[badmonkey]

I posted this in the computers-as-source forum then realized that this was probably the better home for it. There must be something...?

http://www.head-fi.org/forums/showthread.php?t=254747

 

[ADD]

I wouldn't be trying to downsample a high res source yourself - even converters (hardware or software) that claim to be mathematically perfect can often really hurt the sound.

If you go to the Linn Records site, they have their music available in multiple resolutions. For your particular purpose, just look for recordings which are available in both 16-44 and 24-96 - there are quite a few of them.

I recently did this same experiment myself. For a few dollars I bought some music in both resolutions. The 24-96 did sound better, but I did not feel it justified the extra cost.

In the end, I'd rather the balance engineer had moved the microphone 6 inches further back and 2 inches to the left than for me to spend all that extra money on the high res stuff.

http://www.linnrecords.com/

 

[ADD]

I should add that even if you were to find the 24-96 samples "better", it's not necessarily a fair comparison on the part of the humble 16-44. The problem is that an SACD master for instance, might be pure DSD or 24-96, so then a 16-44 version of that has to be created for the ordinary CD or CD layer. The same process applies to downloads as well - for example Linn's 16-44 downloads are the resampled versions of the higher res versions and are used to create the redbook CDs. As soon as you downsample anything you will lose sound quality, no matter how sophisticated the hardware or software is.

The only really fair way to compare the two resolutions would have been if two different storage mediums were used at the recording session and those two different mediums used to create the high res and standard res versions. For example, a 24-96 recorder could have been used concurrently with a DAT machine set to 16-44. Then the final comparison would be much more meaningful because there would be no resampling required of the low res version.

I know that Simon Eadon of Abbas, for example, has employed both the 24-96 and DAT machines at the same time, but I'm not sure if he sets the DAT machines to 16-44 or 16-48.

 

[sejarzo]

Some of the Linn material is actually 24/88.2, which makes much more sense, IMHO. It's trivial to get to 24/44.1 by taking every other sample from a 24/88.2 recording. Dithering to reduce the 24 bits to 16 shouldn't produce huge degradation--in fact, hardly anybody tracks at 16 bits anymore, as far as I know. Virtually everything is tracked/mixed/mastered at 24 bits and dithered down now, right?

I'd be curious to know why the perception is that the benefit of going to 96 kHz from 88.2 kHz exceeds the penalty of downsampling 96 kHz to 44.1 kHz, as it does seem that more hi-rez recordings are available that way....when the majority of the consumers want CD/44.1 kHz format rather than a hi-rez version.

Anthony Michaelson of Musical Fidelity put out an SACD of himself playing the Mozart Clarinet Concerto that has four versions of the same session.....SACD from DSD, followed by SACD from analog concurrently tracked with the DSD recorder, with the CD layer containing the downconverted DSD version plus one analog direct to redbook version.

You can often find one of them for sale on eBay or buy it new here:

http://ssl.blueearth.net/primedia/pr...0&cat=9&page=1