Debate Thread: Scott's Hi-res Audio Test - Page 21 - AVS Forum
Forum Jump: 
 734Likes
 
Thread Tools
post #601 of 2920 Old 06-15-2014, 02:55 AM
Member
 
Kees de Visser's Avatar
 
Join Date: Jan 2009
Posts: 41
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 24 Post(s)
Liked: 18
Quote:
Originally Posted by stereoeditor View Post
All the wideband, unweighted S/N ratios published in Stereophile are measured over the full <10Hz-500kHz bandwidth of my Audio Precision SYS2722. This is routinely mentioned in my testing, just not in the case of the Audio Research SP20 review.
It's not just for DSD fans , but wouldn't it make sense to specify both 500kHz and 20kHz bandwidth results ? (assuming there is more perceptual relevance in the latter)
Kees de Visser is offline  
Sponsored Links
Advertisement
 
post #602 of 2920 Old 06-15-2014, 05:20 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 stereoeditor View Post
Quote:
Originally Posted by arnyk View Post
Quote:
Originally Posted by DonH50 View Post
I noticed the SNR of the ARC SP20 ($9k preamp) phono stage as measured in Stereophile was 55/58 dB for the left and right channels... Unweighted and wideband, the line stage SNR was about 70 dB, and bandlimited to the audio band it was about 79 dB.
any noise measurement that is not presented with information about the measurement bandwidth is by definition faulty. "Wideband" does not suffice. Not your doing Don, but don't put too much credibility in people who stumble at this point.
All the wideband, unweighted S/N ratios published in Stereophile are measured over the full <10Hz-500kHz bandwidth of my Audio Precision SYS2722. This is routinely mentioned in my testing, just not in the case of the Audio Research SP20 review.
Please explain how measuring broadband noise in the 20 KHz-500 KHz band reflects on the audible performance of audio gear.
arnyk is offline  
post #603 of 2920 Old 06-15-2014, 05:31 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 krabapple View Post
This, perhaps, puts warnings about HDMI/gear that inject 'thousands of picoseconds' of jitter, into perspective. Some more exactitude on the word 'thousands' would seem to be important there (and type & spectrum of the jitter)
Both HDMI gear and analog tape recorders may have peak jitter @ 30 Hz. In analog tape machines, the jitter is given in percentage, and 0.05% peak jitter represents really good performance that is generlaly thought to be free of audible consequences. BTW it has been long observed that the FM distortion in loudspeakers is in this same range. 0.05% jitter is the same as 500 microseconds of jitter, 500,000 picoseconds, or 500,000,000 nanoseconds of jitter.

Why is 500,000,000 of jitter at 30 Hz not a problem when it comes from an analog tape recorder or a loudspeaker, but is thought to be a horrifically severe problem and in fact a problem that is so bad that it is never ever even close to being seen in digital audio gear?

Note that a CD, DVD-A or SACD that is transcribed from an analog tape master (as very many are), already has something like 500,000,000 nanoseconds or jitter built into it courtesy of its analog tape sourcing.

And, as soon as any music is played on speakers it picks up another 500,000,000 nanoseconds (more or less) jitter!

Why is speaker and analog tape machine jitter given a pass?
arnyk is offline  
post #604 of 2920 Old 06-15-2014, 06:06 AM
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 Kees de Visser View Post
Quote:
Originally Posted by stereoeditor View Post
All the wideband, unweighted S/N ratios published in Stereophile are measured over the full <10Hz-500kHz bandwidth of my Audio Precision SYS2722. This is routinely mentioned in my testing, just not in the case of the Audio Research SP20 review.
It's not just for DSD fans , but wouldn't it make sense to specify both 500kHz and 20kHz bandwidth results ? (assuming there is more perceptual relevance in the latter)
All Stereophile's reviews of electronics, since I started performing them, include the unweighted S/N ratio measured over both <10Hz-500kHz and 22Hz-22kHz bandwidths and the A-weighted ratio.

John Atkinson
Editor, Stereophile

Last edited by stereoeditor; 06-15-2014 at 07:55 AM. Reason: To correct typo
stereoeditor is offline  
post #605 of 2920 Old 06-15-2014, 06:16 AM
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 arnyk View Post
Please explain how measuring broadband noise in the 20kHz-500kHz band reflects on the audible performance of audio gear.
The wideband ratio will include contributions from flicker noise (<20Hz) and how susceptible the DUT is to RF interference. For example, a preamp we recently reviewed had an unweighted S/N ratio that improved by 30dB when the bandwidth was reduced from 500kHz to 22kHz. Given that many class-D amplifiers use modulators that operate in the 400kHz region and can have up to 500mV of noise at that frequency present at the speaker outputs, this preamplifier might have audible issues when used with such amplifiers.

I am not saying that there will be audible problems with this preamp in such a situation. However, I feel, as with Paul Miller's tests of HDMI-input receivers that have been quoted elsewhere in this thread, the role of an equipment review is to offer as full a picture of the DUT's behavior as is practicably possible.

John Atkinson
Editor, Stereophile
stereoeditor is offline  
post #606 of 2920 Old 06-15-2014, 07:11 AM
AVS Addicted Member
 
Chu Gai's Avatar
 
Join Date: Sep 2002
Location: NYC area
Posts: 14,964
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 317 Post(s)
Liked: 595
Quote:
Originally Posted by stereoeditor View Post
Quote:
Originally Posted by arnyk View Post
Please explain how measuring broadband noise in the 20kHz-500kHz band reflects on the audible performance of audio gear.
The wideband ratio will include contributions from flicker noise (<20Hz) and how susceptible the DUT is to RF interference. For example, a preamp we recently reviewed had an unweighted S/N ratio that improved by 30dB when the bandwidth was reduced from 500kHz to 22kHz. Given that many class-D amplifiers use modulators that operate in the 400kHz region and can have up to 500mV of noise at that frequency present at the speaker outputs, this preamplifier might have audible issues when used with such amplifiers.

I am not saying that there will be audible problems with this preamp in such a situation. However, I feel, as with Paul Miller's tests of HDMI-input receivers that have been quoted elsewhere in this thread, the role of an equipment review is to offer as full a picture of the DUT's behavior as is practicably possible.

John Atkinson
Editor, Stereophile
Can you explain more how this 'could' could happen? Also, what kind of measurement results would suggest there probably would be audible consequences as opposed to unlikely?

"I've found that when you want to know the truth about someone that someone is probably the last person you should ask." - Gregory House
Chu Gai is offline  
post #607 of 2920 Old 06-15-2014, 08:37 AM
Senior Member
 
hevi's Avatar
 
Join Date: Jan 2007
Posts: 365
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 33 Post(s)
Liked: 48
Quote:
Originally Posted by arnyk View Post
Both HDMI gear and analog tape recorders may have peak jitter @ 30 Hz. In analog tape machines, the jitter is given in percentage, and 0.05% peak jitter represents really good performance that is generlaly thought to be free of audible consequences.

Please don't use the term "jitter" for analog signals. Jitter is an unwanted time offset in discrete signal transitions. The correct term to use for the analog domain is phase noise, which is a phase offset that continually changes the timing of the signal.
amirm and stereoeditor like this.

hevi is offline  
post #608 of 2920 Old 06-15-2014, 10:06 AM - Thread Starter
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: 546
Quote:
Originally Posted by arnyk View Post
Both HDMI gear and analog tape recorders may have peak jitter @ 30 Hz.
Good morning Arny. Thanks again for your great fact fill posts! Much appreciated.

I am a bit puzzled though by your use of the word "may." That means we don't know if it does or does not have that frequency. Given that, any follow up explanation may not apply so not sure what the purpose of the exercise was.

That aside, here is my quick sample of three HDMI products: two AVRs and a mid-to-high end Processor:



I have put a cursor on the first peak around our main tone at 12,000 Hz. If we subtract the frequencies of those we get, 30 Hz, 65 Hz and 115 Hz as the highest and first peaks around our test tone. Yes, one of them which I believe where you got from my measurements is your 30 Hz number above. We see however that the other two products have differing peaks so an analysis around 30 Hz is non sequitur if you are using my measurements.

I have not looked at measurements of analog devices enough to know that one sees 30 Hz there. Do you have a few data points like above that we can look at for the sake of argument?

Quote:
Originally Posted by arnyk View Post
In analog tape machines, the jitter is given in percentage, and 0.05% peak jitter represents really good performance that is generlaly thought to be free of audible consequences.
Generally? Is there a double blind test for that which we can read Arny?

Quote:
Originally Posted by arnyk View Post
Why is 500,000,000 of jitter at 30 Hz not a problem when it comes from an analog tape recorder or a loudspeaker, but is thought to be a horrifically severe problem and in fact a problem that is so bad that it is never ever even close to being seen in digital audio gear?
I am sorry to say this but once again you rewrite my post differently and go on to draw conclusions. I did not say HDMI performance is "a horrifically severe problem." The phrasing I would use is sloppy performance meaning it could have been much better if there was more care in its design. I showed examples of other companies doing exactly that. I don't believe analog wow and flutter can be shown to be improved so much as the case in HDMI. But again I am not an analog guy and you would surely present us with data that says analog wow and flutter that you are talking about could be improved by tens of dbs.

Quote:
Originally Posted by arnyk View Post
Note that a CD, DVD-A or SACD that is transcribed from an analog tape master (as very many are), already has something like 500,000,000 nanoseconds or jitter built into it courtesy of its analog tape sourcing.
That is not a factor in our discussions Arny. What is in the music that we get was approved by the talent. If it had that kind of distortion, that is what was liked as their art and presented to us on disc. Our job is not then to keep adding more salt to a salty meal. High Fidelity means faithfulness to what is delivered to us without additional distortions and colorations. It would be quite a defeatist attitude to say that all of these distortions are just fine because the source recording may have another form of them.

Quote:
Originally Posted by arnyk View Post
And, as soon as any music is played on speakers it picks up another 500,000,000 nanoseconds (more or less) jitter!
Well, it doesn't pick up any "jitter" because it is not a digital device. But again, the argument remains as above. This is no license for us to go and distort the sound in different manner on our device. Furthermore, distortions are additive. It would be very difficult to prove that any and all jitter upstream will get masked by FM modulation in the speaker. FM modulation will potentially take those jitter variations and amplify them by adding its own FM modulation. A single sideband jitter of say 1000 Hz will get additional pairs added to it due to FM modulation of the speaker. That latter distortion is not an eraser for upstream distortions.

Quote:
Originally Posted by arnyk View Post
Why is speaker and analog tape machine jitter given a pass?
We are in a digital thread Arny. Whatever analog did or did not is not material as no one has claimed in this thread that analog is transparent to the source.

Back to HDMI, its performance is not predictable in the manner you are supposing. Here is a less zoomed spectrum of HDMI noise/distortion (please ignore the Onkyo Cursor value; I forgot to move it from where it was):



Notice the area that I have highlighted. We have a spike at the same frequency in three different devices and manufacturers. It reasons therefore that the cause of that distortion is the source. Lack of filtering in HDMI input clock of these machines is allowing that distortion to peak through as I explained yesterday. These spikes are way away from our 12 Khz tone at 7.1 Khz. Masking then is much weaker if at all existing at this frequency. Furthermore, these spikes are in a more sensitive area of our hearing than the 12 Khz. So handwaving around audibility of these distortions is much less effective here.

Before we run off and do another paper analysis of 7.1 Khz, keep in mind again that this distortion was caused by the source. This means that the source you use may have a completely different profile. Even for frequencies around our main tone, the response varied from device to device.

I think the only safe assumption at this point is that HDMI performance in mass market devices and some higher end ones is random. It will differ from device to device and will be source dependent. As such you can't proceed to compute numbers and such based on the sampling I am showing. You would need to have measurements of everyone's devices here to make any case at all. I suspect none of you have this kind of measurements so in that regard, you are "blind" to what distortions you are hearing.

What we know and you can take to the bank is that S/PDIF performance is superior to HDMI. You can also take to bank knowing that this didn't have to be this way. That HDMI distortions can be brought down significantly with good design. We also know that good HDMI designs do exist in high-end products. So a fair assessment is that when a capable designer is not under cost constraints, he is able to do much better work. These are the only conclusions in my opinion that we can draw given the data that we have.
stereoeditor likes this.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #609 of 2920 Old 06-15-2014, 10:55 AM
AVS Special Member
 
TVOD's Avatar
 
Join Date: Jun 2003
Posts: 4,881
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 74 Post(s)
Liked: 70
Quote:
Originally Posted by amirm View Post
The phrasing I would use is sloppy performance meaning it could have been much better if there was more care in its design.
I could say that about a lot of expensive professional equipment.
TVOD is offline  
post #610 of 2920 Old 06-15-2014, 08:48 PM
Senior Member
 
UndersAVS's Avatar
 
Join Date: Mar 2008
Posts: 259
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 35
Quote:
Originally Posted by DonH50 View Post
It takes about 2 ns of random jitter to degrade the output to about a 12-bit noise floor (about 74 dB).

The argument can be made that signal-dependent (deterministic) jitter is easier to hear, and I think that is likely, but I have no idea what the threshold of audibility is for it, especially with a musical source instead of test tones.
Don,
Thank you for providing information. You are citing 2 NANOSECONDS or 2,000 picoseconds of random jitter reducing the noise floor. Amir's 'personal' target is a few hundred PICOs. Quite a difference.

I'm appreciate your forwardness, admitting that you don't know what level of jitter is inaudible. It seems to be a complicated topic and I appreciate your honesty. Modesty and an open mind are excellent qualities in my book.

I don't know whether the Acoust. Sci. & Tech. article has been discussed, but these opening comments provide illumination:

https://www.jstage.jst.go.jp/article...1/26_1_50/_pdf

Quote:
Accordingly, the maximum acceptable random jitter can be defined as the jitter that results in a 1-quantizer level error when the signal is of the highest recordable frequency [1,3]. When the highest recordable frequency is 20 kHz and the quantization bit number is 16, the maximum acceptable size of random jitter is 121.4 ps. In other words, jitter has to be smaller than 121.4 ps in order to reproduce a 20 kHz tone with a 16-bit resolution [3].
Even in the case of music sounds, the maximum acceptable size of random jitter can be estimated if the
maximum slope in the sound waveform is known. By
analyzing various signals in a lot of CDs, it was shown that jitter has to be as small as several hundreds ps in some cases to preserve a 16-bit resolution [3].
However, these estimations are not practical because such small distortions are supposed to be masked by the
background noise and the internal noise of the reproduction system in the real environment. It is much more practical to evaluate jitter based on its detection threshold.
UndersAVS is offline  
post #611 of 2920 Old 06-15-2014, 09:22 PM
Senior Member
 
UndersAVS's Avatar
 
Join Date: Mar 2008
Posts: 259
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 35
Quote:
Originally Posted by UndersAVS View Post
http://www.madronadigital.com/Librar...stortions.html

"Because we ignored masking, this may set too high a bar relative to our ability to hear such non-linear distortions. So if you like, you can derate them some. But don’t go too far. It is very hard to build a credible case for example that thousands of picoseconds of jitter as we usually get over HDMI is inaudible. The analysis I just provided shoots way too many holes in that line of reasoning. My personal target is a few hundred picoseconds.."

amir,
How do you arrive at the the personal target? "A few hundred picoseconds?"
Quote:
Originally Posted by amirm View Post
The metric is an engineering one that says if we have a 16 bit DAC, how low does its distortions need to be to equal the noise level of that DAC. Undithered 16-bit signal has a noise floor of about -96 dbFS relative to 0 db reference. We can work backward and determine what sinusoidal jitter we can have that has an amplitude distortion spikes at -96 dbFS. Using 20 KHz as the maximum frequency of interest, the math tells us the jitter amplitude has to be no more than 500 picoseconds or 0.5 nanoseconds.

Note that we can hear through noise. So the above metric does not assure inaudibility of the distortions. On the other hand as noted, real music may mask jitter. Then again distortions that we see are not a single jitter tone but many such tones as my measurements and that of Paul Miller show. So the actual value is pretty complex to determine. If we still insist on a single number, then the above value is a decent metric we can use to determine if our system is well implemented.

As a fun exercise, if we want to have distortions that are below 20 bit DAC, the jitter value shrinks to just 30 picoseconds! If we up the sampling rate to 96 KHz and attempt to reproduce all of its spectrum, the jitter spec now sinks down to just 10 picoseconds. If we now boost the resolution to 24 bits for "24 bit/96 KHz" the jitter spec becomes just 1 picoseconds!!!

The above numbers are computed using math and analysis of the amplitude of jitter sidebands. So the computation is quite precise and accurate. The notion then "why would I worry about picosecond jitter specs when audio is such low frequency" is put to rest. You absolutely need such low values of jitter if you are going to advertise that your DAC is say capable of 24-bit/96 KHz as just about any AVR today says.
Amir,

We have seen your AV Receiver measurements. I believe these are the Paul Miller measurements you are referring to:

Denon AVR-3803A
---------------
SPDIF: 560psec
HDMI: 3700psec

Onkyo TX-NR906
---------------
SPDIF: 470psec
HDMI: 3860psec

Pioneer SC-LX81
---------------
SPDIF: 37psec
HDMI: 50psec

Yamaha RX-V3900
---------------
SPDIF: 183psec
HDMI: 7660psec

I see no evidence of audible distortion in the Receiver measurements that you provided or those of Miller. Starting with your measurements; given the component with the worst performance, the Pioneer AV Receiver, -73dB measured in the most sensitive frequency of human hearing capacity, less than a half octave from the main tone, masking renders the distortion completely inaudible. The outlying distortion spikes barely reach thresholds of audibility and are still within the range of masking. Based upon the various masking graphs that I've pondered, I feel confident that there is no chance of recognition by any listener.

Foggier is the possibility of jitter audibility when summing the presentation of tones presented simultaneously within a couple of octaves from each other at equal amplitude. It seems possible that the sum of the outlying spikes *might* combine to create some audible messiness, which Miller's specs seem to imply. I don't know. Given the information available thus far, Miller's worst case measurement is that of the Yamaha's 7,660 ps, or 7.7ns.

Given the few rigorous listening tests that I've surveyed, picoseconds of jitter are of no consequence, and something on the order of 500 nanoseconds presents a possible threshold.

Acoust. Sci. & Tech. 26, 1 (2005) "Detection threshold for distortions due to jitter on digital audio"

Kaoru Ashihara1, Shogo Kiryu1, Nobuo Koizumi2, Akira Nishimura2,
Juro Ohga3, Masaki Sawaguchi4 and Shokichiro Yoshikawa5

https://www.jstage.jst.go.jp/article...1/26_1_50/_pdf

"It can be concluded that detection threshold for random jitter added to program materials is several hundreds ns even for well-trained listeners under their preferable listening conditions. According to Benjamin and Gannon, sinusoidal jitter as small as 30 ns (r.m.s.) might be detectable under a certain condition. Considering these results, the maximum acceptable size of jitter would be the
order of ns."


A less formal online test offered the same conclusion.

http://hddaudio.net/viewtopic.php?id=15
http://hddaudio.net/viewtopic.php?id=63

Last edited by UndersAVS; 06-15-2014 at 09:35 PM.
UndersAVS is offline  
post #612 of 2920 Old 06-15-2014, 09:42 PM
Senior Member
 
UndersAVS's Avatar
 
Join Date: Mar 2008
Posts: 259
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 35
Quote:
Originally Posted by arnyk View Post
The above statement seems to conflate signal-dependent and deterministic jitter when in fact they can be and often are independent. Much deterministic jitter is caused by modulation of the audio signal by spurious noises inside the AVR such as power line noise. In optical players it is possible for electrical noise and surges from the power supply due to operation of optical servos to be the source of such modulation. The potential sources of this sort of thing form an almost endless list but they are typically addressed well enough to render all related forms of jitter to below everybody's audibility.

While much has been made of signal-dependent jitter due to S/PDIF and interconnecting cables interacting with the audio signal, the more common causes we see today appear to be due to the non-audio properties of the signal such as data block sizes and arrival times.

The best source that I know of for information about the threshold of audibility for FM distortion (AKA jitter) is Zwicker and Fastl (2007) which I've posted here several times. I'm not encouraged to post it again and again because I've posted it before and nobody seems to pay any attention to it. Also the forum seems to have changed its procedures for posting pictures in such a way to make doing so far more difficult.

I've also set up some .wav files with variable amounts of jitter starting with unmistakable and ending with inaudible that involves the same jitter frequencies as we see in AVRs, involving both sine waves and music. Just download, plug into FooBar2000 and ABX. I've published links to them several times but I see zero signs of their usage.

Again, there appears to be zero interest in anything but ignoring the relevant facts, whether by authoritative and recognized scientific references or by means of their own personal experience. Instead people seem to want to wander around muttering nonsense and acting like its all over their heads, I guess.

So anybody who "has no idea" apparently has preserved themselves in that state by means of well practiced ignorance bordering on intellectual sloth. Either that, or they don't believe in Science and/or personal experience. When a strong proponent of high end audio pseudo science acts this way to avoid admitting his past serious mistakes, I get it. When people blunder around apparently pretending that they have no options, I don't.
Arny,
I hope you aren't including me in this group. If so, please let me know and I'll provide you with more details as to my understanding. Hopefully, I can alleviate any concerns that your contributions are wasted on my behalf.
UndersAVS is offline  
post #613 of 2920 Old 06-15-2014, 09:57 PM - Thread Starter
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: 546
Quote:
Originally Posted by UndersAVS View Post
Kaoru Ashihara1, Shogo Kiryu1, Nobuo Koizumi2, Akira Nishimura2,
Juro Ohga3, Masaki Sawaguchi4 and Shokichiro Yoshikawa5

https://www.jstage.jst.go.jp/article...1/26_1_50/_pdf

"It can be concluded that detection threshold for random jitter added to program materials is several hundreds ns even for well-trained listeners under their preferable listening conditions. According to Benjamin and Gannon, sinusoidal jitter as small as 30 ns (r.m.s.) might be detectable under a certain condition. Considering these results, the maximum acceptable size of jitter would be the
order of ns."
I don't know how many times I have explained that we cannot go by single digit jitter numbers. It is like saying the bacteria count that can make you sick in food without saying which bacteria we are talking about.

Above is case in point: it may seem like they are measuring the threshold of hearing for "jitter" but in reality, all they are measuring is the threshold detection for noise. Random jitter, simply generates random noise.

As you have seen from the many measurements I have shown, we are dealing with specific jitter frequencies which are not random. As such, the above test is of no value in this regard.

There are other problems with their tests which I won't go into. Suffice it to say, if you want to look informed about jitter, you don't want to ever cite that paper .

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #614 of 2920 Old 06-15-2014, 10:30 PM - Thread Starter
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: 546
Quote:
Originally Posted by UndersAVS View Post
I see no evidence of audible distortion in the Receiver measurements that you provided or those of Miller.
Well then, congratulations are in order because you know more than the committee on high resolution audio at AES Technical who had this to say again on this topic:

HDMI, the point-to-point connector required for BR and HD video, has excellent bandwidth and an Ethernet data link (HDMI 1.4), but lacks an audio clock. HDMI receivers must derive audio word clock from the video pixel clock, commonly resulting in very high jitter that affects quality and can be audible. Some high end receivers address the jitter and many companies are researching it but current solutions are expensive and uncommon.

This is the list of committee members:

Committee Members

Code:
 Bob Katz 	 Douglas McKinnie 	 Duane Wise 
 James Johnston 	 Jan Berg 	 Robert Schulein 
 Steven Harris 	 Thomas Sporer 	 Vicki R. Melchior 
 John Grant 	 Michael Kelly 	 Josh Reiss 
 Juergen Peissig 	 James J McTigue 	 David Rich 
 Sergio Liberman 	 Hans van Maanen 	 John Dawson 
 Bryan Martin 	 Oles Protsidym 	 Dylan-Thomas Vance 
 Mark Waldrep 	 Peter Craven 	 Partev Sarkissian 
 Ryan Stables 	 Michael Hudson 	 Yonghao Wang 
 David Jones
James Johnston is "JJ" who used to be an audio architecture in my group at Microsoft and whose name is mentioned quite often. He is in our (objectivity) camp.

John Dawson is a friend and the founder of Arcam, the company whose presentation I quoted earlier on Paul Miller measurements. He is also a frequent member here. He actually knows how to design the circuits we talk about here.

Vicki Melchior is the co-chair of the committee. If I remember right, she was part of the Boston Audio Society -- same group that brought us the Meyer and Moran test!

Anyway, what are the odds that you know this topic better than the group above as to have arrived at opposite conclusion? Not good, right?

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #615 of 2920 Old 06-16-2014, 12:29 AM
AVS Addicted Member
 
Chu Gai's Avatar
 
Join Date: Sep 2002
Location: NYC area
Posts: 14,964
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 317 Post(s)
Liked: 595
Quote:
Originally Posted by amirm View Post
Quote:
Originally Posted by UndersAVS View Post
Kaoru Ashihara1, Shogo Kiryu1, Nobuo Koizumi2, Akira Nishimura2,
Juro Ohga3, Masaki Sawaguchi4 and Shokichiro Yoshikawa5

https://www.jstage.jst.go.jp/article...1/26_1_50/_pdf

"It can be concluded that detection threshold for random jitter added to program materials is several hundreds ns even for well-trained listeners under their preferable listening conditions. According to Benjamin and Gannon, sinusoidal jitter as small as 30 ns (r.m.s.) might be detectable under a certain condition. Considering these results, the maximum acceptable size of jitter would be the
order of ns."
I don't know how many times I have explained that we cannot go by single digit jitter numbers. It is like saying the bacteria count that can make you sick in food without saying which bacteria we are talking about.

Above is case in point: it may seem like they are measuring the threshold of hearing for "jitter" but in reality, all they are measuring is the threshold detection for noise. Random jitter, simply generates random noise.

As you have seen from the many measurements I have shown, we are dealing with specific jitter frequencies which are not random. As such, the above test is of no value in this regard.

There are other problems with their tests which I won't go into. Suffice it to say, if you want to look informed about jitter, you don't want to ever cite that paper .
Julian Dunn cited the Dolby paper.

"I've found that when you want to know the truth about someone that someone is probably the last person you should ask." - Gregory House
Chu Gai is offline  
post #616 of 2920 Old 06-16-2014, 12:39 AM
Senior Member
 
UndersAVS's Avatar
 
Join Date: Mar 2008
Posts: 259
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 35
[quote=amirm;24998218]
Quote:
Originally Posted by UndersAVS View Post
I see no evidence of audible distortion in the Receiver measurements that you provided or those of Miller.
Quote:
Originally Posted by amirm
Well then, congratulations are in order because you know more than the committee on high resolution audio at AES Technical who had this to say again on this topic:

HDMI, the point-to-point connector required for BR and HD video, has excellent bandwidth and an Ethernet data link (HDMI 1.4), but lacks an audio clock. HDMI receivers must derive audio word clock from the video pixel clock, commonly resulting in very high jitter that affects quality and can be audible. Some high end receivers address the jitter and many companies are researching it but current solutions are expensive and uncommon.

This is the list of committee members:

Committee Members

Code:
 Bob Katz 	 Douglas McKinnie 	 Duane Wise 
 James Johnston 	 Jan Berg 	 Robert Schulein 
 Steven Harris 	 Thomas Sporer 	 Vicki R. Melchior 
 John Grant 	 Michael Kelly 	 Josh Reiss 
 Juergen Peissig 	 James J McTigue 	 David Rich 
 Sergio Liberman 	 Hans van Maanen 	 John Dawson 
 Bryan Martin 	 Oles Protsidym 	 Dylan-Thomas Vance 
 Mark Waldrep 	 Peter Craven 	 Partev Sarkissian 
 Ryan Stables 	 Michael Hudson 	 Yonghao Wang 
 David Jones
James Johnston is "JJ" who used to be an audio architecture in my group at Microsoft and whose name is mentioned quite often. He is in our (objectivity) camp.

John Dawson is a friend and the founder of Arcam, the company whose presentation I quoted earlier on Paul Miller measurements. He is also a frequent member here. He actually knows how to design the circuits we talk about here.

Vicki Melchior is the co-chair of the committee. If I remember right, she was part of the Boston Audio Society -- same group that brought us the Meyer and Moran test!

Anyway, what are the odds that you know this topic better than the group above as to have arrived at opposite conclusion? Not good, right?
amir,

The quote you're giving was already addressed. "Very high jitter that affects quality" is not measurable. "Can be audible" is not the same as "is audible." So I'm not superceding any opinion when I state that I see no evidence of audible jitter given the information provided. There may be hdmi-jitter issues that are discernable by listeners, but I haven't seen any evidence of this in your product measurements, your explanations, or the vaguely written citation. I referenced several sources that demonstrate inaudibility of the distortion you documented.

Therefore, your reply is merely an example of pointless argumentation, perhaps hoping that a faulty appeal to authority (providing a list of notables) will overcome conventional logic. If you don't understand, I recommend reading the following:

http://en.wikipedia.org/wiki/Appeal_to_authority

"Fallacious examples of using the appeal include any appeal to authority used in the context of logical reasoning, and appealing to the position of an authority or authorities to dismiss evidence,[2][3][4][5] as, while authorities can be correct in judgments related to their area of expertise more often than laypersons,[citation needed] they can still come to the wrong judgments through error, bias, dishonesty, or falling prey to groupthink. Thus, the appeal to authority is not a generally reliable argument for establishing facts.[6]"

Furthermore, the snarkiness noted in your opening statement is not only rude, but an example of hypocrisy. You got bummed out when I pointed out that you 'dunno,' or more mildly, "don't know" in the "Jitter" discussion, vacated the discussion, and derided others for lack of professionalism. Yet you experience no trouble doing the same or worse. The irony is that you don't know the seriousness of your measurements and made it public (in so many words), but find it unacceptable that *I* mention it.

I hope that the future holds something better. Ultimately, my interest in the topic is not about you or anyone else.
UndersAVS is offline  
post #617 of 2920 Old 06-16-2014, 12:53 AM
Senior Member
 
UndersAVS's Avatar
 
Join Date: Mar 2008
Posts: 259
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 35
Quote:
Originally Posted by UndersAVS View Post
Kaoru Ashihara1, Shogo Kiryu1, Nobuo Koizumi2, Akira Nishimura2,
Juro Ohga3, Masaki Sawaguchi4 and Shokichiro Yoshikawa5

https://www.jstage.jst.go.jp/article...1/26_1_50/_pdf

"It can be concluded that detection threshold for random jitter added to program materials is several hundreds ns even for well-trained listeners under their preferable listening conditions. According to Benjamin and Gannon, sinusoidal jitter as small as 30 ns (r.m.s.) might be detectable under a certain condition. Considering these results, the maximum acceptable size of jitter would be the
order of ns."
Quote:
Originally Posted by amirm View Post
I don't know how many times I have explained that we cannot go by single digit jitter numbers. It is like saying the bacteria count that can make you sick in food without saying which bacteria we are talking about.

Above is case in point: it may seem like they are measuring the threshold of hearing for "jitter" but in reality, all they are measuring is the threshold detection for noise. Random jitter, simply generates random noise.

As you have seen from the many measurements I have shown, we are dealing with specific jitter frequencies which are not random. As such, the above test is of no value in this regard.

There are other problems with their tests which I won't go into. Suffice it to say, if you want to look informed about jitter, you don't want to ever cite that paper .
The Ashihara et al study demonstrates that random jitter is inaudible below 500 ns. That is of value to me. And you have demonstrated distortion in a few products that would be rendered inaudible by masking.

My next task it to point out flaws in the article you wrote.
UndersAVS is offline  
post #618 of 2920 Old 06-16-2014, 02:04 AM
Member
 
Kees de Visser's Avatar
 
Join Date: Jan 2009
Posts: 41
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 24 Post(s)
Liked: 18
Quote:
Originally Posted by UndersAVS View Post
The Ashihara et al study demonstrates that random jitter is inaudible below 500 ns. That is of value to me. And you have demonstrated distortion in a few products that would be rendered inaudible by masking.
Ashihara used a jitter simulator for the study. Are we (all) in agreement that jitter simulation is a valid option for listening tests? Of course it requires "jitter profiles" and that's another subject. IMO a great advantage of simulation is that it's in the digital domain and the resulting audio files can be shared with others for evaluation.
(also see the paper of Hawksford)

Last edited by Kees de Visser; 06-16-2014 at 02:07 AM. Reason: added Hawksford
Kees de Visser is offline  
post #619 of 2920 Old 06-16-2014, 04:57 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 amirm View Post
Quote:
Originally Posted by UndersAVS View Post
I see no evidence of audible distortion in the Receiver measurements that you provided or those of Miller.
Well then, congratulations are in order because you know more than the committee on high resolution audio at AES Technical who had this to say again on this topic:

HDMI, the point-to-point connector required for BR and HD video, has excellent bandwidth and an Ethernet data link (HDMI 1.4), but lacks an audio clock. HDMI receivers must derive audio word clock from the video pixel clock, commonly resulting in very high jitter that affects quality and can be audible. Some high end receivers address the jitter and many companies are researching it but current solutions are expensive and uncommon.

This is the list of committee members:

Committee Members

Code:
 Bob Katz 	 Douglas McKinnie 	 Duane Wise 
 James Johnston 	 Jan Berg 	 Robert Schulein 
 Steven Harris 	 Thomas Sporer 	 Vicki R. Melchior 
 John Grant 	 Michael Kelly 	 Josh Reiss 
 Juergen Peissig 	 James J McTigue 	 David Rich 
 Sergio Liberman 	 Hans van Maanen 	 John Dawson 
 Bryan Martin 	 Oles Protsidym 	 Dylan-Thomas Vance 
 Mark Waldrep 	 Peter Craven 	 Partev Sarkissian 
 Ryan Stables 	 Michael Hudson 	 Yonghao Wang 
 David Jones
James Johnston is "JJ" who used to be an audio architecture in my group at Microsoft and whose name is mentioned quite often. He is in our (objectivity) camp.

John Dawson is a friend and the founder of Arcam, the company whose presentation I quoted earlier on Paul Miller measurements. He is also a frequent member here. He actually knows how to design the circuits we talk about here.

Vicki Melchior is the co-chair of the committee. If I remember right, she was part of the Boston Audio Society -- same group that brought us the Meyer and Moran test!

Anyway, what are the odds that you know this topic better than the group above as to have arrived at opposite conclusion? Not good, right?
More evidence that denial isn't just a river in Egypt. The above statement fails on grounds that have been discussed here many times before which is proof by authority. History has shown any number of well-known, highly regarded people who have supported ideas that we now know to be completely false.

It also fails on the grounds that strong assertion cannot make vague statements into specific claims.

I fear for the future credibility of anybody who continues to rest their reputation on obviously fallacious arguments.
arnyk is offline  
post #620 of 2920 Old 06-16-2014, 05:13 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 amirm View Post
Quote:
Originally Posted by arnyk View Post
Both HDMI gear and analog tape recorders may have peak jitter @ 30 Hz.
Good morning Arny. Thanks again for your great fact fill posts! Much appreciated.

I am a bit puzzled though by your use of the word "may." That means we don't know if it does or does not have that frequency.
Incorrect logic. In this context the word "may" means that analog tape recorders always have high amounts of jitter by the standards applied to digital audio, but that the frequency of peak jitter can vary with the design of the analog tape recorder.

The number 0.05% that I used is a typical number for high quality analog tape recorders. For example the following comes from the spec sheet for the Otari MX 5050 which you own and whose superior sound you have claimed on this forum:

http://www.otari.com/download/files/mx5050_ad_e1.pdf

Wow And Flutter Peak Weighted Per DIN 45507
15 ips: max. ±0.06 b%
7.5 ips: max. ±0.08 %
3.75 Ips: max. ±0.12 %

The DIN 45507 weighting curve is as follows:



Its general effect is to reduce reported jitter at all frequencies but 4 Hz. At 30 Hz, the jitter measurement is decreased by 10 dB. Note that in general jitter measurements are unweighted.

This begs the question why your personal analog tape equipment is given a free pass to add jitter?

At any rate I used the number 0.05% in my comments when in fact your own personal equipment is specified to perform significantly more poorly.

So then Amir, please explain the apparent obsession with digital jitter and the concurrent praise for analog equpiment with many orders of magnitude higher jitter?
arnyk is offline  
post #621 of 2920 Old 06-16-2014, 05:37 AM
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 Chu Gai View Post
Quote:
Originally Posted by stereoeditor View Post
The wideband ratio will include contributions from flicker noise (<20Hz) and how susceptible the DUT is to RF interference. For example, a preamp we recently reviewed had an unweighted S/N ratio that improved by 30dB when the bandwidth was reduced from 500kHz to 22kHz. Given that many class-D amplifiers use modulators that operate in the 400kHz region and can have up to 500mV of noise at that frequency present at the speaker outputs, this preamplifier might have audible issues when used with such amplifiers.
Can you explain more how this 'could' could happen?
RF noise is insidious. Ground paths may not have zero impedance at RF; as Martin Colloms has shown, speaker cables are potentially antennae when the assumption that the low-impedance loading of the loudspeaker is not necessarily true at RF; every feedback amplifier has two input ports, the second being the output; and as Bruce Hofer of Audio Precision points out in his paper on the measurements of class-D amplifiers, audio amplification circuits can be driven into non-linearity by low levels of RF pollution.

Quote:
Originally Posted by Chu Gai View Post
Quote:
Originally Posted by stereoeditor View Post
I am not saying that there will be audible problems with this preamp in such a situation. However, I feel, as with Paul Miller's tests of HDMI-input receivers that have been quoted elsewhere in this thread, the role of an equipment review is to offer as full a picture of the DUT's behavior as is practicably possible.
Also, what kind of measurement results would suggest there probably would be audible consequences as opposed to unlikely?
Answering that question is beyond the purview of an equipment report. I have untangled your original posting to show what I believe the function of an equipment report should be. My responsibility is to show that the DUT has a potential problem; exploring the probability of this possibility with every combination of ancillary products with which the DUT might be used is impractical in the context of a magazine review.

John Atkinson
Editor, Stereophile

Last edited by stereoeditor; 06-16-2014 at 05:59 AM. Reason: To correct typo
stereoeditor is offline  
post #622 of 2920 Old 06-16-2014, 07:58 AM
AVS Special Member
 
DonH50's Avatar
 
Join Date: Feb 2010
Location: Monument CO
Posts: 6,301
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 230 Post(s)
Liked: 293
I'm with John on this one. I wish more testing would include wideband response and would like to see it carried through at least the AM band and a little beyond. RFI/EMI is insidious, sneaks in all sorts of ways, and when circuits inside the component "see" it they can do all sorts of things with it, including modulating down to the audio band. In my years as a tech dealing with weird sounds (or even worse distortion difficult to detect, like when the 0.01% rated THD+N preamp exhibits 0.05%) caused by RF were among the most difficult to track down (another was leaky capacitors, but that's another topic). Phono inputs were among the worst. And with power supplies switching from hundreds of kHz to several MHz plus a myriad of internal signals from all the digital processing and displays. Yes, displays; they are sometimes the worst offenders because they use fairly slow but large signals that make all the pretty pictures and numbers appear, a big noise source.

On the random jitter issue, I would hazard a guess that nobody on this thread who knows a bit about jitter thinks random jitter is a problem. If it is large enough it will increase the noise floor (reducing SNR and SINAD) but jitter "tones" from deterministic sources are far more detectable. I can't say how detectable. Spurs >80 dB down seem awfully small but I have not performed any sort of recent tests nor taken time to read some of the articles linked (on my list). My career has mostly focused on much higher rate systems, though we have to have that performance all the way to DC in many systems.

I made a rookie mistake when I provided the jitter number by not citing the signal frequency. It is 2 ns at 20 kHz that results in 12-bit performance. For an N-bit converter and sinusoidal signal, the aperture time is 1/[(2^N)*pi*f], and that is the level of jitter that causes an lsb to be uncertain. Note that f is the signal frequency; the clock frequency does not matter. The reasons why are explained over on WBF ( http://www.whatsbestforum.com/showth...322-Jitter-101 ) or numerous texts on the subject. I attached a chart from that thread here. N.B. This is for random jitter.

HTH - Don
Attached Thumbnails
Click image for larger version

Name:	Jit101F2.JPG
Views:	12
Size:	81.3 KB
ID:	120370  

"After silence, that which best expresses the inexpressible, is music" - Aldous Huxley
DonH50 is offline  
post #623 of 2920 Old 06-16-2014, 09:44 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 hevi View Post
Quote:
Originally Posted by arnyk View Post
Both HDMI gear and analog tape recorders may have peak jitter @ 30 Hz. In analog tape machines, the jitter is given in percentage, and 0.05% peak jitter represents really good performance that is generlaly thought to be free of audible consequences.

Please don't use the term "jitter" for analog signals. Jitter is an unwanted time offset in discrete signal transitions.
Are you under the false impression that time offsets are not produced by analog jitter?

Quote:
The correct term to use for the analog domain is phase noise, which is a phase offset that continually changes the timing of the signal.
If you want to play that game then the correct term to use for jitter in the digital domain is digital domain phase noise.

In fact, the word jitter, and the phrase FM Distortion are appropriate for phase noise in both domains.

If you think that only analog equipment can produce FM distortion that sounds like wow and flutter, I would be happy to correct that misapprehension
arnyk is offline  
post #624 of 2920 Old 06-16-2014, 10:39 AM - Thread Starter
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: 546
Quote:
Originally Posted by arnyk View Post
More evidence that denial isn't just a river in Egypt. The above statement fails on grounds that have been discussed here many times before which is proof by authority. History has shown any number of well-known, highly regarded people who have supported ideas that we now know to be completely false.
Good morning Arny . Your posts often put a smile on my face and for that, I thank you.

What I quoted is not an "idea." They are not saying earth is round when folks thought it was flat. They are making an observation based on hundreds of years of combined experience in the audio field. That observation very much matches the only data we have on hand: that of Paul Miller and mine.

You yourself unfortunately have no data to share. Your case is based on "I know audio science and these distortions are not audible." Well, they know audio science a 100 times more than you and I combined . Perhaps you don't know these people. So please allow me to talk more about their backgrounds.

I have already spoken about JJ. Here is what you had to say about him in another thread here:

Quote:
Originally Posted by arnyk View Post
One of the posters at the Hydrogen Audio forum who commented is actually James Johnson, the Bell Labs researcher who was one of the inventors of MP3 and recently retired as head scientist of DBX. http://home.comcast.net/~retired_old_jj He's a walking encyclopedia of these things and if he didn't recognize it, I don't know who would.
Some of what you attribute to him is incorrect but overall assessment is quite correct. I guarantee you that he will lose you on the second sentence on any topic related to audio . I have talked to JJ about HDMI audio and he says idiots designed it. I trust you have not had any interactions with him on the topic at hand (HDMI).

Here are the bios of a couple of other people that you may not be familiar with:

McKinnie holds a Ph.D. from the University of Surrey (U.K.), where his research at the Institute of Sound Recording focused on the influence of spatial envelopment and localization accuracy on the perceived sound quality of surround-sound playback systems.

McKinnie received his master of music degree in sound recording from McGill University, where he engaged in research on techniques for low-bit-rate audio critical listening tests. While at McGill, he assisted in the selection of critical listening materials for the Electronic Industries Association/National Radio Standards Committee, which were used to assess the sound quality of HD radio. This research was carried out at Canada's Communications Research Centre in Ottawa.

He received his bachelor of arts degree in music from Case Western Reserve University.

Since 1994, McKinnie has been the director of live sound operations at Tanglewood, the summer home of the Boston Symphony Orchestra, where he has worked with artists as diverse as James Taylor, Diana Krall, The Boston Pops, the Kalichstein-Laredo-Robinson Trio and Train. He also was the audio engineer for the Cleveland Institute of Music and a staff engineer at Cleveland's Commercial Recording Studios.

His recording credits include compact discs for Telarc and McGill Records, radio production and demonstration recordings for the BBC, as well as countless other radio broadcasts and independent releases.

Dr Jan Berg is an associate professor in audio technology at Luleå University of Technology, Sweden, where he carries out research and education on different aspects of audio production with special focus on sound quality evaluation. In his dissertation (2002) Dr Berg studied evaluation of spatial quality in multichannel systems. Dr Berg’s research activities include collaboration with both academia and industry, e.g. in codec testing, loudness and development of listening tests. Prior to his research work, Dr Berg worked for more than a decade at Swedish Radio as both recording engineer and maintenance engineer. In 2006, Dr Berg chaired Sweden’s first AES international conference, the AES 28th, on Audio Technology in the Future – Surround and Beyond, taking place in Piteå, Sweden. This was followed in 2010 by the 38th conference on sound quality evaluation. For his work Dr Berg received the Board of Governors Award on two separate occasions. Dr Berg is also active within the AES as a reviewer for the Journal in addition to other duties. He previously held the position as Dean of the Faculty of Arts and Social Sciences at Luleå University of Technology.

Bob Schulein
AES Member Profile
Job Title: President
Company: RBS Consultants
Status: Life Fellow
Member since: 1965
AES Position: Vice-Chair Technical Council
AES Committees: Technical Council, Journal Reviewers

Technical Committees: Automotive Audio, Hearing and Hearing Loss Prevention, High Resolution Audio, Microphones and Applications, Perception and Subjective Evaluation of Audio Signals, Spatial Audio, Technical Council
Standards Committees: SC-04-04 (Microphone Measurement and Characterization), SC-04-04-D (Project AES-X42)

Bob has been active in a number of professional organizations within the audio and hearing industries. He is a Fellow and Past President of the International Audio Engineering Society, where he received the Publications Award in 1977, the Board of Governors Award in 1990 and the Bronze Metal in 2013.

He currently serves as a vice-chair of the AES Technical Council, and is chair of the AES Technical Committee on Hearing and Hearing Loss Prevention. Bob is a member the American Auditory Society as well as the American National Standards Institute S3WG48 working group on hearing-aid measurements.


In these forums we create artificial bubbles that have rules that don't have any association with reality. In real life, people's qualifications matter and matter greatly. Fact that someone has created stupid arguments like "appeal from authority" doesn't mean we become stupid just the same and follow them. You know that Arny.

Let's look at our silly rules and how it falls flat on its face. Both of the two experts above are AES Journal reviewers. Don't we put immense value on "peer reviewed" papers? Well, these are the "peers" that did the reviewing. All of a sudden we don't care what these people have to say??? Worse yet, we put our opinion ahead of theirs?

No, this is not proof by authority. It is reporting on informed opinions from a broad group of experts on the status of HDMI when it comes to jitter and audibility.

You keep talking masking this, and masking that. Do you think masking is an unknown concept to JJ? How about the other two experts above? They got through their education, professional work, and that of AES, and don't know who Zwicker is and what is masked and isn't?

Of course they know masking and rest of science of psychoacoustics. Not only from academic point of view of reading the text but through careful listening tests that gives them a first hand feel for what it really means.

And what is our counter to their combined experience? A cut and past of picture or two we have post from Zwicker's book. Heck, we don't even have the book. Said pictures of masking have come from Google Docs page snapshots or tidbits online as we seem to be too cheap to even buy the book we show as bible of our argument.

As if that was not enough, we go on to quote the HDMI wiki as our one and only source of knowledge about this interface! So let's get real and not keep pretending that we are that stupid.

These are experts in the field and have published an opinion under the guise of audio engineering society. That matters and matters a lot no matter how we try to spin it using our made up online rules.

Quote:
Originally Posted by arnyk View Post
It also fails on the grounds that strong assertion cannot make vague statements into specific claims.
Sorry but no. It is not vague at all. They explain the source of the problem specifically and provide their combined opinion as to whether it is or is not an audible problem. Just because we don't like what they are saying we can't claim "it is vague."

Quote:
Originally Posted by arnyk View Post
I fear for the future credibility of anybody who continues to rest their reputation on obviously fallacious arguments.
And I shed a tear for anyone who in real life dismisses the opinion of world class researchers and practicing professionals in some field and instead, believes what random dudes on an internet forum are saying based on tidbits they have read online.

That is the real crime Arny. We talk about believing in "science" yet the first thing we do is dismiss our scientists. We hold the Meyer and Moran paper up high because it is "peer reviewed" and then turn around and claim the people who may have reviewed it are not qualified to make a simple statement about audio fidelity?

Let's not be partisan all the time. Be interested in the science of audio. Not some religion around it that causes you to deny useful information in front of us just because we hold some extreme position in this field that may be invalidated this way.
stereoeditor likes this.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"

Last edited by amirm; 06-16-2014 at 10:42 AM.
amirm is offline  
post #625 of 2920 Old 06-16-2014, 10:54 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 amirm View Post

Quote:
Originally Posted by arnyk View Post
I fear for the future credibility of anybody who continues to rest their reputation on obviously fallacious arguments.

And I shed a tear for anyone who in real life dismisses the opinion of world class researchers and practicing professionals in some field and instead, believes what random dudes on an internet forum are saying based on tidbits they have read online.
Excluded middle argument noted.

Believe it or not, there is a bit of middle ground between accepting what people who may be authoritative have said as being the unassailable truth and thinking about it critically and letting the pieces fall where they fall.

In this case, there is no evidence that even these people agree with how their statement has been utterly distorted and twisted out of shape and misrepresented here.

So, we're talking about stacking fiction on top of fiction.
arnyk is offline  
post #626 of 2920 Old 06-16-2014, 11:13 AM
Senior Member
 
hevi's Avatar
 
Join Date: Jan 2007
Posts: 365
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 33 Post(s)
Liked: 48
Quote:
Originally Posted by arnyk View Post

Are you under the false impression that time offsets are not produced by analog jitter?
I am very well aware of what jitter is and what phase noise is, and how they relate to each other. The correct term for analog systems is phase noise.


Quote:
Originally Posted by arnyk View Post
If you want to play that game then the correct term to use for jitter in the digital domain is digital domain phase noise.
The correct term for "jitter" is "jitter". There's no need to qualify it with "in the digital domain" because that is already in its very definition.

Quote:
Originally Posted by arnyk View Post
If you think that only analog equipment can produce FM distortion that sounds like wow and flutter, I would be happy to correct that misapprehension

I am just trying to help you use the correct terminology.
stereoeditor likes this.


Last edited by hevi; 06-16-2014 at 12:41 PM. Reason: added lost word
hevi is offline  
post #627 of 2920 Old 06-16-2014, 11:35 AM - Thread Starter
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: 546
Quote:
Originally Posted by arnyk View Post
Quote:
Originally Posted by amirm View Post

Quote:
Originally Posted by arnyk View Post
I fear for the future credibility of anybody who continues to rest their reputation on obviously fallacious arguments.

And I shed a tear for anyone who in real life dismisses the opinion of world class researchers and practicing professionals in some field and instead, believes what random dudes on an internet forum are saying based on tidbits they have read online.
Excluded middle argument noted.
Back to bubble rules again. Please Arny, use them with folks who take it seriously.

Quote:
Originally Posted by arnyk View Post
Believe it or not, there is a bit of middle ground between accepting what people who may be authoritative have said as being the unassailable truth and thinking about it critically and letting the pieces fall where they fall.

In this case, there is no evidence that even these people agree with how their statement has been utterly distorted and twisted out of shape and misrepresented here.
I have quoted their statement in its entirety and with no changes whatsoever. Here it is again:

HDMI, the point-to-point connector required for BR and HD video, has excellent bandwidth and an Ethernet data link (HDMI 1.4), but lacks an audio clock. HDMI receivers must derive audio word clock from the video pixel clock, commonly resulting in very high jitter that affects quality and can be audible. Some high end receivers address the jitter and many companies are researching it but current solutions are expensive and uncommon.

I present their statement exactly as it is. That HDMI has no audio clock. We know that is true. That HDMI has to extract the audio clock from video. We know that is true. They then present their informed opinion that these two factors result in "very high jitter" and "can be audible."

You can't push this aside with debating terms Arny. They are providing foundation and their combined opinion and experience on what is wrong with HDMI.

Here is you playing the role of above experts in precisely the same manner Arny:

Quote:
Originally Posted by arnyk View Post
One reason that HDMI jitter tends to be numerically greater is that HDMI audio is transmitted in widely separated possibly varying and therefore somewhat intermittent fairly long packets, which makes the job of the buffer on the receiving end that much harder. In a SP/DIF connection, the audio information is transmitted in small packets that follow each other closely with consistent timiing.
See? Just like them you present a cause and effect. The cause being "widely separated" audio packets. The effect being "numerically greater" jitter.

How come your explanation doesn't match theirs? Why didn't they mention that cause as you are? Isn't the answer that your explanation of HDMI is incorrect? And aren't you asking us to accept you as "authority" with that paragraph while at the same time, telling us to disown this highly respected group at Audio Engineering Society?

While we all love you and cherish you as our god of audio, you have to admit that in real world your credentials in this matter (HDMI) would not hold to even one person on that panel let alone all of them combined.

Net, net, I think this is a battle that is already lost Arny. People who know a lot more you have declared your position as not being correct. No high-school debating term is going to undo this situation. Your one opinion can't and won't override theirs I am afraid. Let's try to regain any credibility we have left by not continuing to deny what is in front of us. As you say, Denial is not just a river in Egypt.
stereoeditor likes this.

Amir
Retired Technology Insider
Founder, Madrona Digital
"Insist on Quality Engineering"
amirm is offline  
post #628 of 2920 Old 06-16-2014, 01:18 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 hevi View Post
Quote:
Originally Posted by arnyk View Post

Are you under the false impression that time offsets are not produced by analog jitter?
I am very well aware of what jitter is and what phase noise is, and how they relate to each other. The correct term for analog systems is phase noise.
Incorrect and false distinction noted. the correct name for phase noise is phase noise, no matter what kind of system it occurs in.

I think you may need to apply google to your research and see how the rest of the world sees things.

It's an audiophile myth that phase noise isn't phase noise when it occurs in digital system. By enforcing invalid and false distinctions, people are mislead into believing that analog systems are somehow immune to jitter, when in fact they are generally more severely afflicted by jitter than digital systems.

How else to explain the immense denial Amir seems to be in about the relatively immense jitter in his Otari MX 5050 or for that part, his speaker systems?
arnyk is offline  
post #629 of 2920 Old 06-16-2014, 03:14 PM
Senior Member
 
hevi's Avatar
 
Join Date: Jan 2007
Posts: 365
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 33 Post(s)
Liked: 48
Quote:
Originally Posted by arnyk View Post
Quote:
Originally Posted by hevi View Post
Quote:
Originally Posted by arnyk View Post

Are you under the false impression that time offsets are not produced by analog jitter?
I am very well aware of what jitter is and what phase noise is, and how they relate to each other. The correct term for analog systems is phase noise.
Incorrect and false distinction noted. the correct name for phase noise is phase noise, no matter what kind of system it occurs in.

I think you may need to apply google to your research and see how the rest of the world sees things.
Um, funny you should mention Google -however, before giving other people that advice, you might want to try Googling your home brew term "analog jitter" first, and compare the number of hits to "phase noise". I haven't bothered myself but I am pretty sure what the result will be...


Quote:
Originally Posted by arnyk View Post
It's an audiophile myth that phase noise isn't phase noise when it occurs in digital system. By enforcing invalid and false distinctions, people are mislead into believing that analog systems are somehow immune to jitter, when in fact they are generally more severely afflicted by jitter than digital systems.
And there you go again, "analog jitter". In analog domain the correct term is still phase noise.

In the field of mixed signals, jitter is, if not defined as, so at least extremely strongly associated with fluctuations in the times of zero-crossings of a (typically) nominally DC-free periodic waveform.

While phase noise is a continuous function (i.e. φ(t)), jitter, time deviations between zero crossings, comprise a set of discrete values (i.e {τi}). One is a continuous function, the other one is discrete. The are related, but they are not the same thing.

hevi is offline  
post #630 of 2920 Old 06-16-2014, 03:57 PM
AVS Special Member
 
CharlesJ's Avatar
 
Join Date: Jan 2006
Posts: 3,408
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 188 Post(s)
Liked: 117
[quote=amirm;24998218]
Quote:
Originally Posted by UndersAVS View Post
...

Committee Members

Code:
 Bob Katz      Douglas McKinnie      Duane Wise 
 James Johnston      Jan Berg      Robert Schulein 
 Steven Harris      Thomas Sporer      Vicki R. Melchior 
 John Grant      Michael Kelly      Josh Reiss 
 Juergen Peissig      James J McTigue      David Rich 
 Sergio Liberman      Hans van Maanen      John Dawson 
 Bryan Martin      Oles Protsidym      Dylan-Thomas Vance 
 Mark Waldrep      Peter Craven      Partev Sarkissian 
 Ryan Stables      Michael Hudson      Yonghao Wang 
 David Jones
James Johnston is "JJ" who used to be an audio architecture in my group at Microsoft and whose name is mentioned quite often. He is in our (objectivity) camp.

John Dawson is a friend and the founder of Arcam, the company whose presentation I quoted earlier on Paul Miller measurements. He is also a frequent member here. He actually knows how to design the circuits we talk about here.

Vicki Melchior is the co-chair of the committee. If I remember right, she was part of the Boston Audio Society -- same group that brought us the Meyer and Moran test!

...?
So, none of these members, least of all as a technical group committee are infallible and must be accepted as beyond reproach???


I sat in on such a committee once as a visitor and didn't get that feeling at all. Like any other committee, back and forth of ideas and more or less come to some agreement, regardless of facts in evidence.
CharlesJ is offline  
Closed Thread Audio Theory, Setup, and Chat

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