[blades]

I guess that settles it.

 

[sandab]

The main difference is in the frequency cutoff for the analog low-pass filter on the DAC's output.  The more gradual an analog filter can cut off, the less impact it will have on phase.  So higher output rate and a digital low-pass will sound better than relying on a very steep (brick) analog filter.  Of course, you can do this in the DAC, assuming it has a DSP.  This will add cost and complexity though since the DSP itself is a source of HF ground noise so will need its own separate power source.  Or you can do it in the transport or computer and feed a higher rate to the DAC over optical (which makes for nice isolation) or USB (which may potentially be noisier, but not as bad as a DSP).  Or put it in the source material itself.

 

[SilverEars]

Interesting, didn't expect this to come off in a sampling rate thread.    
 
I have read that Wolfson has low pass built in, possibly analog?  Actually, the WM8741 differentiates fomr WM8740 in that the filter can be switched via software.   ESS implementation I've seen is with low pass outside the DAC and people have said the LPF is not implemented in the chip like some.  
 
 

http://www.mezzohifi.com/private_folder/608e348712076d73f11a929240fa3d78.jpg
 

 

[castleofargh]

  The main difference is in the frequency cutoff for the analog low-pass filter on the DAC's output.  The more gradual an analog filter can cut off, the less impact it will have on phase.  So higher output rate and a digital low-pass will sound better than relying on a very steep (brick) analog filter. 

I think that is exactly why NOS DACs have(should have) disappeared. most DACs already use the best sample rate at the best moment, whatever the original sample rate of that track is(prononce dothraki). so I can hardly see that as a reason to use hirez.
 
people who truly believe 20 or 25khz are part of the music should go for hirez(or learn a little more about how human hearing works). if we admit hearing those, it does make sense to go hirez for several technical reasons if you're positive you headphone/speakers can actually deal with those frequencies.
now if you are a human being older than 15, chances are hirez is a waste of space and money for you.
 
all I know is that if 22 or 30khz were necessary to hear the "real" music, that would mean I'm ****ed, because at my age I can't hear above 17khz.
the fact that I still enjoy music very much is to me a clear proof that I don't mind missing 30khz sounds and in fact I would live very well with music cut off at 14khz. my headphones showed me that long ago. but everybody's free to believe what he wants and buy what he likes. the human race is not in danger from more hirez files, so who cares?

 

[Mambosenior]

In the hi-rez version of "Let It Be," the message "Yoko broke up the Beatles" can be heard continuously at around 23khz. Playing the album through my Audio-gd DAC, the message is heard in Chinese.

And I was ready to ditch hi-rez!

 

[esldude]

In the hi-rez version of "Let It Be," the message "Yoko broke up the Beatles" can be heard continuously at around 23khz. Playing the album through my Audio-gd DAC, the message is heard in Chinese.

And I was ready to ditch hi-rez!

I don't see why it is so hard to see how hirez has more info to work with.  All anyone has to do is play back hirez files at half the sample rate to hear a difference.  Depending on your hearing you gain an additional 12-20 khz of bandwidth that way. 

 

[nick_charles]

 
I don't see why it is so hard to see how hirez has more info to work with.  All anyone has to do is play back hirez files at half the sample rate to hear a difference.  Depending on your hearing you gain an additional 12-20 khz of bandwidth that way. 

 
The issue is not whether high res has more data but whether the difference is audible.
 
To date when the audible difference between high res and non high res (from the same source)  has been tested with material with ultra high frequencies that has been nobbled by downsampling or low pass filters the difference is not normally audible. This has been supported by several published papers going back to the late 70s including from professional broadcasting bodies and has not been contradicted by any of the industry proponents of high res with carefully controlled listening tests. Yes there are a bunch of anecdotes out there but at present there is only one paper which might possibly support audibility of high sampling rates (Pras and Gustavino) but it has somewhat questionable stats and method and a small sample, the largest scale test the Meyer and Moran study found nobody who could reliably detect the presence of a secondary A/D/A loop inserted after a high res player output.
 
Humble red book competently captures above 20K and even for those lucky few who can hear above 20K anything above 20K filtered out from musical content is not missed

 

[esldude]

   
The issue is not whether high res has more data but whether the difference is audible.
 
To date when the audible difference between high res and non high res (from the same source)  has been tested with material with ultra high frequencies that has been nobbled by downsampling or low pass filters the difference is not normally audible. This has been supported by several published papers going back to the late 70s including from professional broadcasting bodies and has not been contradicted by any of the industry proponents of high res with carefully controlled listening tests. Yes there are a bunch of anecdotes out there but at present there is only one paper which might possibly support audibility of high sampling rates (Pras and Gustavino) but it has somewhat questionable stats and method and a small sample, the largest scale test the Meyer and Moran study found nobody who could reliably detect the presence of a secondary A/D/A loop inserted after a high res player output.
 
Humble red book competently captures above 20K and even for those lucky few who can hear above 20K anything above 20K filtered out from musical content is not missed

I knew I should have used the facetious smiley face. 
 
You misunderstood me.  The last sentence should have been the tipoff.
 
By playing back at half the sample rate I didn't mean resample say 96khz to 48 khz.. I meant taking the 96 khz data and playing it back at a 48 khz rate.  Which cuts all frequencies by half.  So 40 khz content becomes 20 khz.  Then ultrasonics become truly audible.  Everything else about it is messed up.  But the crazy claims about ultrasonic content can at least be heard in a fashion.  Then compared to 48 khz played back at 24 khz you really would hear a "hirez" difference. 

 

[nick_charles]

 
I knew I should have used the facetious smiley face. 
 
You misunderstood me.  The last sentence should have been the tipoff.
 
By playing back at half the sample rate I didn't mean resample say 96khz to 48 khz.. I meant taking the 96 khz data and playing it back at a 48 khz rate.  Which cuts all frequencies by half.  So 40 khz content becomes 20 khz.  Then ultrasonics become truly audible.  Everything else about it is messed up.  But the crazy claims about ultrasonic content can at least be heard in a fashion.  Then compared to 48 khz played back at 24 khz you really would hear a "hirez" difference. 

 
Sorry, I'm obviously not paying enough attention !

 

[bigshot]

Comedy is serious business. It should be left to professionals... like me!

 

[ab initio]

 
I knew I should have used the facetious smiley face. 
 
You misunderstood me.  The last sentence should have been the tipoff.
 
By playing back at half the sample rate I didn't mean resample say 96khz to 48 khz.. I meant taking the 96 khz data and playing it back at a 48 khz rate.  Which cuts all frequencies by half.  So 40 khz content becomes 20 khz.  Then ultrasonics become truly audible.  Everything else about it is messed up.  But the crazy claims about ultrasonic content can at least be heard in a fashion.  Then compared to 48 khz played back at 24 khz you really would hear a "hirez" difference. 

 
I thought the original statement was so good that I couldn't think of a reply worthy enough to follow it!
 
Cheers

 

[BirdManOfCT]

 
And the high resolution one may actually contain LESS information than the CD resolution one, in that many, many high resolution downloads, especially those purchased from HDTracks, do not come with full information booklets containing information such as recording data (time and place of the recording, the equipment used to make the recording, producer, recording engineer, mastering engineer) and musicains, etc. - i.e. LESS information.
 
But hey it costs more and the high end audio press just love high resolution digital audio so high resolution just has to be BETTER

Some of the re-masterings are welcome. Provided it's not tweaking to make it into a different animal. Some of the songs before re-mastering have cymbals that sound like glass shattering, while the re-mastered version sounds much more like cymbals. If someone can't tell the difference, even with cheap gear, they should give up listening critically and ONLY enjoy the music. I like to do both.

 

[BirdManOfCT]

I don't have a pony in this race: frankly, I couldn't care less if 24/192 is proven to be better than lossless 16/44.1 (that's why, in my initial post, when I said my HDTracks purchases sound better than CD lossless counterparts, I added "for whatever reason", leaving open the possibility of better masters). Musical memory is a funny thing--for those of us who don't have perfect pitch, it's all too brief. By "musical memory," I don't, of course, just mean remembering melody or drum rhythm: I mean remembering pitch, dynamics, timbre, soundstage, etc. For most songs, unless they're remarkably minimalist, it's a mass of information all at once. Moving on--I think it's perfectly valid to administer these tests to Joe Schmo, so long as your goal is to determine whether such file differences matter to Joe Schmo. But if your goal is to guage whether or not 24/192 is better than 16/44.1, then you need a far more rigorous methodology than any I've seen: for me, you'd need, say, 50 professional musicians--preferably with perfect pitch. I'm familiar with all the arguments--some very well informed--that say anything over 16/44.1 is pointless. I understand the scientific claims--and yet, my ears tell me there's a difference. I've long believed that, for me, bit depth was key. So I decided to put that hypothesis to the test (of course, this is hardly conclusive--just anecdotal). For me, high res music invariably sounds more polished, more mellifluous (even 16/44.1, by comparison, sounds the tiniest bit grainy). It's subtle, but, for me, definitely there. For an informal test (I use a Mac, so can't use Foobar), I scaled an HDTracks 24/44.1 file back to 16/44.1 (Coldplay's "Ink"--my wife's unusual choice). My wife played the versions back to back, switching at my request, and I noted which track I thought was which. We were in separate rooms. I spent 30 mins before the test familiarizing myself with the song. She played the song in pairs, one version 16/44.1, the other 24/44.1, ten pairs total. I "guessed" the high res version 100% of the time. I'm not offering this as proof of any kind. Maybe my DAC manages 24 bit better than 16. Who knows. As I said, I'm not going to believe that 24/192 is definitively better than 16/44.1 until a test is devised with a methodology I can respect (I don't care what Joe Schmo hears). But until then, I'll trust my own ears.

Good stuff. I'm aware of placebo effect, too, but it's amazing how over time the there-is-no-difference starts giving ground when things improve. What was "impossible" 10 or 20 years ago is now "common knowledge" since the science caught up and "fixed" it.

 

[tarnumf]

Not sure if this was mentioned in this topic before.. But theoretically, couldn't frequencies in inaudible range still distort audible frequencies through interference effect, producing beats?

 

[Head Injury]

  Not sure if this was mentioned in this topic before.. But theoretically, couldn't frequencies in inaudible range still distort audible frequencies through interference effect, producing beats?

It's called intermodulation distortion (IMD) and yes it can. However, if it's present from how the recording was made (i.e. what we'd theoretically want to produce being all into hi-fi and crap) then the parts that distort the audible range will still be present with a 44.1 kHz sampling rate. Anything higher sampling rates and frequencies can add will be unwanted distortion not present on the recording, just flaws in the playback chain. This is why higher sampling rates can actually be harmful, as opposed to the standard.