Tuesday, September 3, 2013

Hirez - Treasure Island - Part II

Today we cover HiRes (that covers DSD as well) vs. Redbook.

In the first blog about Hirez I covered the base HiRez material.

My conclusion was.:

Hirez material quite often seems to get resampled and remastered from unknown raw masters (formats).
There is usually nothing like a real "master" for sale out there.

Only that fact alone will make many Hirez vs. Redbook comparisons rather useless.
I mean. If you compare a Redbook  CD and its remaster 5 years later, you'll
also hear a huge difference.

What would you expect from a "24bit" first of all and a higher resolution done with even better algorithms. Yep. Just another re-issue. They all sound different.

...which might not be a bad thing.

There've been tests run with professionals, who can easily tell SW resampling algorithms apart.
That means if you use a different resampler or resampling parameters, a different attenuation
and dithering algorithm, your results will always sound different , even with the same base material transfered to the same HiRez sampling frequency.

Numerous forums and numerous sites discuss if 44.1 vs. 88.2+ makes a difference.

There is no common sense. And usually discussions end up in flame wars.

Obviously there are audiophiles who can hear a difference, and there are others who don't.
There a group therapies, usually held at events like RMAF, with some kind of potentially scientific
approach (doubleblind and ABX) who "proved" that only 50% could tell a difference.
Thus there's none -- at least from that particular """scientific""" perspective.

And there are the physicians who try to sell the message "We can't hear above 20khz".
Hires is nonsense.

There are others who tell us that in theory it is impossible to process these high sample rates through any DA conversion process, since the noise floor will always be higher then those tiny bits.

Guess what. They are all right. In a way.

Depending on each and every specific source material, situation and setup, you'll experience a difference or not, even without being able to listen above 16khz at best.

Very often people ignore the HW side.  Which is, beside the re-mastering aspect, the key issue.
IMO the main argument supporting HiRez is the fact that HW implementations can be made simpler.
You need less intrusive filters, you can run NOS Dacs without any filters, you might pass by the internal resampling algorithms, which exist in many DACs.

With HiRes. You basically introduce some kind of "offline" filtering, prior to the realtime processing through your DAC.
If we talk asynchronous resampling (ASRC) inside a DAC, you'll see 44.1 converted to 192 or higher instead of 176,4. In many cases 44.1 will suffer more from an asynchronous conversion to 192 than 96khz base material to 192.

I'll stop here for the time being. I just wanted to outline that there is no black or white. As usual.
There's always an "It depends".  And. Yep. Sometimes I can hear differences and sometimes I can't.


1 comment:

  1. Hi Klaus, nice to see a sane person writing about audio and high-res.

    Regarding high samplerates:

    If there are 2 impulses/sounds:
    One is e.g. an echo of the other, or you might think of them as identical, but both are half a samplerate apart in time (low res ADC).

    What would be recorded?
    Now the same thing for high-res?