Topping E30 MKII - Wonder DAC?

sorry for the late reply I was away for Xmas 

The number of bits a digital system has will determine its dynamic range - i.e the difference between the quietest sound possible to the loudest sound possible.

CD has 16 bits so 96dB of dynamic range. This is massive overkill for pretty much every domestic playback situation and there are very few recordings that exceed this range anyway.

We know vinyl has a best case dynamic range of around 70dB - so converting backwards we can get to the equivalent in bits which as stated previously is about 12 bits on a notionally perfect vinyl playback system with a notionally perfect pressing.

In practice this is adequate for good quality replay as I'm sure we have all found over the years.

Vinyl does not have a sampling rate as we are all aware but this is not an advantage. Sampling rate determines the highest frequency response that can be encoded - half the sampling frequency - so sampling at  44.1Khz gives frequency out to 22.05 Khz - beyond audibility. Technically vinyl does not have a maximum FR but in practice it is flat to maybe 15Khz on a good day, after that the signal starts to disappear into the noise as it attenuates.

So why do some prefer 'the sound' of vinyl?

My own take is this is based on some of its failings rather than any inherent strengths (as it really has none at all). 

Bass is mono on most cuts below about 100Hz - may give impression of stronger, more vibrant low frequencies

There's lots of noise - some correlated to the signal (which is bad) but some uncorrelated noise too, and this may give an impression of  enhanced 'air and space'

Phono stages vary wildly in their RIAA curve recreation, some are way out but this may enhance some recordings same as making a smiley face on our graphic equalizer did for AC/DC when we were fifteen.

And the same for cartridges.

put simply the bit depth dictates maximum amplitude and the sampling frequency dictates maximum frequency.

Being an electrical signal it only has three properties - frequency, amplitude and phase.

This shouldn't be confused with digital video where you have an increasing number of pixels in a given area as image resolution increases. The two systems have no connection.

The lower the resolution of the audio signal the lower the maximum dynamic range will be, but within those parameters there is nothing missing from the signal. I.e more bits does not allow you to see 'deeper into the grain'.

It is counter-intuitive which is why it doesn't seem logical.

If you consider that the whole signal is just frequency and amplitude it might make more sense. I can keep reducing the upper limit of FR down quite a way before it starts to sound dull - if you consider FM radio has a cut off at 15KHz


Likewise I can reduce the dynamic range down a lot before you notice, there's usually compression applied to recordings to do that deliberately since you don't actually want a lot of dynamic range for domestic replay and even less in a car. You have to turn it up to hear the quiet bit then the loud part blasts you out - and maybe clips the amp.

There are no gaps between the samples and there is no 'grain' to get deeper into. Just an upper limit to FR and a maximum peak level. This is why you will not be able to distinguish between 16/44 and 24/192 in a controlled test. (Yes I know - I thought I could too until I tried it). 

24 bit makes sense for recording though as you can bounce about a lot without raising the noise floor to audible levels.

I get that distinguishing between 16bits and 24 bits would be very difficult. That is the usual debate.

I was trying to work out when the number of bits is too low to do decent job in a very good audio system. Essentially, (as I understand it) the number of bits gives the resolution, which the amplitude can be 'broken down' in to. So 4 bits only gives a poor resolution (16 levels). Although, I bet there are clever ways to to improve the resolution with lower number of bits.

For arguments sake, 1bit it obviously insufficient to add enough resolution. I would assume that 4bits is inadequate too. 8 bits getting better... 16 and above very good.

I did do digital electronics in my degree, but it is hazy now! I'm being lazy really as know a few boffins I can ask too - they're probably too clever and would give too much detail though!

So, correct me if I'm wrong...

Nyquist is the minimum rate at which a signal can be sampled without introducing errors, which is twice the highest frequency present in the signal, which is typically 20hz-20khz, hence 44.1khz

However, doesn't that meant that mathematically you are taking far less samples from the HF end of that spectrum than you are MID and LF end?




If 44100 samples are taken per second/cycle....there is a difference in what is being captured between the frequencies?

Why do so many speaker manufacturers pomote their super tweeters and the fact their speakers go to 30,40, 50 and 60kHz? Is there a benefit? I know Martin will say it's all rubbish. Perhaps it's all a marketing ploy to sell more expensive kit. I've never heard an 40khz FR vs a max 20khz FR.

If there is something in ultrasonic hifi then the sample rate would be justified at 96kHz and above.

*I am NOT an advocate of speakers that go to ultrasonic frequencies. I will be if someone can reliably demonstrate the benefits.

No there really is a benefit!

Ha Ha just kidding - yes it's nonsense, as Jerry says super tweeters have output in the audible band, usually around 12khz up, so you will hear a difference with them deployed. Unless you have very bad hearing.

When SACD and DVD-A came out there was a rush of speakers to the market with response out to 40Khz (e,g Tannoy Sensys www.soundandvision.com/floorloudspeakers/204tannoy)  In the hope that people would replace perfectly good speakers with new ones that were 'SACD ready.'  It mostly never happened and the trend died out.

You will get some who say that frequencies from acoustic instruments that are above our ability to hear will modulate frequencies in the audible band - and this is true!

But the net effect of this is already captured by the microphone when you record the instrument. You don't need to replicate it again on playback - and indeed there is no benefit in doing so.

Half past 10 dear!