24-192 a scam?

Is there a reason you think the uncorrelated signals will be inaccurate in excess of 4us in a waveform constructed with known values at 22us intervals.

If a transient arrives shortly after the ADC s/h gate closes, then it's influence will be delayed by up to 22us, so, yes.

I don't care much about the science, but I do care about my ears.

The LSO Live series allows subscribers to download both 16 and 24 bit versions and comparing the two is very simple. The 24bit versions are clearly better than the 16bit versions, and I blind tested this on a friend using Dvorak 9 and Bruckner 9. He unhesitatingly picked out the 24bit versions as sounding better.

He's an old git like me, so it's not like our ears are golden.

P.S. Never had the chance to compare sample rates. Less convinced by that.

I'm surprised these new fangled dacs don't have some way of predicting the start of a transient.

Well I thought they did, by holding many samples in memory, looking ahead, then reconstructing the original waveform by clever interpolation algorithms.

I'm surprised these new fangled dacs don't have some way of predicting the start of a transient.

Aren't a certain manufacturer working on a telepathic one with an infinite number of taps?

Jan,

I believe the algorithms are limited by the Fourier Uncertainty limit, whereas the brain appears to circumvent this in some way, possibly by using some non-concious post processing of the signals from the ears.  As you know there's some astonishing parallel processing going on in the background all the time, I think this may be another example.

http://phys.org/news/2013-02-h...ainty-principle.html

(In reply to fatcat) Not so much prediction as reconstruction by interpolation. Polynomial interpolation in the example below.

With interpolation, things could be idealised to some extent, though these interpolations could never be guaranteed with absolute certainty [but only statistical probability] for all that - IF WE ACCEPTED a certain degree of latency in the system

I don't find latency an issue so long as pitch is steady. If pitch is steady then time - by definition - so would timing [rhythm and pulse] also be stable, and relayed with as much precision as interpolation would allow.

But the issue does come back to what the microphone can catch ...

At its best replay is a splendid and convincing illusion that presents something quite accurately close to what the players did in the recorded performance. 

But every link in the chain has a negative effect on the result.

In reality - in my opinion - we have long since passed the point where such deficiencies are significant, and therefore it is unlikely that most people will gain much from improving the replay chain beyond Redbook standards. 

I do not doubt that some people can accurately define the differences in blind listening tests, but nineteen out of twenty cannot. At that point it is most unlikely that commercial recording ventures will bother.

In a way we have reached the point with replay where the improvement of it is more or less without a good point. Without a good point to it, then it is not going to become, commercially, an imperative!

ATB from George

When I've seen video's of how the Vinyl master disc's are made, then cut, and then the Vinyl puck is heated and 'squashed' onto the master disc, I'm amazed that the records even play on a turn table... given when all that was created (and has hardly changed) it's pretty incredible really.

Short of going to a Dr of Otology to have my hearing range's validated, I downloaded the mutli encoded versions of some music over the holidays :

Studio 24bit flac/alac 192kHz

Studio 24bit flac/alac 96kHz

CD Quality 16bit flac/alac 44.1kHz

Compressed 320k mp3 44.1kHz

{from Linn and B&W - Naim weren't giving away freebies :-( }

I plan to listen through different setups, headphone/PC + hifi system to see if I can hear the difference, and then whether those difference justify any cost expenditure on equipment/downloads in the future.

At the end of the day, it's our ears and the systems we use that will determine whether we can hear the difference, regardless of whether someone on a forum or hifi magazine can.

Having wasted many hours encoding music since the mid 90's ( when it took my PC overnight to process 3 WAV files into 128k MP3 via a DOS program ) I plan on a 'good enough' approach, especially when I have a basement with 6,000+ CDs, 300 Vinyl and 200 Tapes.

Agreed.

Is this:

a) factual and fully contextualised

b) regurgitated from one academic's theory to create something from which we can only be saved by more hi-fi?

Wow, is that......a graph?? Maths and stuff? Must be true.

It can be difficult to grasp that with 44,000 samples in a second, what happens in the tiny gaps between them is not something you're going to be able to separate.

As an experiment, draw 44,000 dots on each of two A4 sheets of paper. On one, conform to the dots in joining them. On the other, feel free to freestyle a bit between the dots - whilst still ensuring you join them properly. Now, examine both sheets of paper and describe the differences.

That's right, there are none. The dots are too close together to provide for differences.

You can talk about the master and its resolution all you like dear George but some still want to believe you can add everything which is supposedly lost back in afterwards. That you, the humble audiphile, may be delivered from such injurious sounds by a time-machine which will take you back into the concert hall where the recording was first played.

One fears they may not only be in defiance of logic but also in denial of their task in-hand. That is, going round all the record companies to persuade them to re-master all records made in their preferred format

Correct me if I'm wrong, but I think there may be a difference between the way the eye-brain combination works and the way the ear-brain combination works.

Which seems to me to be carte blanche for the hifi snake oil salesmen!!

The argument that highres only makes sense if the recording was done in the same or better quality makes sense. You can take a tape master and make it into a 192 24 bit file, but in reality it has nothing to do with highres technically speaking. it is interesting or worrying to see how much old music from the 70's or earlier that is rereleased by hdtracks and proclaimed to be highres.

So a valid comparison must be with music that was mixed and mastered in 192 24 bit from the start.

What about all the music from the 70's where the masters are on analog tape.  That's a lot of what I see showing up in 24/192 following remastering.

Indeed, and a lot of it (bought through Qobuz) sounds very good too.

It really does sound very good to me.  Recent 24/192 remasters purchased from HDTracks, and now from the Pono store, sound fantastic.  I do credit the remastering engineer more than the sheer number of bits, that's for sure. 

Some of the material in the video above comes from the Xiph site. The video in the link below (unable to embed it here) lasts 24 minutes and is always a useful refresher on things digital. Timing is covered near the end of the presentation.

http://xiph.org/video/vid2.shtml

Earlier in the thread I made the points raised in this linked video.

Jan,

I hope that you will not be offended by me making that point.

The reason the  normal [nowadays] 24 bit production prior to mastering to 16 bit occurs is the leeway it gives for level setting and noise-ratio that might be brought in by sub-optimal level setting with 16 bit production-work. However, as our ear will soon let us know, even in the days of 16 bit recording, many recordings were wonderful, such as the 1979 New Year's Day Concert from Vienna on Decca. That was there first digital issue, and today remains not only a superb recording, but one of the very best New Year's Day Concerts to be commercially issued. 

ATB from George

That's a very good point Mark, if it's impossible to distinguished anything between samples at a rate of 44,000 per sec, it's pointless producing music at higher sampling rates. There must be something other than higher sampling rates causing a perceived improvement.