[firedog]

What do you think of these posts?

See: http://www.head-fi.org/forums/f133/24bit-vs-16bit-myth-exploded-415361/
http://www.avguide.com/forums/red-book-riddle-solved

Does this make sense?

Danny

[ted_b]

I think, when you listen to hirez music, you listen to increased resolution AND bit depth.  I have no clue what one would do without the other.  To spend a lot of time debating 16 vs 24 seems to be debating one hand clapping.  I guess we could listen to 16 bit downsamples (like what comes out of Oppo's coax while playing DVD-A) and compare.  i like what I've heard, so I guess resolution is more important than bit depth.  But how many 16/96 recordings are out there??

[FullRangeMan]

...so I guess resolution is more important than bit depth.

so I guess resolution is more important than bit depth.  >> You are right, the important is the ''resolution''  or Sampling Rate per second(96Khz, 192khz etc), the bit rate is less important, it is just a way to packing the sampling rate data.
Obviously with DSD2.8Mhz and DSD5.6Mhz all PCM recordings formats are obsoleted in sound quality, aside this all PCM formarts generate a lot of data and use too much space in the disc than DSD.
Regards

[bdiament]

Hi Ted,

I think, when you listen to hirez music, you listen to increased resolution AND bit depth.  I have no clue what one would do without the other.  To spend a lot of time debating 16 vs 24 seems to be debating one hand clapping.  I guess we could listen to 16 bit downsamples (like what comes out of Oppo's coax while playing DVD-A) and compare.  i like what I've heard, so I guess resolution is more important than bit depth.  But how many 16/96 recordings are out there??

Having played with all sorts of combinations for a long time, I would submit that both sample rate and word length (what some call "bit depth") are both measures of resolution.  One is horizontal and the other vertical. 

By this I mean increasing sample rate (all other things being equal, though they seldom are) increases resolution in terms of frequency.  Increasing word length increases resolution in terms of dynamics.

Personally, if I had to choose only one (and I'm so glad I don't), I might be tempted to go with word length, i.e. 24/44.  Then again, I have recently heard one or two A-D converters that, unlike many others that claim 192k on their spec sheets, attain performance levels I would not have predicted from any recording medium in the foreseeable future (i.e. they no longer sound digital or even like "good" digital - they no longer sound analog either, as they now sound like the mic feed).

Still, if I had to, I just might go 24/44 over 16/192.  Those extra 8 low order bits bring back the sound of the space, the focus of the space, the low level details in the instruments and the complex harmonics of instruments that are not present with 16-bit, even the best 16-bit.

Just my perspective.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

[ted_b]

Barry,
Thanks.  As i stated I have no clue but your experiences and your explanation make all the sense in the world.  I guess tradeoffs are always the audiophile's obstacle, but at least we know which tradeoffs count more in this example.  All I know is that my 24/96 and 24/192 recordings are quite wonderful.  Bring on more please.      thx
Ted

[jsaliga]

All I know is that my 24/96 and 24/192 recordings are quite wonderful.

That's been my experience as well.  Though some people have suggested the reason most hirez content sounds better is due to better mastering rather than greater resolution.  I think it's probably a little bit of both.

--Jerome

[Quiet Earth]

I would submit that both sample rate and word length (what some call "bit depth") are both measures of resolution.

Totally agree!

I always record to 24 bits and 44khz when I make my own amateur recordings because I can hear a significant improvement over 16/44. I doubt that I actually use all 24 bits but the extra headroom seems to help on my hard disk recorder.

I also prefer the crude truncation of bits down to 16 when making a redbook CDR instead of trying to "dither" it down to 16 bits. But then again, I'm just an amateur.

[bdiament]

Hi Danny,

...Does this make sense?

Danny

Not in the least.  The author makes assumptions about what can and cannot be heard. 

He also completely neglects the fact that low level information is not encoded with all the available bits.  So, on a CD of a dynamically uncompressed classical recording for example, a part that is played at ppp might be 30, 40 or 50 dB down from full scale.  At -50 dBFS, we'd be capturing that information using only 10 bits.  Perhaps such is inaudible to the author but most folks I know would notice this right away.  ("Is that a Steinway?  or a Casio?"  ;-})

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

[bdiament]

Hi Jerome,

That's been my experience as well.  Though some people have suggested the reason most hirez content sounds better is due to better mastering rather than greater resolution.  I think it's probably a little bit of both.

--Jerome

The way to verify this is by comparing 16-bit and 24-bit versions of the same recording from the same mastering session.

One example can be found here:
http://www.soundkeeperrecordings.com/format.htm

I intend to put up more in the future.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

[bdiament]

Hi Quiet Earth,

I would submit that both sample rate and word length (what some call "bit depth") are both measures of resolution.

Totally agree!

I always record to 24 bits and 44khz when I make my own amateur recordings because I can hear a significant improvement over 16/44. I doubt that I actually use all 24 bits but the extra headroom seems to help on my hard disk recorder.

I also prefer the crude truncation of bits down to 16 when making a redbook CDR instead of trying to "dither" it down to 16 bits. But then again, I'm just an amateur.

This may be a function of the dither algorithm you are using.
In my experience, most (but not all) dither algorithms tend to cloud the soundstage and alter instrumental timbre.

Then again, simply truncating from 24-bits to 16-bits throws away an awful lot of low level detail and other information.  Listen to what happens to clean 24-bit reverb tails when you truncate to 16-bit.

The criterion I use to evaluate dither/noise shaping algorithms (as well as sample rate conversion algorithms) is to compare the results with the unprocessed original. 

To date, of the few dozen dither and few dozen src algorithms I've tried, I find MBIT+ from iZotope (as well as their 64-bit SRC) to provide results that sound more like the original, unprocessed files than I've gotten from any of the competition.

MBIT+ is wonderful at preserving much of the low level detail of the 24-bit original in the 16-bit result, without the loss of all the "air" that I find occurs with truncation to 16-bits.

Just my perspective.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com