24bit vs 16bit, the myth exploded!
Apr 5, 2015 at 8:17 PM Post #3,241 of 7,175
The threshold of perception for timing error (i.e.: group delay) is 1 to 3 ms in the core frequencies. All other things being good, that speaker you have the measurement on would still sound fast enough for human ears.
 
Apr 5, 2015 at 8:19 PM Post #3,242 of 7,175
  The threshold of perception for timing error (i.e.: group delay) is 1 to 3 ms in the core frequencies. All other things being good, that speaker you have the measurement on would still sound fast enough for human ears.

 
I was responding to analoguesurvivor's ridiculous statement that speakers/headphones were faster than the CD format.
 
Apr 5, 2015 at 8:22 PM Post #3,243 of 7,175
  I was responding to analoguesurvivor's ridiculous statement that speakers/headphones were faster than the CD format.


He says a lot of stuff like that. I used to read and reply to it, but stating self evident things over and over got tiresome. Now I read everyone else's comments and blow past his.
 
Apr 5, 2015 at 9:06 PM Post #3,244 of 7,175
   
Yes, and the same goes for audio illusions. Luckily we can take measures to eliminate human error.

Granted, then it's a semantic issue over the word "wrong".  I don't think it's the best word to describe sensory phenomena that arise completely as a consequence of how our brains are wired to lie to us under some circumstances, since "wrong" has a little bit of a sense of moral judgement to it.  In my opinion it is a special class of error since it does not involve making a mistake or overlooking something.  In other words, sensory illusions are not caused by dishonesty or miscalculation, and we are all subject to them.  It does perhaps involve letting go of a little bit of ego to admit that your ears, or eyes, etc. might be fooling you, but that applies to everyone.

Yes I'm getting into philosophical territory, and I'm flogging the horse more than a little bit, but only because I think this is where we might start to find common ground: that we are all subject to these tricks and sensory illusions, audio included.  Entrenched disagreements like this one usually trace down to conflicts at fundamental levels, so I'd like to see how far down that disagreement carries.

I'd really like to hear whether people on both sides of this bridge acknowledge that audio illusions exist, and that they personally are subject to them.  I know that I sure am.

I listen to redbook FLAC on my playback devices of choice, and believe (backed by all the science I have seen), in terms of the source data, I am getting the best possible audio that I am able to hear.  Does that confidence in the suitability of the source file influence how it sounds to me?  Probably.  Hi-res (etc) advocates clearly feel the same thing only when playing hi-res files (etc).
 
Apr 5, 2015 at 9:32 PM Post #3,245 of 7,175
  Yes I'm getting into philosophical territory, and I'm flogging the horse more than a little bit, but only because I think this is where we might start to find common ground: that we are all subject to these tricks and sensory illusions, audio included.  

 
I feel like I am in the Matrix and I keep seeing that same black cat walk by, over and over...
 
Apr 6, 2015 at 2:46 AM Post #3,246 of 7,175
  I listen to redbook FLAC on my playback devices of choice, and believe (backed by all the science I have seen), in terms of the source data, I am getting the best possible audio that I am able to hear.

 
And scientific tests will tell you that you probably would get the exact same sound quality from AAC 256. Feel free to follow science instead of your gut feelings.
 
Apr 6, 2015 at 3:42 AM Post #3,247 of 7,175
  CD-ROM error correction is different from Redbook - Redbook (i.e. CD audio) will actually attempt to handle read errors by interpolating data and presenting an approximation of the expected bits; CD-ROM (i.e. what computers use for data reads/writes also provides error detection, via Reed-Solomon, but any failure is simply missing bits - i.e. catastrophic - and the reading software is so notified via the I/O stack).

 
While CD-ROM data tracks do use more error correction than Red Book (each 1/75 s = 2352 bytes sector includes 276 bytes of error correction code, 4 for error detection, and only 2048 bytes of actual data without the header), the latter does in fact already have significant error protection. For each 24-byte frame of raw audio data, there is 8 bytes of Reed-Solomon error correction code (= additional 784 bytes per 2352-byte sector), and the actual physical encoding uses eight-to-fourteen modulation, which further improves error tolerance. Additionally, the data is interleaved in a way that makes a large error at a single physical location (for example, because of a scratch) distributed over multiple frames as small, easily corrected errors. Overall, with the error correction, EFM, synchronization, and subchannel data, the total data size as physical pits on the CD is more than three times larger than the raw Red Book audio.

 
In short, it should normally be possible to read a good quality disk that is not damaged without bit errors, and it is not true that every microscopic scratch or dust particle will turn into an error that has to be interpolated.
 
Apr 6, 2015 at 4:02 AM Post #3,248 of 7,175
  ABX is a methodology not a specific piece of software or hardware. To ABX DSD and PCM would require more work and hardware.

 
One possible solution is to convert the DSD to PCM, and then back to DSD, and compare that to the original DSD. Although this may in theory be worse than listening to the PCM version directly, if no difference is detected in ABX testing, then that suggests that the lower sample rate of the PCM does not audibly degrade the sound quality.
 
Apr 6, 2015 at 5:38 AM Post #3,249 of 7,175
   
Beauty of digital is that it's either all right or all wrong. Your CD mat might help with making it all right, but that's provided there's something wrong to begin with.
 
------------------------------------
 
Correct me if I'm wrong, but I thought the burden of proof was necessary here? If you challenge what is established, get proof.

For some reason, it won't let me to comment on the first half. Since you are relatively new here - I AM the binaural guy here. It is the (commercial) presure of necessity to play music over loudspeakers that is keeping me from getting exclusively binaural - but it is my personal favourite. And I doubt I will ever use multimiking - two mike recordings and no mixing desk straight into the recorder. Post production limited to pasting together the best (bits of) takes. But if the producer wants his "mic minefield" or no recording at all ... - then it will be somebody else doing the recording. I meant these multimiked recordings are beyond salvation if realistic sound is to be obtained. 
One can mix and master ad nuseaum - the result will sound artificial no matter what.
 
I, too, thought that the beauty of the digital it is either right or all wrong. Then, I started to worry what will happen to my CD-R masters in the long run - it never happened to you that CD or CD-R went bad and could no longer be read ? First, I tried gold CD-Rs - claimed to have longevity of 100 years +, meant for archiving.
The first time I was recording direct to this gold CD-R in a church, I cried LOUD some most unholly words - so MUCH better did this CD-R sound. As these CD-Rs were expensive and on their way out, alI  I could get was the remaining stock of 20-30 CD-Rs - clearly not sufficient for any serious work. I went into search mode for archival CD-Rs, particularly those that do not cost an arm and a leg. It was a three months search on the internet - and I did stumble upon then emerging CD-Rs that perform even better than gold ones and cost the same as regular ones - AND  guaranteed to be good for 100 years to boot. Adding CD mat was the final icing on the cake.
 
I asked my friends to participate in CD-R shootout listening - among original pressed CD and copies made to "regular" CD-R, gold CD-R and this "super regular priced" CD-R. And it was established the SQ is in the same, ascending order. NO ifs and buts. CD mat was used at every stage, from ripping to playback.
 
So much for the digital being either right or completely wrong. It obviously does have shades of grey. At least when disc is involved.
 
Recently, this subject has been brought up on this pages. As my stock of these "super regular priced" CD-Rs has dwindled to a single 100pcs box, I went into search mode - again. And was finally able to procure some more - by now, they have become so rare ( everything that is too good at too low price gets discontinued...) that search even on ebay worldwide produces - zero results. Now it is down to somebody who abandoned CD-Rs altogether is selling off his/hers leftover - totally sporadic - so it is understandable why I do not wish to share the brand and model of these discs. They will be used for direct  to CD-R recording - exclusively. For copies, lesser CD-Rs will have to be used from now on.
 
You are correct in establishing the proof is my burden.
 
Apr 6, 2015 at 5:51 AM Post #3,250 of 7,175
  For some reason, it won't let me to comment on the first half. Since you are relatively new here - I AM the binaural guy here. It is the (commercial) presure of necessity to play music over loudspeakers that is keeping me from getting exclusively binaural - but it is my personal favourite. And I doubt I will ever use multimiking - two mike recordings and no mixing desk straight into the recorder. Post production limited to pasting together the best (bits of) takes. But if the producer wants his "mic minefield" or no recording at all ... - then it will be somebody else doing the recording. I meant these multimiked recordings are beyond salvation if realistic sound is to be obtained. 
One can mix and master ad nuseaum - the result will sound artificial no matter what.
 
I, too, thought that the beauty of the digital it is either right or all wrong. Then, I started to worry what will happen to my CD-R masters in the long run - it never happened to you that CD or CD-R went bad and could no longer be read ? First, I tried gold CD-Rs - claimed to have longevity of 100 years +, meant for archiving.
The first time I was recording direct to this gold CD-R in a church, I cried LOUD some most unholly words - so MUCH better did this CD-R sound. As these CD-Rs were expensive and on their way out, alI  I could get was the remaining stock of 20-30 CD-Rs - clearly not sufficient for any serious work. I went into search mode for archival CD-Rs, particularly those that do not cost an arm and a leg. It was a three months search on the internet - and I did stumble upon then emerging CD-Rs that perform even better than gold ones and cost the same as regular ones - AND  guaranteed to be good for 100 years to boot. Adding CD mat was the final icing on the cake.
 
I asked my friends to participate in CD-R shootout listening - among original pressed CD and copies made to "regular" CD-R, gold CD-R and this "super regular priced" CD-R. And it was established the SQ is in the same, ascending order. NO ifs and buts. CD mat was used at every stage, from ripping to playback.
 
So much for the digital being either right or completely wrong. It obviously does have shades of grey. At least when disc is involved.
 
Recently, this subject has been brought up on this pages. As my stock of these "super regular priced" CD-Rs has dwindled to a single 100pcs box, I went into search mode - again. And was finally able to procure some more - by now, they have become so rare ( everything that is too good at too low price gets discontinued...) that search even on ebay worldwide produces - zero results. Now it is down to somebody who abandoned CD-Rs altogether is selling off his/hers leftover - totally sporadic - so it is understandable why I do not wish to share the brand and model of these discs. They will be used for direct  to CD-R recording - exclusively. For copies, lesser CD-Rs will have to be used from now on.
 
You are correct in establishing the proof is my burden.

Great! I love binaural too (except for synthetic or pop): very mind-screwy in the beginning to listen to, but becomes beautiful once you get used to it.
 
That being said, I hate to say it, but I have to agree to disagree with your statement about digital having shades of grey. If it does, then it's called analog. The very definition of digital is that it's either a 1 or 0. Of course, real life is never that discrete (pun intended), so we just have a bunch of transistors and gates that interpolate everything as either a 1 or 0. Hence my extreme dubiousness when you said the gold CD's were superior in quality (and I'm not talking about after a few dozen years).

If you insist though, what's your theory on why gold CD's are superior? I highly doubt you just gobbled it up: you must've done some research or at least some cursory reading to help explain your experience.
 
Apr 6, 2015 at 6:33 AM Post #3,251 of 7,175
  In short, it should normally be possible to read a good quality disk that is not damaged without bit errors, and it is not true that every microscopic scratch or dust particle will turn into an error that has to be interpolated.

 
Yes, it is now, today, the case. I was just making the point that CD-ROM will only go so far to recover from read errors and CD Audio will "wing it" with interpolation when it can't read all the pits. I'm sure you don't think I believe that every scratch results in a CD-ROM read failure, due in large part to improvements in the pick-up lens and the laser, as well as tracking and the software.
 
Apr 6, 2015 at 6:38 AM Post #3,252 of 7,175
   
Nice try but no cigar (and obviously I was talking about speakers not cartridges...) . Frequency response is NOT the only factor in determining how quickly a system can start and stop, and a fast tweeter is not enough if the other drivers are slower. Even a high-end speaker's (JMLab Utopia, $30K) step response looks like this:
 

 
A Stax headphone is faster but not faster than the CD format itself. Top-of-the-line SR-009:
 

There IS one use of PCM/DSP I will be (grundgingly...) perhaps forced to adopt - filters for loudspeakers. It has been demonstrated a really decently made PCM/DSP filter (running at least at 192/24, preferrably twice that speed ... > VERY powerful computer just for the crossover ! ) can produce an almost perfect SQUARE WAVE, not
only a pulse - measured by the microphone at the listening position. There are articles floating in the net.
 
Stax DO have a problem - their stators are too thin and too prone to resonance - the bigger, the worse - and their fastest phones are, quite understandably, the baby Stax IEM - SR001MK2 or their new succesors. Simply because they are the smallest surface and least prone to get mechanically excited. I have Lamda Pro and SR001MK2 - and although the IEM can never image nearly as well as something that presents our pinna with for all practical purposes undisturbed free air response, the ringing in Lambda(s) is clearly audible and for some critical listening I will always use SR001. They are also less fatigueing over long periods of time - like watching long movies - for the same reason. I will have to make adapter to be able to use them with desktop amp instead of the limited portable amp. I can not find the measurements for the 001 at the moment, but I do remember it is superior in pulse to the 009.
 
It is ridiculous but true - Xiaomi Piston 2.0 ($25 delivered worldwide) HAS better pulse response than SR-009 : http://www.innerfidelity.com/images/XiaomiPiston2.pdf
Needless to say - I own this one too. And will order the Piston 3, according to the first impressions MUCH improved IEM, ASAP - and it costs < 30$ delivered.
 
Regarding speakers - you obviously have chosen (deliberately?) something poor in pulse response. All it takes is approx 2K$ - some Magnaplanar in that price range. Higher up, Quad ESLs are CERTAIN to run rings around Utopia regarding pulse response - as do old Acoustats etc. And to make it even more riduculous, go out and measure this little guy (that will set you back approx whole $25 or so per driver - plus enclosure, of course) http://www.visaton.com/en/industrie/breitband/frs8_8.html As far as pulse response is concerned, it will drive Utopia into the ground so hard that nothing will be left to see above the ground level...
 
It is THE speaker for vocals - as much as I like ESLs, this 80 mm paper cone taught me a lesson that is impossible to forget. It is not large, it does not play loud, it requires VERY powerful amp relative to its diminutive size ( I drive it with 75W/ch ), it obviously can not deal with (loud) bass, it will not court my resident bats - BUT in the all important midrange and pulse, it has VERY few peers. 
 
Needless to say, it can be augumented with a decent bass driver and some supertweeter into a full range system intended for small(er) rooms. And crossover arranged using DSP in order to pass perfect pulse and square wave at the listening position. True - I did not make it reality - YET.
 
Apr 6, 2015 at 6:54 AM Post #3,253 of 7,175
  Yes, it is now, today, the case. I was just making the point that CD-ROM will only go so far to recover from read errors and CD Audio will "wing it" with interpolation when it can't read all the pits.

 
The point was that Red Book does use Reed-Solomon error correction, it just does not have the extra layer of error correction that is present on CD-ROM data. This did not seem to be clear from your post, so I tried to explain it better for those who could have thought that Red Book error tolerance is limited entirely to good optics and interpolation. The same applies to the comment regarding any small scratches (not) resulting in data errors.

 
Apr 6, 2015 at 7:12 AM Post #3,254 of 7,175
   
One possible solution is to convert the DSD to PCM, and then back to DSD, and compare that to the original DSD. Although this may in theory be worse than listening to the PCM version directly, if no difference is detected in ABX testing, then that suggests that the lower sample rate of the PCM does not audibly degrade the sound quality.

 
It is quite wrong to compare only the sampling rates when comparing PCM and DSD. The DSD's rate is faster but it takes a lot more samples for it to express a transient, for example, than PCM, which can express it in a single sample. DSD offers no advantage but does have the disadvantage that it is not compatible with DSP. I wish DSD would die already...
 
As for your test, yes if people can't ABX a double conversion then it's safe to say the single conversion is also inaudible.
 
Apr 6, 2015 at 7:26 AM Post #3,255 of 7,175
  There IS one use of PCM/DSP I will be (grundgingly...) perhaps forced to adopt - filters for loudspeakers. It has been demonstrated a really decently made PCM/DSP filter (running at least at 192/24, preferrably twice that speed ... > VERY powerful computer just for the crossover ! ) can produce an almost perfect SQUARE WAVE, not
only a pulse - measured by the microphone at the listening position. There are articles floating in the net.

 
Finally, something we can agree on. Digital crossovers are seriously good, eliminating all sorts of problems in traditional passive crossover, simple amp configurations. Only thing holding me back is the considerable extra cost.
 
Originally Posted by analogsurviver /img/forum/go_quote.gif
 
Simply because they are the smallest surface and least prone to get mechanically excited.

 
This is also a false conclusion. If it were true, then every IEM would demonstrate exceptional impulse response. Size does not necessarily make a driver faster or slower - there are other factors involved.
 
Originally Posted by analogsurviver /img/forum/go_quote.gif
 
Regarding speakers - you obviously have chosen (deliberately?) something poor in pulse response.

 
No, I literally posted the first high end speaker I found.
 

Users who are viewing this thread

Back
Top