R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: 1 ... 8 9 [10]   Go Down

Author Topic: Bandwidth of the Ear (was '192kHz fs for audio')  (Read 18425 times)

Johnny B

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1134
Re: Bandwidth of the Ear (was '192kHz fs for audio')
« Reply #135 on: July 13, 2004, 09:28:37 PM »

Oh, I dunno. Is there not room for a new format that dispenses with all the legacy problems? Why not come up with a detailed proposal that solves the problems? If it really sounded that great, not only would the recording engineers, ear people, musicians and others love it, but it would give the record companies yet one more opportunity to sell their back catalogs to the consumers. Oh, and then there might be all that re-mastering work too. Sounds like an all-around winner.

But I would hope you could come up with a specific proposal.


 
Logged
"As far as the laws of mathematics refer to reality,
they are not certain; as far as they are certain,
they do not refer to reality."
---Albert Einstein---

I'm also uncertain about everything.

saxist

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 71
Re: Bandwidth of the Ear (was '192kHz fs for audio')
« Reply #136 on: July 13, 2004, 10:21:55 PM »

Johnny B wrote on Tue, 13 July 2004 20:28

Oh, I dunno. Is there not room for a new format that dispenses with all the legacy problems?


Absolutely!  You and a number of others on the site are clearly interested in it, so it should sell.  If it really does get rid of ALL the problems with current gear, and introduce no new problems of its own, that would be fantastic.  

Quote:

Why not come up with a detailed proposal that solves the problems? If it really sounded that great, not only would the recording engineers, ear people, musicians and others love it, but it would give the record companies yet one more opportunity to sell their back catalogs to the consumers. Oh, and then there might be all that re-mastering work too. Sounds like an all-around winner.

Absolutely!  Though its worth questioning who, exactly, those consumers are who will buy their catalog of recordings all over again.  The vast majority of the consumers I know are more than happy with the sound quality of CDs; they're thrilled with low-bit-rate mp3 files.  DVD-A and SACD are both struggling in the marketplace, when compared with the success of the CD introduction, and both represent a significant improvement in sound quality over standard CDs.  Why should we think that the general populace is going to be willing to invest in a new format, even if it does sound great?

Quote:


But I would hope you could come up with a specific proposal.


Here's where I have the same thought I had when you suggested to Bob Cain that he contact people at UCSC to set up and perform the experiment he proposed: quite seriously, why don't you do it yourself?  It's clearly something you'd very much like to have happen; go for it!

Eric H.
Logged
Eric Honour
Composer/Audio Engineer/Professor

Paul Frindle

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 380
Re: Bandwidth of the Ear (was '192kHz fs for audio')
« Reply #137 on: July 14, 2004, 05:54:31 AM »

Zoesch wrote on Wed, 14 July 2004 01:42

Paul Frindle wrote on Wed, 14 July 2004 09:14

Yes I have had many ideas on the subject over the years - most notably starting when I first encountered the potential effects of Mic cables being used to transfer digital audio - and timing!


I'd like to hear your findings, since most of my professional EE (Not AE) life has been spent in telco design (And analog design) I'd be more than happy to share and discuss this as this has been one of my recurring headaches for the past 12 years.

Quote:

Remember - timing sync and sample clocks are inherantly analogue signals!


Indeed, and I've always been baffled why we choose RZ encoding for those signals knowing how critical they are.

Quote:

However such ideas have never met with much sympathy as you may surmise - in a world where designers of some of the systems we use have actually even been offended by the idea of a CD drive not playing data out in exact sync with the sample clock! Or even that lossless data coding inherantly changes the sound!


The joys of isochronous systems and clock derivation, CD players are notorious for how badly in sync they remain, and you have not just the issue of clock integrity to the DAC but also mechanical issues from variably spin rates that affects the integrity of the signal being read from the lasers, issues which can be resolved but are costly to implement.

Quote:

It really is a perceptual thing in this industry IMVHO.


It's also a thing around expectations, for example, in my time with Newbridge and other equipment manufacturers showed that people are only expecting to pay a premium if the premium is tangible. Deploying a Stratum-1 clock is only viable when the sum of your revenues makes that clock worthwhile.

If we talk about a portable CD player with a simple oscillator cristal (If any, I've seen worse things like buffers and NOT gates used as self-generating clocks) the price to stabilize that clock to say 1-2 ppm accuracy is higher than the price of the system, and who would buy a CD player then?

Quote:

All we actually need to do is deliver the data securely and verifiably from A to B in a timely and correct manner and send the rate at which it is to be played out to something that generates the required timing at high accuracy and integrity Smile



And we have plenty of experience doing that, experience that shows that in order to achieve data integrity you need to make your encoding jitter-proof, that discards RZ and NRZ encodings and deriving clock from the encoded signal, Manchester coding for example is less succeptible to jitter and clock instability but it's also more complex to implement.

If you change from RZ-encoded PCM to Manchester-encoded PCM you'll notice an improvement due simply to clock regeneration accuracy, however that means changing each and every DAC and each and every CD manufactured to accomodate a new encoding mechanism.

Legacies, in any industry is what stops the industry from improving.



Dear Zoesch.

You make some salient and good points that illustrate the issues I would address - all this stuff I have witnessed and struggled with for years.
No time to discuss now - but actually you haven't got the drift of what I am driving at. I am not talking about simply changing the coding method to reduce jitter or asking for equipment to grow expensive electronics or people to run out and buy expensive clock generators. Soon as you pass that stuff down an analogue cable you've potentially had it anyway.
What I am on about is addressing the whole issue of transfer of data and clock in one swoop where none of this would be an issue (either it works or it doesn't - no in between) - an upheaval of the whole way we think about digital audio interconnection protocols. Probably not something I should labour on these forums if I want to avoid flame wars galore Sad
Logged

Zoesch

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 269
Re: Bandwidth of the Ear (was '192kHz fs for audio')
« Reply #138 on: July 14, 2004, 06:00:50 AM »

Paul Frindle wrote on Wed, 14 July 2004 19:54

Dear Zoesch.

You make some salient and good points that illustrate the issues I would addredd - all this stuff I have witnessed and struggled with for years.
No time to discuss now - but actually you haven't got the drift of what I am driving at. I am not talking about simply changing the coding method to reduce jitter or asking for equipment to grow expensive electronics or people to run out and buy expensive clock generators. Soon as you pass that stuff down an analogue cable you've potentially had it anyway.
What I am on about is addressing the whole issue of transfer of data and clock in one swoop where none of this would be an issue (either it works or it doesn't - no in between) - an upheaval of the whole way we think about digital audio interconnection protocols. Probably not something I should labour on these forums if I want to avoid flame wars galore Sad



Well color me VERY interested, inband clock regeneration is one of the things I deal with regularly and have my own issues around that (Specially around transmission since as you correctly point out, that's when problems arise)

If you want to take this offline it's fine by me, pm me if that's the case.
Logged
It has always been Ringo's fault

bobkatz

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2926
Re: Bandwidth of the Ear (was '192kHz fs for audio')
« Reply #139 on: July 14, 2004, 11:49:35 PM »

I dunno about a brain trust, Ted, maybe it's a lot of brains on overload!

ted nightshade wrote on Tue, 13 July 2004 16:31



This transfer issue plagues my atavistic distrust of digital- Will what's going out from device A really arrive in good order at device B? Will the digital signal suffer from being transferred? I live in fear... the basic promise of digital is that the signal can and will make it through basic unity operations fully intact.




Your paranoia is based on good logic, but fortunately, the SPDIF and AES/EBU interconnects work, most of the time. There is no fancy error checking and this is a pity, but most times it's either working or it's not, there's no in between. So you can usually tell right away that there's a problem.

When there is a problem, one symptom encountered is a complete loss of lock, which reminds me of the slogan "death is nature's way of telling you to slow down."

For me a complete loss of lock is far better than the more dangerous and subtle problem,  manifested by tiny clicks in the sound, which probably represent phase locked loops on the threshold of unlock. Happens around here whenever I repatch the input to one of my boxes without power cycling the box afterward. It's predictable, the mechanism for its cause is known, so it's just an annoyance, not a work-stopper.

There are diagnostic tools for AES/EBU, and like any well-kept professional studio in the analog days, having the right tools around, like an eye-pattern checker in the tool locker would not be a bad thing. You know, though, I don't own one; I could build one, but fortunately, the bottom line is the number of times we meet up with one of the above problems in a working studio is pretty infrequent. Usually substitution of a cable, or just plugging and unplugging or power-cycling clear up 99% of the problems we encounter. Are interconnect problems more frequent in the digital age than the old hum and noise problems we had with analog?  I think not, so actually I think digital has been an improvement in that respect. Plug and play, plug and play.

So, take a deep breath, make clean, short connections with good short cables, and relax, the risks are low and the world isn't coming to an end (yet).

Eventually AES/EBU will be replaced by wider bandwidth systems such as Firewire, yet another connector designed to pull out at a drop of a hat. Count your blessings.

BK
Logged
There are two kinds of fools,
One says-this is old and therefore good.
The other says-this is new and therefore better."

No trees were killed in the sending of this message. However a large number of
electrons were terribly inconvenienced.

Level

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1811
Re: Bandwidth of the Ear (was '192kHz fs for audio')
« Reply #140 on: July 15, 2004, 12:15:30 AM »

Those tiny "ticks" and "pops" that happen prior to getting a project can be an utter nightmare to fix in mastering. I actually traced them back to a USB port adapter the recording engineer had in use. The archive was littered with them...during the recording phase. It did not get caught in the mixdown either. How it got unnoticed until mastering blows my mind. Hundreds if not thousands of pencil tool edits later...finally...finally, I had useable product. It really should have been recut.

This was a luxury I did not have.

Nightmare of epic proportions.

Bob, I guess you experienced a simalar situation as well....
Logged
http://balancedmastering.com

"Listen and Learn"
---Since 1975---
Pages: 1 ... 8 9 [10]   Go Up
 

Site Hosted By Ashdown Technologies, Inc.

Page created in 0.079 seconds with 19 queries.