Why CD quality (as through Tidal) is all you need and higher res a waste of money

Posted by: gramophone on 01 September 2018

I used to wonder why whereas Tidal is fully integrated by Naim, Qobuz isn't. I was thinking of subscribing to Qobuz's sublime service (attracted by how you could buy highres albums for the price of the equivalent mp3's.) I sent a question to their customer services about integration with Naim - no response. This is the first issue with them - the often stated poor quality of their customer service.

But more important is the issue of the validity of highres itself. Surely you'd only want Qobuz for it's library of highres files.  In this video

https://www.youtube.com/watch?v=IiZqYnd5g8M

a guy gives a fantastic, clear and fully informed analysis of the maths and acoustic realities of the issues (like the uncontested science regarding the hearing parameters of the human ear). He's a sound engineer. He explains how, when Philips set the 16bit 44.1 KHz standard for CD just before the '80's, they set the parameters of the standard up to the best optimum and not down to any market based compromise. He does this really elegantly drawing and annotating sine wave and audio signal graphs with his ink pen. Camera just on the page showing numbers, equations and theories like that of Nyquist (all of it comprehensible to the layman). No larger field shots like you often get on youtube of would be hifi gurus who prattle on coffee in hand trying to develop their brand.

Naim are right to specify their streamers so that they can play all types of highres files - if the customer has such files they want to be able to play them.

 

Posted on: 03 September 2018 by Simon-in-Suffolk
Huge posted:
joerand posted:

I wouldn't lose sight of the fact that redbook CD is an accepted, functional replay medium and not a calculus quiz.

Indeed Red Book is an extremely effective replay standard; capable of very good, highly enjoyable results.

Technically speaking, it doesn't quite manage to cover all the auditory capabilities of human perception - but it does come fairly close to doing so.

Remember Redbook or CD-DA is a storage and data indexing, and error recovery format for PCM audio on optical discs. The sample rate of 44.1kHz and the bandwidth of 150kbytes per second is defined which was agreed upon  for fairly arbitrary reasons around marketability and production ... but there is no definition on how that data rate is to be replayed and converted into audio... tat is left to implementors.

Posted on: 03 September 2018 by joerand
Simon-in-Suffolk posted:  The sample rate of 44.1kHz and the bandwidth of 150kbytes per second is defined which was agreed upon  for fairly arbitrary reasons around marketability and production ...

I've read various layman explanations for the decision to go with the Redbook CD format and trust it is a rational SQ standard, even for audiophiles. I wonder how or where marketability affected that decision. Space on a disc? Size of a disc? The real SQ metric for me has nothing to do with Hz's or bits. It's all about the recording and mastering quality - done before the music ever finds its final replay outlet to the masses.

Posted on: 04 September 2018 by pete T15

An interesting read regarding sample rates , the limitations of the original CD format and their latest solutions can be found on the Benchmark website . It is a problem they describe as "Intersample Overs" , I don't think I can share the link here but its an easily found article thats quite in depth as well as being understandable . 

Posted on: 04 September 2018 by jlarsson

Originally PCM was encoded on normal VCR:s (Mainly sony beta), x samples per video line (dont remember x right now). So we got 44.1kHz.

16-bits is good as a distribution fomat. Your noise floor at home (if you live in a very quiet house)  we can set at 24dB. In a flat maybe 30dB. Add the 96dB 16 bits gives and you are talking about maximum sound levels of 120-126dB. 

The discussions about timing errors have been around since the beginning of the 80’s. Dont forget the filters and how everything translates to analog when discussing this. We dont listen to the bitstream.

”intersample overs” is not a problem with 44/16. it was discovered over 15 years ago that most oversampling DAC chipsets used interpolation that created peaks over the maximum peak level (0dB). This then led to clipping in the analog part of the chipset which was not designed handle levels above 0dB. It started to show withe ”loudness war” and happen even at 192/32bit. Oversampling DAC also multiplies jitter-problems.

Remember that vinyl has a dynamic range of what?  70-75dB tops? And music still sound pretty fine. And then we can start talking about timing errors of an elliptical stylus vs. tracking error. Vinyl still sound pretty impressive to me.

A 24-bit recording with natural dynamics wouldnt be listenable at home. You could listen at natural levels ... but your system probably cant play that loud. Or you could turn the volume down and loose all those important details into the natural noise floor. So the real magic (music aside) comes from experienced mixing/mastering engineers. Not mathematica.