AVS/AIX High-Resolution Audio Test: The Results So Far - Page 9 - AVS Forum
Forum Jump: 
 147Likes
Reply
 
Thread Tools
post #241 of 457 Old 08-07-2014, 11:26 PM
AVS Special Member
 
m. zillch's Avatar
 
Join Date: Oct 2006
Posts: 4,352
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 335 Post(s)
Liked: 284
Quote:
Originally Posted by tubetwister View Post
Krab , can someone easily discern 0.2dB ? I was under the impression that the threshold was somewhat higher or no ?
I can't resist but to jump in. It depends on if the level change is:

- up or down [up is easier to hear, certainly for me at least]
- music, the kind of music, a sine wave test tone, random noise like pink noise, or the best signal is a narrow band of noise centered where the ear is both most sensitive to weak sounds and small changes in sound, around 3.5kHz
- the dBSPL level you started the test at
- is it a steady state tone which changes or is it comparing two bursts with a time space between them?
- what's the frequency, Kenneth?

Roughly speaking:
3dB = dead obvious
1dB = not a big difference but noticeable
.5 dB = only under good conditions like a lightning fast switch box that is held by the test subject so they know when to expect it.
.3 dB = some groups under ideal conditions have been able to hear it
.2 dB = very sketchy, no consistent group can hear it and usually close enough to deem it "level matched" for casual testing, but if your gear can adjust it in tenths of a dB anyways, why not go for...
.1 dB = what you really should shoot for to be published but studies with much worse have been deemed "level matched"


Test your self with this 440 Hz tone [not a frequency where humans are very sensitive to small changes, by the way, so don't be surprised if .5dB is super hard or even impossible]:
http://www.audiocheck.net/blindtests_level.php?lvl=1

In A/V reproduction accuracy, there is no concept of "accounting for taste". We don't "pick" the level of bass any more than we get to pick the ending of a play. High fidelity is an unbiased, neutral, exact copy (or "reproduction") of the original source's tonal balance, timing, dynamics, etc..


Last edited by m. zillch; 08-07-2014 at 11:32 PM.
m. zillch is offline  
Sponsored Links
Advertisement
 
post #242 of 457 Old 08-08-2014, 12:13 AM
AVS Special Member
 
tubetwister's Avatar
 
Join Date: Feb 2013
Location: Sacramento delta N. Cal. US Don't trust any air I can't see ☺
Posts: 3,207
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 634 Post(s)
Liked: 382
Quote:
Originally Posted by m. zillch View Post
I can't resist but to jump in. It depends on if the level change is:

- up or down [up is easier to hear, certainly for me at least]
- music, the kind of music, a sine wave test tone, random noise like pink noise, or the best signal is a narrow band of noise centered where the ear is both most sensitive to weak sounds and small changes in sound, around 3.5kHz
- the dBSPL level you started the test at
- is it a steady state tone which changes or is it comparing two bursts with a time space between them?
- what's the frequency, Kenneth?

Roughly speaking:
3dB = dead obvious
1dB = not a big difference but noticeable
.5 dB = only under good conditions like a lightning fast switch box that is held by the test subject so they know when to expect it.
.3 dB = some groups under ideal conditions have been able to hear it
.2 dB = very sketchy, no consistent group can hear it and usually close enough to deem it "level matched" for casual testing, but if your gear can adjust it in tenths of a dB anyways, why not go for...
.1 dB = what you really should shoot for to be published but studies with much worse have been deemed "level matched"


Test your self with this 440 Hz tone [not a frequency where humans are very sensitive to small changes, by the way, so don't be surprised if .5dB is super hard or even impossible]:
http://www.audiocheck.net/blindtests_level.php?lvl=1
Zillch

Thanks for info just as I suspected all along . IMO what matters (or should matter ) in the market place with hires is not what a trained listener who might as well be watching paint dry can maybe,or maybe not discriminate under very controlled conditions using controlled content , but how hires can provide a significant improvement to playback to joe six pack and enthusiasts alike entirely on it's own merits without extraordinary effort or special training which has not been proven in over a decade . I think at this point we are watching paint dry and trying to see the primer underneath if you will.

regards

Hires Music formats ..............."Why does it sound like a CD ?" ............. can we make it louder "?
"The wireless music box has no commercial value. Who would pay for a message sent to nobody in particular?"
- David Sarnoff's associates at RCA the 1920's -

Last edited by tubetwister; 08-08-2014 at 12:47 AM.
tubetwister is offline  
post #243 of 457 Old 08-08-2014, 01:25 AM
Advanced Member
 
jj_0001's Avatar
 
Join Date: Sep 2006
Location: In the rain
Posts: 761
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 59 Post(s)
Liked: 40
0.2 dB should be detectable, but not as a level difference (it comes across as a difference in quality, or other sensation than level for nearly everyone, well, everyone I know of, at least), rather as just a difference, if the listening room is quiet, the equipment is good, the listener has reasonably normal hearing, and the listener has rather a lot of practice.

Up or Down does not matter if you have proper test switching, because you get to go both ways.

If you have any time delay (say more than 1 second) between the two presented signals, that will erase the difference for all practical purposes.

Now, to this test, I seem to recall seeing somebody say that they had a set of replacement signals that were level equalized, and that the first ones weren't. That needs to be paid attention to.

What I haven't seen is any discussion on the exact response of the SRC that was used, which is, to put it mildly, of paramount importance. The fact that the 0.2dB came about due to SRC in the first place seems to suggest that there may be some passband slope or ripple, and frankly, that needs to be tested before any kind of listening test can be even judged to have meaning.

It would be good to put an 0.2dB level difference in a test as one of the test conditions, and make sure it actually shows up, too, as well as a duplicated file running as both A and B, (from two different files), just to ensure both positive and negative controls.

It is very easy to get both false positives and false negatives in this kind of test, very, very easy, to say the least. Time alignment can also bite you. If the time alignment is a few milliseconds off, the switching sensation will give it completely away, for instance.
FreeFire likes this.

James D. (jj) Johnston
jj_0001 is offline  
post #244 of 457 Old 08-08-2014, 02:58 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,387
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 762 Post(s)
Liked: 1175
Quote:
Originally Posted by jj_0001 View Post
0.2 dB should be detectable, but not as a level difference (it comes across as a difference in quality, or other sensation than level for nearly everyone, well, everyone I know of, at least), rather as just a difference, if the listening room is quiet, the equipment is good, the listener has reasonably normal hearing, and the listener has rather a lot of practice.

Up or Down does not matter if you have proper test switching, because you get to go both ways.

If you have any time delay (say more than 1 second) between the two presented signals, that will erase the difference for all practical purposes.

Now, to this test, I seem to recall seeing somebody say that they had a set of replacement signals that were level equalized, and that the first ones weren't. That needs to be paid attention to.
Here is an amusing example of how not to level match properly, coming from people I personally expected more from:

http://www.analogplanet.com/content/...les-identified

"On January 2nd analogplanet.com posted five 96/24 bit files, each containing the same minute’s worth of John Williams’ “Liberty Fanfare” performed by the National Symphonic Winds conducted by Lowell Graham excerpted from the album Winds of War and Peace originally issued in 1988 on Wilson Audio Specialties Records (W-8823) and used with permission.
Recorded by Wilson Audio Specialties’ Dave Wilson along with Bruce Leek and Joseph Magee using a pair of Sennheiser microphones and a John Curl Ultramaster™ 30IPS ½ inch tape recorder, the sound is spectacular, particularly the giant bass drum struck by what Mr. Graham said recently was the “M.O.D.” or Mallet of Death.”

"The excerpts were digitized using the $3950 Ayre QA-9 USB A/D converter. Four of the five files were produced using a $2995 Pro-Ject XTension10 turntable fitted with a $4249 Transfiguration Phoenix MC phono cartridge, each using one of four moderately priced phono preamplifiers: the Rega Aria ($1500), the Lehmann Black Cube SE II ($1099), the Musical Fidelity M1ViNL ($1199) and the P.S. Audio NuWave™ Phono Converter ($1895), which also includes a double DSD resolution A/D converter (the M1 ViNL was reviewed in Stereophile, not analogplanet, hence no hyperlink).

"The fifth file contained the same musical excerpt recorded using the $150,000 Continuum Caliburn turntable, $6500 Kuzma 4 Point turntable, $6995 Lyra Etna cartridge and Ypsilon VPS-100 MM phono preamp"

The following statistical analysis shows that the levels of the files fall over a range of more than 1 dB. The test just coincidentally was done in such a way that careful listeners would find an audible difference even if the items being compared were the same identical component. ;-)

Attached Thumbnails
Click image for larger version

Name:	preamp test 1-4.png
Views:	135
Size:	38.6 KB
ID:	202585  

Last edited by arnyk; 08-08-2014 at 03:36 AM.
arnyk is offline  
post #245 of 457 Old 08-08-2014, 03:11 AM
AVS Special Member
 
NorthSky's Avatar
 
Join Date: Jan 2013
Location: Canada - West Island: Vancouver, South Direction: Go East
Posts: 4,730
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 2044 Post(s)
Liked: 598
Is that good, Arny?
NorthSky is offline  
post #246 of 457 Old 08-08-2014, 03:17 AM
AVS Special Member
 
NorthSky's Avatar
 
Join Date: Jan 2013
Location: Canada - West Island: Vancouver, South Direction: Go East
Posts: 4,730
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 2044 Post(s)
Liked: 598
@ home I have a volume dB meter (analog) that records 0.001dB level variations. I'm quite used to it, and late @ night I bet I can differentiate level differences as little as 0.1dB. ...Perhaps even less if I concentrate hard enough, like 0.05dB. ...Maybe.
NorthSky is offline  
post #247 of 457 Old 08-08-2014, 03:34 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,387
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 762 Post(s)
Liked: 1175
Quote:
Originally Posted by NorthSky View Post
Is that good, Arny?
Mismatching the levels in a comparison of audio components is not the least bit good, and in the quantities measured, it suggests carelessness, incompetence or malevolence. Tough words, but true.

BTW full disclosure: The people involved in doing this test are not exactly my BFFs. We have been acquainted for a long time.

Someone observed that the final listener ranking of the equipment put the most expensive component first, and the poor level matching made it loudest. If true, that is probably best interpreted as evidence to support the idea that there is a tendency to perceive slightly louder sounds as sounding better.
tubetwister likes this.

Last edited by arnyk; 08-08-2014 at 04:26 AM.
arnyk is offline  
post #248 of 457 Old 08-08-2014, 07:04 AM
Senior Member
 
Gaugster's Avatar
 
Join Date: Feb 2005
Location: Chicago, IL
Posts: 325
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 31 Post(s)
Liked: 16
Quote:
Originally Posted by granroth View Post
My (limited) software shows that all six files have a dynamic range of between 13 and 17 dB and for each A-B pair, they are essentially equivalent. In any case, 17 dB is trivially within the redbook specs, so this is likely not what you are talking about.

If the songs don't have a dynamic range that exceeds either standard, then my assumption is that a perfect conversion (no conversion anomalies) would reproduce the dynamics with 100% accuracy.

Clearly they were not reproduced 100% since you could detect differences.

So first, do you disagree with my assumption that you would not be able to detect differences in dynamics between a redbook quality version of a song and a high-res version if the full dynamic range was handled by redbook?

You mention that it could be related to the conversion process. What is your actual opinion (not me putting words in your mouth) on why you could detect differences in dynamics in those samples?



I would love it if you did. It would either answer some questions or add fuel to our existing ones, both worthwhile results. If you could still detect those differences in dynamics, then they clearly would have nothing to do with redbook vs high-res. If you no longer could hear those, though, then the mystery of what is going on continues.
Granroth, what software do you use (or recommend) to evaluate the dynamics etc... for these digital files. I would like to understand how you are digitally comparing them. I am thinking to measure the real output (complete audio chain) via REW if possible for comparison/verification. However I will have to study how to play a specific segment of the music as opposed to sinewave sweeps normally generated by REW.

Will try to muck around with ABX software this weekend. Plenty of examples of the go to software on this front.

Gaugster

Last edited by Gaugster; 08-08-2014 at 07:22 AM.
Gaugster is offline  
post #249 of 457 Old 08-08-2014, 07:18 AM
Senior Member
 
Gaugster's Avatar
 
Join Date: Feb 2005
Location: Chicago, IL
Posts: 325
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 31 Post(s)
Liked: 16
Quote:
Originally Posted by tubetwister View Post
Quote:
Originally Posted by Gaugster
If the songs don't have a dynamic range that exceeds either standard, then my assumption is that a perfect conversion (no conversion anomalies) would reproduce the dynamics with 100% accuracy.
I'll leave that one to the resident experts !
I was misquoted per above. Not my words.

No harm done. We're all friends here.

Gaugster
Gaugster is offline  
post #250 of 457 Old 08-08-2014, 07:37 AM
Senior Member
 
Gaugster's Avatar
 
Join Date: Feb 2005
Location: Chicago, IL
Posts: 325
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 31 Post(s)
Liked: 16
Quote:
Originally Posted by krabapple View Post
Also good to know if he was using the *first* set of AIX files (which had a 0.2dB level difference) or the second , 'corrected' set.
Good point. I am using the second version of the files that are reported to be level matched. However I did submit my thoughts on the first set before the test flaw was found. Never learned the results.

I am 'Taking the Bait' and will fiddle with ABX testing and digital analysis tools to better understand how I could have detected either a fundamental difference and/or mere preference. Sounds like a good time!!!!!!!!

Gaugster
Gaugster is offline  
post #251 of 457 Old 08-08-2014, 08:09 AM
Advanced Member
 
granroth's Avatar
 
Join Date: Jan 2012
Location: Gilbert, AZ
Posts: 631
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 185 Post(s)
Liked: 149
Quote:
Originally Posted by Gaugster View Post
Granroth, what software do you use (or recommend) to evaluate the dynamics etc... for these digital files. I would like to understand how you are digitally comparing them. I am thinking to measure the real output (complete audio chain) via REW if possible for comparison/verification. However I will have to study how to play a specific segment of the music as opposed to sinewave sweeps normally generated by REW.
I did the quick route and used the software from the Dynamic Range Database site (http://dr.loudness-war.info/). They have versions for Mac and Windows (including foobar component).

That site lists the dynamic range of some 65,000 albums (minimum and maximum by track). It's notable to me that the maximum dynamic range of any track in their database is still "only" 34 dB -- comfortably within the 96 dB range allowed by redbook.

Quote:
Originally Posted by Gaugster View Post
Will try to muck around with ABX software this weekend. Plenty of examples of the go to software on this front.
Awesome
granroth is offline  
post #252 of 457 Old 08-08-2014, 11:15 AM
AVS Special Member
 
m. zillch's Avatar
 
Join Date: Oct 2006
Posts: 4,352
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 335 Post(s)
Liked: 284
My favorite way to level match, which anyone can do with the right test signal, is sometimes affectionately referred to as the "Larry Klein method". Anyone who gets my reference [I'm betting only one or two here, who like me has examined this principal for decades] is probably laughing a bit right now and thinking, "It's funny and ironic, but its probably for the most part pretty accurate!"


Larry was a great guy and one of the last reviewers on the scene here for science and the truth, rather than the shameless promotion of snake oil and the advertising dollars it brings (a trend more prevalent now than it ever has been in the history of hifi). We lost his great brilliance last year:
http://www.strata-gee.com/2013/01/15...view-magazine/

In A/V reproduction accuracy, there is no concept of "accounting for taste". We don't "pick" the level of bass any more than we get to pick the ending of a play. High fidelity is an unbiased, neutral, exact copy (or "reproduction") of the original source's tonal balance, timing, dynamics, etc..


Last edited by m. zillch; 08-08-2014 at 12:47 PM.
m. zillch is offline  
post #253 of 457 Old 08-08-2014, 12:42 PM
AVS Special Member
 
tubetwister's Avatar
 
Join Date: Feb 2013
Location: Sacramento delta N. Cal. US Don't trust any air I can't see ☺
Posts: 3,207
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 634 Post(s)
Liked: 382
Quote:
Originally Posted by Gaugster View Post
I was misquoted per above. Not my words.

No harm done. We're all friends here.
sorry bud I'll edit or take the quote out since I obviously can't attribute it correctly

I still think we are watching paint dry here though some of it just dries a little faster than others that's all

Hires Music formats ..............."Why does it sound like a CD ?" ............. can we make it louder "?
"The wireless music box has no commercial value. Who would pay for a message sent to nobody in particular?"
- David Sarnoff's associates at RCA the 1920's -

Last edited by tubetwister; 08-08-2014 at 12:51 PM.
tubetwister is offline  
post #254 of 457 Old 08-08-2014, 02:12 PM
AVS Special Member
 
tubetwister's Avatar
 
Join Date: Feb 2013
Location: Sacramento delta N. Cal. US Don't trust any air I can't see ☺
Posts: 3,207
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 634 Post(s)
Liked: 382
Quote:
Originally Posted by arnyk View Post
Mismatching the levels in a comparison of audio components is not the least bit good, and in the quantities measured, it suggests carelessness, incompetence or malevolence. Tough words, but true.

BTW full disclosure: The people involved in doing this test are not exactly my BFFs. We have been acquainted for a long time.

Someone observed that the final listener ranking of the equipment put the most expensive component first, and the poor level matching made it loudest. If true, that is probably best interpreted as evidence to support the idea that there is a tendency to perceive slightly louder sounds as sounding better.
mm? ... audiophool website rigs the testing so the more expensive gear tests better (maybe with a hidden or at least thinly disguised agenda ).mmm LMT ? ......................yes more than plausible IMO

Hires Music formats ..............."Why does it sound like a CD ?" ............. can we make it louder "?
"The wireless music box has no commercial value. Who would pay for a message sent to nobody in particular?"
- David Sarnoff's associates at RCA the 1920's -
tubetwister is offline  
post #255 of 457 Old 08-08-2014, 02:34 PM
AVS Special Member
 
krabapple's Avatar
 
Join Date: Apr 2003
Location: in a state bordered by Kentucky and Maine
Posts: 5,360
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 177 Post(s)
Liked: 229
Quote:
Originally Posted by jj_0001 View Post
What I haven't seen is any discussion on the exact response of the SRC that was used, which is, to put it mildly, of paramount importance. The fact that the 0.2dB came about due to SRC in the first place seems to suggest that there may be some passband slope or ripple, and frankly, that needs to be tested before any kind of listening test can be even judged to have meaning.


I'm not familiar with the SRC that was used to convert the AIX files (Sonic Solutions Sonic Process ), and I'm not sure which SRC Arny used for his sets of files (possibly Cool Edit's?)



As there are excellent SRCs available 'for free' (e.g Audacity 2.0.3's , Adobe Audition 3's , Sox -- see SRC comparisons at http://src.infinitewave.ca) this step should not have caused as much fuss as it has.
krabapple is offline  
post #256 of 457 Old 08-08-2014, 02:55 PM
Advanced Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 672
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 215
Quote:
Originally Posted by krabapple View Post
I'm not familiar with the SRC that was used to convert the AIX files (Sonic Solutions Sonic Process ), and I'm not sure which SRC Arny used for his sets of files (possibly Cool Edit's?)

As there are excellent SRCs available 'for free' (e.g Audacity 2.0.3's , Adobe Audition 3's , Sox -- see SRC comparisons at http://src.infinitewave.ca) this step should not have caused as much fuss as it has.
For the AIX material, they refused to use anything except the Sonic Solutions SRC. At that point, I pretty much blew them off as not acting in good faith. The PDF manual of the Sonic Solutions SRC is here.
andyc56 is offline  
post #257 of 457 Old 08-08-2014, 03:03 PM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,387
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 762 Post(s)
Liked: 1175
Quote:
Originally Posted by krabapple View Post
I'm not familiar with the SRC that was used to convert the AIX files (Sonic Solutions Sonic Process ), and I'm not sure which SRC Arny used for his sets of files (possibly Cool Edit's?)
I used the SRC in Cool Edit Pro 2.1, which is AFAIK as blameless as any. It will certainly pass the tests used by src.infinitewave.ca with results equal to the best.
arnyk is offline  
post #258 of 457 Old 08-08-2014, 04:34 PM
Senior Member
 
stereoeditor's Avatar
 
Join Date: Feb 2010
Posts: 343
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 140 Post(s)
Liked: 39
Quote:
Originally Posted by granroth View Post
the Dynamic Range Database site (http://dr.loudness-war.info/) . . . lists the dynamic range of some 65,000 albums (minimum and maximum by track). It's notable to me that the maximum dynamic range of any track in their database is still "only" 34 dB -- comfortably within the 96 dB range allowed by redbook.
You're confusing two different usages for the term "Dynamic Range." The DR database figure is actually what used to be called the "crest factor" or the ratio between the peak and mean levels in the file. It is not the same as the 96dB often quoted as the dynamic range of the CD medium, which, simply put, is the ratio between the maximum level and and the 16-bit noise/quantization floor.

John Atkinson
Editor, Stereophile
RichB, amirm, krabapple and 1 others like this.
stereoeditor is offline  
post #259 of 457 Old 08-08-2014, 04:48 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by stereoeditor View Post
You're confusing two different usages for the term "Dynamic Range." The DR database figure is actually what used to be called the "crest factor" or the ratio between the peak and mean levels in the file. It is not the same as the 96dB often quoted as the dynamic range of the CD medium, which, simply put, is the ratio between the maximum level and and the 16-bit noise/quantization floor.

John Atkinson
Editor, Stereophile
Indeed. There are tracks there that they say only has 6 db of dynamic range. That is equiv. to a single bit in digital domain (where that 96 db number comes from). Who here wants to claim that if we shrink that file down to 1 bit, it would be transparent to its source? I hope no one is saying that.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #260 of 457 Old 08-08-2014, 05:08 PM
AVS Special Member
 
RayDunzl's Avatar
 
Join Date: Sep 2013
Location: Tampa, Florida
Posts: 1,327
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 458 Post(s)
Liked: 582
Quote:
Originally Posted by amirm View Post
Indeed. There are tracks there that they say only has 6 db of dynamic range. That is equiv. to a single bit in digital domain (where that 96 db number comes from). Who here wants to claim that if we shrink that file down to 1 bit, it would be transparent to its source? I hope no one is saying that.
That would be 1 MSB for 6db, not one LSB.

The big MSB is worth a whole lot of LSBs. Do the math.

I'll be back later...


1.5RQ > digits > OpenDRC-DI > DEQ2496 > DAC2 > KCT > FPB 350mcx > reQuest > Sweetspot
RayDunzl is online now  
post #261 of 457 Old 08-08-2014, 06:20 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by RayDunzl View Post
That would be 1 MSB for 6db, not one LSB.

The big MSB is worth a whole lot of LSBs. Do the math.
For 6 db of "dynamic range" MSB and LSB are one in the same as we only have one bit.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #262 of 457 Old 08-08-2014, 06:53 PM
AVS Special Member
 
RayDunzl's Avatar
 
Join Date: Sep 2013
Location: Tampa, Florida
Posts: 1,327
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 458 Post(s)
Liked: 582
6db of dynamic range at -90'ish is one LSB 0000000000000000 to 0000000000000001 = 6db

6db of dynamic range of 16/44 at full scale to -6db is 16,384 bits 0011111111111111 to 0111111111111111 for a positive sample


Resolution - number of bits per 6db - is reduced as the level of the signal is decreased.

Go think about it some more.
tubetwister likes this.

I'll be back later...


1.5RQ > digits > OpenDRC-DI > DEQ2496 > DAC2 > KCT > FPB 350mcx > reQuest > Sweetspot
RayDunzl is online now  
post #263 of 457 Old 08-08-2014, 07:26 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by RayDunzl View Post
6db of dynamic range at -90'ish is one LSB 0000000000000000 to 0000000000000001 = 6db

6db of dynamic range of 16/44 at full scale to -6db is 16,384 bits 0011111111111111 to 0111111111111111 for a positive sample


Resolution - number of bits per 6db - is reduced as the level of the signal is decreased.

Go think about it some more.
There is nothing to think about as you are talking about the role of one bit in 16 bit sample, and I am talking about 1 bit out of 1 bit which was the counter to the original argument. That if all we need is 35 db, then 16 bits is way more than sufficient. I simply changed that to the track that has 6 db of "dynamic range" and asked if there is a need for more than 1 bit then. The role of least significant bit in 16 bits is not material to this discussion.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #264 of 457 Old 08-08-2014, 07:38 PM
AVS Special Member
 
NorthSky's Avatar
 
Join Date: Jan 2013
Location: Canada - West Island: Vancouver, South Direction: Go East
Posts: 4,730
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 2044 Post(s)
Liked: 598
Quote:
Originally Posted by amirm View Post
Indeed. There are tracks there that they say only has 6 db of dynamic range. That is equiv. to a single bit in digital domain (where that 96 db number comes from). Who here wants to claim that if we shrink that file down to 1 bit, it would be transparent to its source? I hope no one is saying that.
DSD? ...One bit.

* 20 bits is sufficient (Bob Stuart from Meridian); not 16 (not enough), no need @ 24, @ 32, and @ 64.

...And double of 44.1kHz => 88.2kHz is also sufficient (Bob Stuart again).
NorthSky is offline  
post #265 of 457 Old 08-08-2014, 09:20 PM
AVS Special Member
 
tubetwister's Avatar
 
Join Date: Feb 2013
Location: Sacramento delta N. Cal. US Don't trust any air I can't see ☺
Posts: 3,207
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 634 Post(s)
Liked: 382
Quote:
Originally Posted by NorthSky View Post
DSD? ...One bit.

* 20 bits is sufficient (Bob Stuart from Meridian); not 16 (not enough), no need @ 24, @ 32, and @ 64.

...And double of 44.1kHz => 88.2kHz is also sufficient (Bob Stuart again).
For playback or no ? .............................. dunno I've got some pretty good sounding 16/44.1

Quote:
Meridian Audio is an English manufacturer of high-performance, high-fidelity audio and video components and systems founded in 1977 by Bob Stuart and Allen Boothroyd. -wikipedia-
Quote:
20 bits is sufficient (Bob Stuart from Meridian); not 16 (not enough), no need @ 24, @ 32, and @ 64.
Consider the source and the possible motivations of anything they say .................. just saying ☺☺☺☺

OTOH for recording/editing/ mastering yes more bits are usually good . ♫ ☺☺☺

Smokey the Bear will be 70 tomorrow he will not be celebrating with birthday candles for obvious reasons ☺☺ ☺

My boys (S.C. 49'ers ) got they $&% handed to them in pre season ball by Baltimore today 23-3 !
........................Raiders lost to Minnesota 10-6

Hires Music formats ..............."Why does it sound like a CD ?" ............. can we make it louder "?
"The wireless music box has no commercial value. Who would pay for a message sent to nobody in particular?"
- David Sarnoff's associates at RCA the 1920's -

Last edited by tubetwister; 08-08-2014 at 10:14 PM.
tubetwister is offline  
post #266 of 457 Old 08-08-2014, 09:47 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,375
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 981 Post(s)
Liked: 547
Quote:
Originally Posted by NorthSky View Post
DSD? ...One bit.
It is "1 bit" but running at a sampling rate of 2,800 Khz instead of CD's 44 Khz. Think about what happens if you sample faster and faster. The amount of change that occurs becomes smaller and smaller and hence we need fewer bits.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #267 of 457 Old 08-08-2014, 10:40 PM
AVS Special Member
 
NorthSky's Avatar
 
Join Date: Jan 2013
Location: Canada - West Island: Vancouver, South Direction: Go East
Posts: 4,730
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 2044 Post(s)
Liked: 598
Quote:
Originally Posted by tubetwister View Post
For playback or no ? .............................. dunno I've got some pretty good sounding 16/44.1
Both; record, and playback. ...Me too, I got some excellent sounding Red Book CDs.

Quote:
Originally Posted by tt
Consider the source and the possible motivations of anything they say .................. just saying ☺☺☺☺
Always. ...For best sound reproduction. ...What they say and how we interpret it are sometimes contradictory.

Quote:
Originally Posted by tt
OTOH for recording/editing/ mastering yes more bits are usually good . ♫ ☺☺☺
Sampling rate (frequency) is more important than bit depth (quantization). ...According to audio gurus.

Quote:
Originally Posted by tt
Smokey the Bear will be 70 tomorrow he will not be celebrating with birthday candles for obvious reasons ☺☺ ☺

My boys (S.C. 49'ers ) got they $&% handed to them in pre season ball by Baltimore today 23-3 !
........................Raiders lost to Minnesota 10-6
Happy Birthday! ...70 is still very young, my granma kept @ it till 99 and 1/2.
Geraldine, was/is her name. ...A pianist and a singer.

Not into ball, sorry. ...I played though, in a real club, when I was about 50 years younger.
{I was a master @ playing marbles back then; you played too?}

Have a great weekend my friend.
NorthSky is offline  
post #268 of 457 Old 08-08-2014, 10:42 PM
AVS Special Member
 
m. zillch's Avatar
 
Join Date: Oct 2006
Posts: 4,352
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 335 Post(s)
Liked: 284
Quote:
Originally Posted by jj_0001 View Post
What I haven't seen is any discussion on the exact response of the SRC that was used, which is, to put it mildly, of paramount importance. The fact that the 0.2dB came about due to SRC in the first place seems to suggest that there may be some passband slope or ripple, and frankly, that needs to be tested before any kind of listening test can be even judged to have meaning...

... Time alignment can also bite you. If the time alignment is a few milliseconds off, the switching sensation will give it completely away, for instance.
[emphasis mine]

I expressed a time alignment concern myself earlier that day (but in another forum) as a possible explanation as to why I can hear an extremely subtle distinction and can pass the foobarABX test:

Quote:
Originally Posted by m.zillch
So in the end what causes there to be a difference? Not sure, but my current theory is the multiple SRC steps to make these cuts not only caused a minor fraction of a dB level change, which was said to have been manually corrected for in the "Take 2" version of the test, but maybe there was also a slight temporal change and one of the two cuts is slightly misaligned on the time axis, by a few samples. I'm not an expert on this but if someone with better skills than me could open them in Audacity, etc. and magnify it to look for that, it would be cool.
Since no one there stepped up to the plate, I decided to try to learn how to use Audacity on my own, having almost no prior experience in its functionality, and as it turns out I was right!

I called it! Took me forever to figure out how to "split stereo tracks" and show just one of them, the left channel, taken from two different stereo files, on top of each other, but eventually I did and here they are. The newer, level matched versions of Mosaic A2 and B2 (without the .2 dB level discrepancy) have a time discrepancy of what looks to be almost exactly a hundredth of a second, or 10 milliseconds:


As you can see, everyone, this loud peak in the L ch appears at about 1m53.04s in Mosaic A2 versus 1m53.03s in Mosaic B2. That's subtle but perceptible in A/B switching.

I encourage Audacity veterans to attempt to replicate my work.

Clearly we need a "Take 3" version of this test, properly level matched and time aligned, and all test scores currently posted should be discarded.
RichB, CharlesJ and andyc56 like this.

In A/V reproduction accuracy, there is no concept of "accounting for taste". We don't "pick" the level of bass any more than we get to pick the ending of a play. High fidelity is an unbiased, neutral, exact copy (or "reproduction") of the original source's tonal balance, timing, dynamics, etc..


Last edited by m. zillch; 08-09-2014 at 01:32 PM.
m. zillch is offline  
post #269 of 457 Old 08-08-2014, 10:49 PM
AVS Special Member
 
NorthSky's Avatar
 
Join Date: Jan 2013
Location: Canada - West Island: Vancouver, South Direction: Go East
Posts: 4,730
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 2044 Post(s)
Liked: 598
Quote:
Originally Posted by amirm View Post
It is "1 bit" but running at a sampling rate of 2,800 Khz instead of CD's 44 Khz. Think about what happens if you sample faster and faster. The amount of change that occurs becomes smaller and smaller and hence we need fewer bits.
Abso!utely Amir. ...Everything clocking faster and spinning faster too. ...Less errors less quantization needed.
...The higher in the ultrasonic frequency range the less noise in the human audible portion of it.

Main key is; a master recording engineer (recordist), with the skills to position the mic(s) and the musicians
and quality recording machines and mics. ...Plus that beautiful magical music playing, of course; that's the essence.
NorthSky is offline  
post #270 of 457 Old 08-08-2014, 10:54 PM
AVS Special Member
 
NorthSky's Avatar
 
Join Date: Jan 2013
Location: Canada - West Island: Vancouver, South Direction: Go East
Posts: 4,730
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 2044 Post(s)
Liked: 598
Question: Where harmonic distortion is mainly coming from?

And: Quantization noise; why does it even exist?

And: Intermodulation noise; can it be derived from phase shifts.
NorthSky is offline  
Reply Audio Theory, Setup, and Chat

Tags
frontpage

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off