Recording Equipment

and worst, digital played at loud volume causes ear's pain- even on big digital systems- whereas you can listen to good analog -a simple Rega Planar 3 with a good cartridge is enough- very loud during hours without being tired.

What better proof do you want of analog's superiority over numeric?
And magnetic particles are not a limiting factor at all, IMO!

goose wrote: Certif1ed wrote: Analogue also both loses and adds "information" (noise) during the copy process - something that digital ABSOLUTELY does not do, unless the software used for copying has a CODEC routine of any kind. Quantisation noise at high levels and quantisation distortion level at low levels, while not actually recorded as such, are still an inherent part of digital signals that can't be removed.

Quantisation errors are what Oliver is talking about, if I'm not mistaken - the digital clock becomes "out of synch" with the audio frequency, and samples the analogue sound at the "wrong" times, producing what sounds like random noise, as the system truncates the values it cannot work out. This is because digital can only sample at set rates per second - and the amplitude of any given waveform may increase or decrease as well as simply be out of synch for that time period.

Increased bit depths help significantly, but most people use software techniques such as dithering to get rid of the random noise - but you're still left with the missing sound that analogue would not have missed, resulting in a slightly grainy sound, occasionally with tiny "artifacts" - like zooming in on a digital picture, and noticing that the sampling results in odd "splodges" in some areas where boundaries cannot be precisely identified.

I think you're mixing up some different things here ... quantisation occurs when an analog signal (at a single point in time) has to be translated to a numerical value. The only fix to this problem is increasing the resolution, and that's why SACD uses 24bit (16.7 million steps) instead of 16bit (65k steps).

The other big problem with digitisation is aliasing - which happens when a signal is sampled at less than half the sample rate of the original (the Nyquist Frequency). This results in a frequency that becomes part of the recorded sound. The effect of this can be heard in samplers - or by speeding up the playback of a digital music file.

Of course ... that is precisely why CDs use a sampling frequency of 44.1khz (DAT uses 48khz): It is intended to reproduce audio up to 22khz. But the problem is that this is the lowest possible way to reproduce the sound. So SACD doubles that to 96khz, and modern sound cards (Creative X-Fi and professional hardware) also allows for 192khz.

Digital music is attenuated using anti-aliasing to filter out the aliasing, which has the same effect as over-compression - you end up killing some of the "live" feel of the sound.

The main problem is with reconciling the nature of analogue wave forms (music) with the nature of digital; At the top end of the dynamic spectrum, digital samples less frequently - and this is typically (if you examine any sine wave) where information is densest.

VS

For example, a 15Khz tone recorded at CD quailty has only 3 samples per second - and so a saw wave cannot be distinguished from a square wave - two options common on synths and fuzz boxes. Dense harmonic clusters simply get lost, unless they've been digitised first.

I'm not sure if I understand what you mean by that. CD quality audio always has 44.1 thousand samples per second ...

BTW: saw/square waves are not single frequencies ... a sine wave is a single wave, and the more complex waveforms are always the result of many simultaneous (sine) waves.

goose wrote: Please bear in mind it is impossible to discern which format is "better" by listening tests, only which one the majority of people prefer listening to. Just because it's more pleasing to the ear doesn't mean it's a more accurate recording.

The above is one good way of telling the two apart - aliasing and attenuation come across as "helium sounds", "digital noise" and compressed, lifeless music, and quantisation produces artifacts and "graininess".

As I suggested to oliver, a good way to get rid of that is to cut off high frequencies alltogether.

You'd need to do some recordings to train your ear - using a 14 ips Reel to reel and comparing it with a recording made via a reasonable soundcard. Choose common but complex sounds, like bells, whistles, and guitars with fuzz boxes that have square and saw wave options

There's a cool web page that contains some demo samples of digital shortcomings ... can't remember the URL right now.

Obviously, most people wouldn't be able to tell the difference - but for rabid audiophiles - especially those with access to top-end equipment - it's not too hard. The difficulty comes in explaining the qualitative differences, and why one is better than the other.

If you really want to hear the difference, you need to train your ears to the specific sounds, or you can just happily live in ignorance and enjoy the music

I choose the latter. Maybe it would sound better to me with a 10k system ... but I'm totally sure that it would not improve the music itself.

I prefer analogue. It makes me feel warm and fuzzy, and the artwork is BIIIIG. 

Of course the analog systems look way cooler.