Serious Question: How can DAC's have a SOUND SIGNATURE if they measure as transparent? Are that many confused?

[valve_integrator]

My point in posting cymbal captures done at I think it was 176 khz wasn't to show there is nothing above 20 khz

Prof Schneider always held that bit depth is the important thing.
The red book idea was always flawed, because 44.1khz 16 bit (CD), could be proven to truncate the reverb.
Transient response and FR and even S:N/dynamic range were not the issue even at that resolution.
It was "came after" that made all the difference.

Having said that, it was usually B + K with their (old laboratory standards) 130V mic range that could exceed the dynamic range of all digital systems.

 

[Thorsten Loesch]

Prof Schneider always held that bit depth is the important thing.
The red book idea was always flawed, because 44.1khz 16 bit (CD), could be proven to truncate the reverb.
Transient response and FR and even S:N/dynamic range were not the issue even at that resolution.
It was "came after" that made all the difference.

Having said that, it was usually B + K with their (old laboratory standards) 130V mic range that could exceed the dynamic range of all digital systems.

16/44 is MOSTLY adequate, but it falls just a trifle short. Decca used 18/48 for recording, it is audibly better.

It would have been better to have used 18 Bit's and 64kHz with a transition band of 20kHz -32kHz instead of 20kHz - 22.05kHz. Also, 16kHz "flat" response and as much as 0.5dB rolloff at 20kHz. It would have have produced a format essentially transparent.

The bit rate of 2,307,970 bpS vs. CD's 1,413,632 bpS (including overhead) would have cut playback time from 72 minutes to 44 minutes per single CD, equivalent to two sides of LP.

A great benefit would have been that Artists did not pad out their album's to 70 minutes with fillers.

More on the early history of digital audio recordings.

https://www.aes.org/aeshc/pdf/fine_dawn-of-digital.pdf

Details of Decca's system are here:

https://theartofsound.net/forum/showthread.php?8386-The-Decca-Digital-Audio-Recorder

Thor

 

[antcollinet]

16/44 is MOSTLY adequate, but it falls just a trifle short. Decca used 18/48 for recording, it is audibly better.

It would have been better to have used 18 Bit's and 64kHz with a transition band of 20kHz -32kHz instead of 20kHz - 22.05kHz. Also, 16kHz "flat" response and as much as 0.5dB rolloff at 20kHz. It would have have produced a format essentially transparent.

The bit rate of 2,307,970 bpS vs. CD's 1,413,632 bpS (including overhead) would have cut playback time from 72 minutes to 44 minutes per single CD, equivalent to two sides of LP.

A great benefit would have been that Artists did not pad out their album's to 70 minutes with fillers.

More on the early history of digital audio recordings.

https://www.aes.org/aeshc/pdf/fine_dawn-of-digital.pdf

Details of Decca's system are here:

https://theartofsound.net/forum/showthread.php?8386-The-Decca-Digital-Audio-Recorder

Thor

I"d modify the statement slightly.

Your proposed standard would be "totally transparent". CD is already "essentially transparent" - yes I know we are talking opinion of what essentially means. In this case I'm meaning in virtuallly all listening conditions, and for virtually all people.

I suspect the benefit of the increased time available on a CD outweighed the imperceptible reduction in transparency in the minds of the developers. It would have for me too.

 

[valve_integrator]

Well we always held that 48V phantom wasn't adequate.
The recording of Percussion de Strasbourg won a prize, - thanks to 130V Bruel & Kjaer and more.

Prof Schneider also held that a lot of bad things happened when digital came along (quite apart from the truncated reverb). I won't go into the details, as it would make sound engineers uncomfortable.
He said his good old Studer Analogue mixing desk was vastly superior to the Yamaha digital that came after.
Having used another Swiss made gorgeous mobile desk (sonosax), in music festivals I have to agree.

CD ? Transparent?
Grief, - to make up for the obvious defects everyone and their dog started to add loads of "effects".
First we had Lexicon, then Avid, and now of course full on Pyramix.

(sadly a product we got on great with, was a thing called wavewarp which twinned with matlab but vanished.
I dunno why. The Gendarmerie forensics lab couldn't get even close to the results we got with that.)

 

[Thorsten Loesch]

I suspect the benefit of the increased time available on a CD outweighed the imperceptible reduction in transparency in the minds of the developers. It would have for me too.

“Those who forget their history are condemned to repeat it.”
— Carlos Santayana, The Life of Reason, Constable London 1905

“Computers in the future may weigh no more than 1.5 tons.”
— Popular Mechanics, forecasting the relentless march of science, 1949

“640K ought to be enough for anybody.”
— Bill Gates, 1981

"16 Bit sampled at 44.1kHz ought to be good enough for anybody."
— Herbert von Karajan and Akio Morita (Sony),1981

Seems 1981 was a good year for "good enough".

Funny, someone remind me why we no longer use Computers with 640k memory? Maybe an absence of ABX tests that show that 4MB memory is indistinguishable from 640k when running Lotus 1-2-3 on IBM DOS 1.0?

Thor

 

[valve_integrator]

"16 Bit sampled at 44.1kHz ought to be good enough for anybody."
— Herbert von Karajan

Karajan was so far "over the hill" by 1981, even he must have been hitting the dementia curve.
Some of his last Henry wood proms performance were awful in the 80s.

 

[pma]

Thor likes apple to oranges comparisons .

 

[restorer-john]

The red book idea was always flawed, because 44.1khz 16 bit (CD), could be proven to truncate the reverb.

Dither. Look it up.

CD ? Transparent?
Grief, - to make up for the obvious defects everyone and their dog started to add loads of "effects".

No obvious defects except in the minds of people with a) short memories, b) weren't there or c) had vested interests in the vinyl status quo.

 

[restorer-john]

Funny, someone remind me why we no longer use Computers with 640k memory? Maybe an absence of ABX tests that show that 4MB memory is indistinguishable from 640k when running Lotus 1-2-3 on IBM DOS 1.0?

Huh?

 

[valve_integrator]

Dither. Look it up.

No obvious defects except in the minds of people with a) short memories, b) weren't there or c) had vested interests in the vinyl status quo.

I don't need to look up dither.
Did you read what I wrote?

There is inadequate resolution on CD format to avoid truncating (live) reverb, hence why people use and abuse convolution and "plug ins".
It's a well known issue in pro sound circles.
There is also inadequate resolution on 48V phantom for large dynamic range, so again why compression is used and abused.
I have no vested interest in vinyl, not even 45RPM SAMs in UHJ surround made by people like Nimbus (have you heard of them?).
I have one of the very few (BBC) Matrix H decoders in the world ever made for that format.
So what?
The BBC set world wide standards for broadcast audio, and one of the first adopters of PCM FM stereo relays

Analogue is what it is.
Digital is what it is.
Both of them have flaws.

 

[Thorsten Loesch]

Having said that, it was usually B + K with their (old laboratory standards) 130V mic range that could exceed the dynamic range of all digital systems.

I have nothing against high voltage microphones, BUT most 1/2" Capsules I know (Yes, I make my own Mike's, but I buy capsules) are 10mV/PA or 10mV/96dB at 60V Bias. This gives 100mV 116dB and 1V 136dB. Using 120V Bias would double the sensitivity.

The self noise of a 1/2" capsule is around 18dB(A) and thus around 78dB below 10mV, let's round up to 80dB and state that the self noise of the capsule will be around 1uV RMS A-weighted, or around 7nV|/Hz.

Incidentally, this gives dynamic range for 136dB SPL (ridiculously high even for recording individual instruments at sensible distance) of 118dB A-weighted.

Using 48V Phantom Power with schoeps style electronics (updated with 1.5nV|/Hz P-mos followers and more modern J-Fet's) and set up for maximum dynamic range we have around 4.7mA current and 16V across our load resistors (and for our frontend). A 10 Ohm Fet into 6.8kOhm has minimal distortion at ~ 10% signal swing.Output would be 20mV/PA.

Running our J-FET at 8V across the FET as split load phase-splitter (maximum dynamic range) and buffer will start showing a bit of distortion, but we have a fair bit current and thus adding a P-MOS to our J-FET as compound feedback pair can help control distortion well and will also reduce the apparent input capacitance of the J-Fet, which if using a modern low noise FET is quite large.

We now have a microphone that will handle 136dB SPL with minimal distortion excluding the capsule itself and will have noise dominated by the Capsule as well as J-FET's with << 7nV|/Hz are not hard to find. And all that with 48V Phantom power.

If you are a Muntzer like me, you limit the Mic current to 2mA and use the resulting 41.2V at the mic for bias, losing 3dB sensitivity. With a Schoeps style electronic we now have 14mV/PA and with ~ 20V for follower and for the frontend each, even a sole FET (not compound feedback pair) will easily handle the resultant 700mV/136dB with minimal distortion and low capacitance << 5nV|/Hz @ 2mA J-FET's are not hard to find.

So I think 48V Phantom Power (P48) is fine if the Microphone is designed to maximally exploit the available voltage.

Edit: It does however leave scant room for poor design or unreasonable performance requirements.

Thor

 

[Mnyb]

I don't need to look up dither.
Did you read what I wrote?

There is inadequate resolution on CD format to avoid truncating (live) reverb, hence why people use and abuse convolution and "plug ins".
It's a well known issue in pro sound circles.
There is also inadequate resolution on 48V phantom for large dynamic range, so again why compression is used and abused.
I have no vested interest in vinyl, not even 45RPM SAMs in UHJ surround made by people like Nimbus (have you heard of them?).
I have one of the very few (BBC) Matrix H decoders in the world ever made for that format.
So what?

Arent you confusing recording and consumer formats ? everyone likes to use 24bit of some sample rate (48/96/192 ?) to recording stuff at concert halls and studios.
headroom is great for production everyone's wants that, you can record at safe levels and leave good margins.

But for final delivery for consumption at home i think CD is good enough , I rarely see records even close to limit of this system ?

 

[Thorsten Loesch]

Thor likes apple to oranges comparisons .

Actually, oranges to oranges. Which for eastern europeans are easy to compare, we could not get any most of the time, until 1989. Oranges were always great if you had any.

Thor

 

[IAtaman]

Audio world is fascinating. You never see people claiming they can see ultraviolet, or monitor designers marketing their products based on their on infrared performance - this kind of stuff only happens in audio world. I am not sure why but I love it. It is engineering with a touch of wine connoisseur BS - love it.

 

[SIY]

I don't need to look up dither.

But perhaps the professor you cite does, unless you're misrepresenting him. The claim is ridiculous.

 

[Killingbeans]

The red book idea was always flawed, because 44.1khz 16 bit (CD), could be proven to truncate the reverb.

Actual truncation? It doesn't just fade into the noise floor?

 

[pma]

Oranges were always great if you had any.

Not true for those that were imported from Cuba .

 

[valve_integrator]

Arent you confusing recording and consumer formats ? everyone likes to use 24bit of some sample rate (48/96/192 ?) to recording stuff at concert halls and studios.
headroom is great for production everyone's wants that, you can record at safe levels and leave good margins.

But for final delivery for consumption at home i think CD is good enough , I rarely see records even close to limit of this system ?

NO.
Fact is if you want to go to comsumer format, you already have to start at HR, so of course it's 24bit.
Not many have access to LIVE high resolution audio from source, but it's the real problem with doing everything from A-Z. (end to end).
Anyone with decent sound recording experience will hear the difference blind from 24bit to 16.
It's really obvious.

FInding a system which doesn't colour reproduction, seems to be working often in reverse, and that includes the DAC layer.

Headphones are as good as useless for getting any idea of the final result, (we were told not ot use them!) and studios are supposed to have dead dry acoustic.
Disappointment is the term when getting involved with the Hi-end audiophile mob a+ their shops and listening rooms. I haven't found a decent one yet!
Nothing ever worked out there, so I had to make my own.
That is all very well for plain vanilla stereo.

When it comes to surround it's a whole shed load of extra hassle and trouble (of course because so few have even approaching a decent surround system), and then what, most of the team say why bother?

We did a surround recording with the choir off stage in the foyer (that was a challenge), then guess what, all the audience in the front rows started talking and opening sweet papers as they sang...(nothing to see on stage eh!)
I can't remember which it was (surround DVD was done), maybe The Tsar's Bride or Iolanta.

I remember doing a Winterreise in the super acoustic of Metz Arsenal, but on that evening it was thick fog and Udo R sang poorly.
That with the rehearsals and set up was a fascinating lesson in setting up a surround recording system, and we did a Brahms Requiem in a similar way, using the (much underrated and beautiful) calrecs for the rear channels..
'nuff said, it's sliding off topic.
Just saying really, optimising any long chain is tough.
Some DAC have proven to be quite audible over the years, and now we are living in the all digi AES67/Ravenna/Dante world. hmmm.

 

[DSJR]

16/44 is MOSTLY adequate, but it falls just a trifle short. Decca used 18/48 for recording, it is audibly better.



Thor

I'm told by a retired Decca mastering engineer friend of mine that their system was developed for enablement of editing and so on without audible issues as the early editors weren't good. I believe Sony and so on caught up by the early 90's, but I'd have to check on that. NOTHING I gather to do with 'audiophile sound quality' as such, but to make their work in editing and so on easier and less audibly intrusive.. I think that was stated more or less in the article linked to (and arguably ignored by many who used to post on that site, what with their modified Technics turntables and valve amps into old Tannoys and so on).

You really think our ears are *that* good? No proof, but I bet what 'we human males' pick up as a cymbal crash or jangling keys is well short of 22kHz and arguably more like 13 - 15kHz for a male over thirty not listening via headphones! Having had three conventional UK based audiology tests done last year, they only test up to 8kHz or so..

 

[BDWoody]

Anyone with decent sound recording experience will hear the difference blind from 24bit to 16.
It's really obvious.

Uh huh.