct, there is equipment that can be purchased which is designed to detect whether you are being subjected to sonic attack.  Even though it may be above the frequencies considered within hearing range, that does not mean it does not affect you.  Subsonics are known to do so, so why not supersonics?

I think you might have been watching too much CSI

Below the noise threshold there can be no information. All that you will find is white noise, pure entropy and no information whatsoever. If you could you would be breaking the laws of thermodynamics. 

If you are interested to understand why this is then read up on the Shannon Hartley Theorem.

http://en.wikipedia.org/wiki/Shannon–Hartley_theorem

You can though extract a signal beneath the sample noise floor by using a sufficiently large sampling time. Not that it has any bearing. (It has to be a constant signal within the system bandwidth anyway and it's not something your ears can do.)

Anyway, you can't hear supersonic noise by definition (otherwise it would be sonic noise). The limits for hearing are the points at which the hearing threshold drops beneath the pain threshold, i.e. you can sense it if it's intense enough, but mainly through discomfort. As you might imagine this isn't very relevant to audio reproduction outside of Guantanamo bay.

You can use spread spectrum as well, but human ears can't!

Skipped

monquixote said:

A stereogram has information encoded in it to be found.

Danny1969

Now those things sounded warm 

richardhomer

Cirrus;422725" said:

All sounds are made up of sine waves, and 44.1k is enough to accurately capture ANY sound up to 22.05kHz.

My understanding was that 44.1 kHz was chosen as it was the lowest sampling rate which could reproduce up to 20 kHz, which is accepted to be the 'limit' of audible sounds.

I believe that the sinewave is extrapolated from the sampling points by (to coin a phrase) 'joining the dots' in the digital to analogue converter - hence the higher sampling (and bit rate) - the more accurately the digital representation of the analogue wave form is.

I take the point that the higher you move the frequency needed to aggressively filter to avoid aliasing, the less intrusive it's audible inpact will be.

I'm no expert - the knowledge I have is from reading up on digital sound reproduction over the years. I have no education or professional training in it - just a fascination with how things work....

monquixote

richardhomer said:

Cirrus;422725" said:

All sounds are made up of sine waves, and 44.1k is enough to accurately capture ANY sound up to 22.05kHz.

My understanding was that 44.1 kHz was chosen as it was the lowest sampling rate which could reproduce up to 20 kHz, which is accepted to be the 'limit' of audible sounds.

Pop trivia fact 44.1 was chosen as it is a multiple of the framerates of the NTSC and PAL video standards as digital audio used to be recorded on specially adapted video recorders.

richardhomer

monquixote;422765" said:

Pop trivia fact 44.1 was chosen as it is a multiple of the framerates of the NTSC and PAL video standards as digital audio used to be recorded on specially adapted video recorders.

You probably need to correct these guys as well:

http://en.m.wikipedia.org/wiki/44,100_Hz

They seem to believe that was a coincidence, rather than the reason.

Danny1969

Everythings recorded at 48 these days as that suits the video boys too

Cirrus

richardhomer said:

I believe that the sinewave is extrapolated from the sampling points by (to coin a phrase) 'joining the dots' in the digital to analogue converter - hence the higher sampling (and bit rate) - the more accurately the digital representation of the analogue wave form is.

Yeah it is really interesting! The deal is even without the extra 'dots ' of a higher sample rate, the maths should extrapolate the same curve that would be described by the extra dots anyway... In theory at least. I seem to remember reading that the ear can distinguish timing differences that are much smaller than the ~0.05 milliseconds it takes for a single cycle of a 20kHz sound to occur - a quick google shows one paper suggesting 10-20 MICROseconds which would mean the timing resolution of our ears is up to 40kHz+ even if our cochlear aren't tuned to pick up repeating sine waves that high. I'm just freewheeling really, I've got no idea if our how this impacts anything, I just find it interesting that we can hear time shifts much more accurately than frequencies.

Edit; between autocorrect, typing errors and format problems, I hate ipadding on forums!

monquixote

richardhomer said:

monquixote;422765" said:

Pop trivia fact 44.1 was chosen as it is a multiple of the framerates of the NTSC and PAL video standards as digital audio used to be recorded on specially adapted video recorders.

You probably need to correct these guys as well:

http://en.m.wikipedia.org/wiki/44,100_Hz

They seem to believe that was a coincidence, rather than the reason.

No it doesn't it agrees completely with what I said (and is taught by the BBC as part of their broadcast engineering qualification)

 "It is simplest if the same number of lines are used in each field, and, crucially, it was decided that a sample rate that could be used on both NTSC (monochrome) and PAL equipment. Since NTSC has a field rate of 60 Hz, and PAL has a field rate of 50 Hz, their least common multiple is 300 Hz, and with 3 samples per line, this yields a sample rate that is a multiple of 900 Hz. For NTSC the sample rate is 5m × 60 × 3, where 5m is the number of active lines per field, which must be a multiple of 5 (the rest used for synchronization), and for PAL the sample rate is 6n × 50 × 3, where 6n is the number of active lines per field, which must be a multiple of 6. The sampling rates that satisfy these requirements – at least 40 kHz (so can encode 20 kHz sounds), no more than 46.875 kHz (so require no more than 3 samples per line in PAL), and a multiple of 900 Hz (so can be encoded in NTSC and PAL) are thus 40.5, 41.4, 42.3, 43.2, 44.1, 45, 45.9, and 46.8 kHz. The lower ones are eliminated due to low-pass filters requiring a transition band, while the higher ones are eliminated due to some lines being required for vertical blanking interval; 44.1 kHz was the higher usable rate, and was eventually chosen."

ICBM

Skipped said:

monquixote said:

A stereogram has information encoded in it to be found.

Danny1969 said:

Now those things sounded warm 

That's because they have no tweeters... so the frequency response is limited to under 10Khz.

They do sound nice though. That's the sort of thing I started out with - a donation from my grandparents, because I loved music so much they felt I should have it. Either that or they were just sick of me playing Charley Pride's 'Crystal Chandeliers' over and over whenever I went to their house. To be fair I was about six I think… but that really was the record I liked best in their collection!

Danny1969

Yep no tweeters just a 7" driver each end in the one we had. Ours had pride of place in the front room and was normally pressed into service on a Sunday

chillidoggy

Our HMV stereogram was a full 8 watts per Channel. Music power.