advertisement


Analogue bit rate equivalent?

Whaleblue

Whaleblue

Southbound
Rather than hi-jack the the tuner thread that raised this question I have given it a thread of its own..

And I'm wondering - what is the answer? How would you calculate such a thing?

Some thoughts. Analogue has to be presented on a physical media. Let's choose vinyl as the obvious one to explore. Let's also assume a fully analogue process from recording to the laying down of the bumps and wiggles that form the groove.

What accuracy does the equipment used in the manufacture and playback allow? And the materials? To what accuracy can a 1Khz simple sound wave be formed in the final vinyl pressing? And then what about the >20KHz frequencies?
 
I read somewhere that tape in a perfect world would give 6 bits but realistically is about 4-5 bits and vinyl is somewhere around 11/12 bits.
 
The only reason we can calculate and compare definition between digital systems is the fact the sound has been broken up (say) 44,100 times per second, and each time the signal has been represented by a word of up to 16 binary digits. A system of different sampling rates or bit depths allows direct comparison with these figures.

This hasn't happened with a truly analogue chain from master tape to vinyl, so essentially the question has no meaning.
 
Bits are equivalent to SNR to put it simplisticaly 1 bit = 6db - so if a vinyl system can achieve a 66 db SNR = 11 bits
 
Rodney gold said:
Bits are equivalent to SNR to put it simplisticaly 1 bit = 6db - so if a vinyl system can achieve a 66 db SNR = 11 bits
Click to expand...

Furthermore, bandwidth is related to sample rate, and vinyl can achieve 20kHz at best due to cutter head limitations, so one can say that the LP is 40kHz sampling at 11bit.

Note:- I'm aware that LPs can be cut at half-speed, as was indeed done in the 70s for cutting CD4 quadraphonic LPs, but all that does is to move the problem of bandwidth limitation to the bass end, as half-speed cut LPs then roll-off early in the bass.

S
 
Nyquist defines the sample rate based on the analogue bandwidth and the bit depth is defined by SNR as stated above - simples :)

So for FM radio at say 15kHz audio bandwidth and 60dB SNR, the sample rate would be 30kHz (2x analogue bandwidth) and 10bit depth (6dB per bit), so 300kbps per channel or 600kps in total for stereo with no compression.
But with digital you could then apply lossless compression to get this down by a factor of around 3 to end up at around 200kbps.
 
But the real question here is can digital accurately record/replay an analogue signal.

Vinyl has lots of SNR and bandwidth issues but is the stylus ability to reproduce the overall signal better than that of digital?
 
Stu_cj said:
Nyquist defines the sample rate based on the analogue bandwidth and the bit depth is defined by SNR as stated above - simples :)

So for FM radio at say 15kHz audio bandwidth and 60dB SNR, the sample rate would be 30kHz (2x analogue bandwidth) and 10bit depth (6dB per bit), so 300kbps per channel or 600kps in total for stereo with no compression.
But with digital you could then apply lossless compression to get this down by a factor of around 3 to end up at around 200kbps.
Click to expand...

The better tuners quote higher SNR figures than that though. The Denon TU-800 specs for example quote mono 96 dB and stereo 88 dB. Now I know there might not be many stations using the full dynamic range but that suggests that in theory FM should be more like 14 or 15 bits does it not?

Regards,
Nick
 
The analogue v digital argument and that word "accuracy" neatly differentiates those who
want to measure their hi-fi, and those like me who prefer to trust my ears.
To my ears analogue nearly always sounds better than digital (does anybody enjoy music on DAB radio?) even if some meter says it's not as accurate.
 
I asked a very similar question a couple of years ago and I think the consensus was that vinyl's specs under ideal circumstances are the bit- and sampling-frequency equivalent of 12 bit and 40 kHz, give or take, but vinyl sounds better.

Joe
 
I think the idea of high bit-rate lossy compression is interesting too. We always think of lossy compression as poor. However if for example we take a 192KHz 24bit recording and reduce the data we could do it in two ways. We could downsample and dither to 44.1KHz 16bit for a bit-rate of 1411Kbs. We could also use lossy compression such as AAC to reduce the bit-rate to 1411Kbs the same as CD, but wouldn't this actually sound better than CD since it is selective about what it throws away unlike downsampling and dithering?

Perhaps it makes no difference since CD is all we need for perfect audio, but it does seem possible to have lossy compressed high definition audio that is better than CD.
 
I remember the White Stripes coming to London so they could use an "Analogue" recording studio. 'nuff said.
 
Joe P said:
I asked a very similar question a couple of years ago and I think the consensus was that vinyl's specs under ideal circumstances are the bit- and sampling-frequency equivalent of 12 bit and 40 kHz, give or take, but vinyl sounds better.

Joe
Click to expand...

I doubt that very much

.
.
.

Consensus on PFM? Surely not!
 
Tenson said:
I think the idea of high bit-rate lossy compression is interesting too. We always think of lossy compression as poor. However if for example we take a 192KHz 24bit recording and reduce the data we could do it in two ways. We could downsample and dither to 44.1KHz 16bit for a bit-rate of 1411Kbs. We could also use lossy compression such as AAC to reduce the bit-rate to 1411Kbs the same as CD, but wouldn't this actually sound better than CD since it is selective about what it throws away unlike downsampling and dithering?

Perhaps it makes no difference since CD is all we need for perfect audio, but it does seem possible to have lossy compressed high definition audio that is better than CD.
Click to expand...
I don't think any lossy compression is likely to be better than CD, even if lossed from a high bitrate source - it's still choosing for me what part of the original I am going to be allowed to hear.

Today I might agree that one of two close frequencies can indeed be considered "masked" by the other which happens to be somewhat louder, and hence removed by a lossy compression algorithm before I get the chance to hear it, but tomorrow I might disagree that it was right to remove the quieter frequency before my ears and brain were allowed the opportunity to sift through the soundscape for whatever they might discover for themselves (and me) to enjoy (or not, as the case may be).

Downsampling (to CD rate) and dithering (at the 16th bit level) aren't doing quite the same thing as that as they really shouldn't result in any significant losses within the audible bandwidth.
 
You must log in or register to reply here.


advertisement


Back
Top