Optical cable or hdmi? - Page 4 - AVS Forum
Forum Jump: 
Closed Thread
 
Thread Tools
post #91 of 144 Old 05-06-2014, 08:45 PM
AVS Special Member
 
David Susilo's Avatar
 
Join Date: Aug 2002
Location: Markham, Canada
Posts: 9,628
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 311 Post(s)
Liked: 417
To me, it's very simple in terms whom to trust LESS (or not to be trusted at all)

1. A certain individual, when you google his name, the list contains near 100% complaints regardless whichever forum you go
2. Can't spell 20+ year old popular technical term correctly
3. Making up terms as he goes along
4. Indirectly saying ALL musicologists are wrong
5. Can't differentiate between digital vs analog distortion
6. Constantly contradicting himself

Usually in baseball, 3 strikes and you're out. This person struck all 6 times (and repeatedly too) and he still portray himself as the king of the hill? What a joke!
amirm and Riffmeister like this.

follow my A/V tweets @davidsusilo

ISF, THX, CEDIA, Control4 & HAA certified
Reviewer for TED, QAV, AUVI & DownUnder Audio Magazine


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

David Susilo is offline  
Sponsored Links
Advertisement
 
post #92 of 144 Old 05-07-2014, 05:00 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by David Susilo View Post

To me, it's very simple in terms whom to trust LESS (or not to be trusted at all)

1. A certain individual, when you google his name, the list contains near 100% complaints regardless whichever forum you go

A real concern if you think that science is determined by popularity contests.
Quote:
2. Can't spell 20+ year old popular technical term correctly

A real concern if you think that science is determined by spelling bees.
Quote:
3. Making up terms as he goes along

Proof?
Quote:
4. Indirectly saying ALL musicologists are wrong

Never happened. What I did say is IME that not many musicologists are also deeply involved in audio technology. If you have this much time on your hands, you might want to disprove that statement.

BTW when did musicologists become a protected species that are always infallible? ;-)
Quote:
5. Can't differentiate between digital vs analog distortion

Never happened.
Quote:
6. Constantly contradicting himself

Never happened.

BTW I don't know who you are talking about. The point is that you can use whatever criteria you want, but it would help you if you stuck to logic, reason and verifiable facts.

BTW it doesn't matter who is "The king of the Hill". It should be about who presents audio science and art the most accurate and reasonable way.

Too bad that it seems so strongly that you are nursing all these pent up hurts in public.
arnyk is offline  
post #93 of 144 Old 05-07-2014, 05:09 AM
FMW
AVS Special Member
 
FMW's Avatar
 
Join Date: Jun 2007
Posts: 4,955
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 313 Post(s)
Liked: 725
Quote:
Originally Posted by David Susilo View Post

To me, it's very simple in terms whom to trust LESS (or not to be trusted at all)

1. A certain individual, when you google his name, the list contains near 100% complaints regardless whichever forum you go
2. Can't spell 20+ year old popular technical term correctly
3. Making up terms as he goes along
4. Indirectly saying ALL musicologists are wrong
5. Can't differentiate between digital vs analog distortion
6. Constantly contradicting himself

Usually in baseball, 3 strikes and you're out. This person struck all 6 times (and repeatedly too) and he still portray himself as the king of the hill? What a joke!

Debating what someone says is far more effective than engaging in personal attacks. The personal attacks simply say that you have no effective counter so you have fallen into desperation, for some unexplained reason. I congratulate Arny for taking the slings and arrows with such aplomb.
FMW is offline  
post #94 of 144 Old 05-07-2014, 06:19 AM
AVS Special Member
 
David Susilo's Avatar
 
Join Date: Aug 2002
Location: Markham, Canada
Posts: 9,628
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 311 Post(s)
Liked: 417
Quote:
Originally Posted by arnyk View Post

A real concern if you think that science is determined by popularity contests.
A real concern if you think that science is determined by spelling bees.
Proof?
Never happened. What I did say is IME that not many musicologists are also deeply involved in audio technology. If you have this much time on your hands, you might want to disprove that statement.

BTW when did musicologists become a protected species that are always infallible? ;-)
Never happened.
Never happened.

BTW I don't know who you are talking about. The point is that you can use whatever criteria you want, but it would help you if you stuck to logic, reason and verifiable facts.

BTW it doesn't matter who is "The king of the Hill". It should be about who presents audio science and art the most accurate and reasonable way.

Too bad that it seems so strongly that you are nursing all these pent up hurts in public.

1. It is not about popularity contest, but it's been stated on various forums that your posts don't reflect scientific approach including one proof that you were caught fudging an ABX test.
2. Spelling is important when writing scientific claim. For example: "sprinkle on water" means completely different from "sprinkles on walter"
3. You started coining "analog jitter"
4. There was no "IME" stated when you claim that "vocal vibrato is a form of analog jitter"
5. See point 4. You didn't say "IME", but now that you've been proven wrong, you backtrack by adding "IME"
6. See item 1-5, you didn't "present audio science and art the most accurate and reasonable way" at all.
Riffmeister likes this.

follow my A/V tweets @davidsusilo

ISF, THX, CEDIA, Control4 & HAA certified
Reviewer for TED, QAV, AUVI & DownUnder Audio Magazine


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

David Susilo is offline  
post #95 of 144 Old 05-07-2014, 08:41 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by arnyk View Post

It may surprise you Amir but my bowel movements are very important to me. They keep me alive like everybody else. But, I don't need to provide published or peer-reviewed documentation of my bowel movements to be credible when I say they exist. I don't even have to say they exist! Much of the audio knowledge I have acquired in my life are about as consequential to me at the time as bowel movements. Guess what, they aren't documented!
Good morning Arny! Thank you so much for the detailed response. Much appreciated.

Your analogy here put a smile on my face although if I were any smarter, I would have read it when I was not eating breakfast. eek.gifsmile.gif

It reminds of an interview with Ted Turner, founder of CNN and TBS networks on TV where he said, "I am no different than anyone else. I get diarrhea just like everyone." That and two billion dollars would equate him to anyone else. biggrin.gif

By the same token, don't you think that referencing your own double blind tests that showed there were audible differences in amplifier over your 20-30 years of having these arguments would be important? You and I have been in a number of threads where I cited LTS blind tests that showed most amplifiers color the sound when testing their input vs output. You were adamant that their tests must have been faulty because the outcome was not believable in your mind. We could have certainly used the information in your amplifier double blind tests to show that they are actually in agreement with the one and only documented test of amplifier that you have been kind enough to participate in.

Isn't this like a trial where the prosecutor has withheld material evidence that would show the accused to be potentially innocent? How would that prosecutor look when the evidence comes to surface? Say that there was no need to show such evidence because it would be like documenting his own bowel movements?
Quote:
No, its this idea that there are laws for me that don't seem to apply to anybody else so that I am not even allowed to misplace a slash without there being a Federal Case.
Sorry about that. Ratman likes to pull your tail on that and I thought we help him along with that tease smile.gif. Sorry if it went too far.
Quote:
Umm, the quote is up there Amir. It isn't about amplifiers it is about digital connections. Have you lost your mind?
I often misplace my mind. I try to remember to always put it in the same drawer but once in a while I forget and leave it somewhere else. Naturally without my mind, I am helpless in finding it and have to ask my wife to search for it.

But yes, with or without my mind I understand OP asked about interconnects. But you chimed in and said:
Quote:
Originally Posted by arnyk View Post


Where is that research? The research into audible differences in digital connections or amps operating within their linear ranges has the same probability of success as research disproving Newtonian physics to greater degree than Einstein did in 1918.

I thought it was very odd to say there is no research into audible differences between amps where you have fully documented an instance of this yourself. It seems we are still operating as if that evidence had not come to surface. Now that it has, I hope you agree that we can no longer say what you said above regarding amplifiers.
Quote:
And as far as all amplifiers sounding the same - how many times have I corrected that misapprehension? Heck I even got an article published about it in an internationally known magazine!
You have said that. Problem is you also dispute any notion of anyone hearing differences between amplifiers as being biased testing. And have never provided any details regarding that test. I think it is material for members here to know you have found such differences and hence we should not so quickly rush to judgment that anyone with similar outcome must be wrong.
Quote:
Someone said something about people with too much time on their hands, right?

Yup!
I am OK with that accusation. I do have the time. And I enjoy such interactions as there is opportunity to learn something from experts like you. Without these discussions every question would end with: "buy monoprice cable." biggrin.gif I don't take that comment as a negative at all.

Don't let people dissuade you Arny from sharing your vast insight and knowledge with the rest of us. We are here because we have chosen to be. Anyone who says they are not having too much time yet reading and posting in this thread have next to zero self-awareness and are just adding noise to an otherwise technical discussion.

Kind regards,
Amir
Riffmeister likes this.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #96 of 144 Old 05-07-2014, 09:04 AM
AVS Addicted Member
 
Ratman's Avatar
 
Join Date: May 2002
Location: Collingswood, N.J.
Posts: 14,494
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 289
Quote:
Originally Posted by amirm View Post

Ratman likes to pull your tail on that and I thought we help him along with that tease smile.gif. Sorry if it went too far.
Thank you for the considerate reference in your typical professional manner. I merely would like to follow in the footsteps of you and Arny to continue maintaining the accuracy with each individual's post in every thread.

Respectfully yours,
Ratman
Ratman is online now  
post #97 of 144 Old 05-07-2014, 10:51 AM
Member
 
Riffmeister's Avatar
 
Join Date: Jul 2003
Location: Indianapolis
Posts: 194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 34
Quote:
Originally Posted by spkr View Post

Pardon me sir, may I ask what your background is in electronics and audio? Being a relatively new member, I would like to be clear on who is who around here. Thanks in advance.
This would be a really good question to address to Arny Krueger, since he is so highly regarded by many noobs such as yourself. What exactly is Arny's professional background in electronics and audio? We know he invented double blind testing some time back in the early 1900s, but beyond that not much is known. He has refrained from answering questions in this regard before, perhaps out of sheer modesty.
Riffmeister is online now  
post #98 of 144 Old 05-07-2014, 11:13 AM
AVS Addicted Member
 
Ratman's Avatar
 
Join Date: May 2002
Location: Collingswood, N.J.
Posts: 14,494
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 186 Post(s)
Liked: 289
Arny has provided that info in another thread. Search around... you'll find it.
Ratman is online now  
post #99 of 144 Old 05-07-2014, 11:37 AM
Member
 
morphemes's Avatar
 
Join Date: Mar 2014
Posts: 36
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10

TOSLINK looses lossless TrueHD or Master Audio.  I don't have anything that outputs to both toslink & HDMI, otherwise I'd do some measurements.  Unless your making measurements from an audio device which outputs to both HDMI & TOSLINK, there really isn't any use in making measurements. all you're really doing is comparing each particular audio device, not the connections themselves.  As far as I'm concerned with optical, you are limited to "lossy" surround formats (not that it makes any audible difference).  HDMI means fewer cables & lossless TrueHD/Master Audio streams.

HDMI

morphemes is offline  
post #100 of 144 Old 05-07-2014, 12:30 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by morphemes View Post

TOSLINK looses lossless TrueHD or Master Audio.  I don't have anything that outputs to both toslink & HDMI, otherwise I'd do some measurements.  Unless your making measurements from an audio device which outputs to both HDMI & TOSLINK, there really isn't any use in making measurements. all you're really doing is comparing each particular audio device, not the connections themselves.  As far as I'm concerned with optical, you are limited to "lossy" surround formats (not that it makes any audible difference).  HDMI means fewer cables & lossless TrueHD/Master Audio streams.

HDMI
All the measurements I showed earlier was from one device. TrueHD and DTS MA are issues for watching movies and BD concerts. For listening to music, that is not a limitation.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #101 of 144 Old 05-07-2014, 08:06 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 amirm View Post

In absence of such listening tests (and I know of none anywhere else), I am guided by my knowledge of the implementation of these standards and unfortunately the worse HDMI measured performance of all the AVRs I have tested relative to SPDIF. Here is the comparison of an Onkyo AVR comparing HDMI vs S/PDIF:

i-tn5rJfr-X2.png
An ideal implementation would have just our main tone in the middle at 12 Khz sharply defined. As we see unfortunately HDMI graph in red shows considerably more distortion spikes than SPDIF. Given the fact that I found the same results on all but one high-end implementation of HDMI (the $30,000 Mark Levinson), I think we have no choice but to agree that measured performance of HDMI is worse than SPDIF. Given the choice of both interfaces, why would I want to listen using the one that has worse measured performance? At best that extra distortions are not audible. At worst, they are. In neither case does HDMI work better whereas SPDIF does. This is behind my recommendation to OP.

The hdmi distortion peaks are -90dB. Is that audible? Do the decibels on the chart translate directly to room SPL measures?
UndersAVS is offline  
post #102 of 144 Old 05-08-2014, 05:18 AM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 3,958
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 95 Post(s)
Liked: 154
Quote:
Originally Posted by UndersAVS View Post

The hdmi distortion peaks are -90dB. Is that audible? Do the decibels on the chart translate directly to room SPL measures?

the spectrum shown is for line level signals. The top of the peak, not shown, would just touch zero line, presumably. Lopping off the top third of the chart makes it easier to see what's happening at low levels but also has the unfortunte byproduct of making the low level stuff look relatively bigger than it is.

Anyway, how loud that particular tone would be in your room depends on the output of the device reproducing it and one's volume setting.

But assuming no compression, the spurious sounds will come from the device at 90 dB below the main tone. Assuming linear amplification and preamplification (no clipping, no compression) the signal keeps its spectral content ll the way through the amplification chain, so the spurious tones are 90 dB below whatever level the main tone is in your room (assuming no compression in the speaker). So if the min tone is 105 dB in your room, the spurious tones are at 15 dB, well below the noise floor in my house.
JHAz is offline  
post #103 of 144 Old 05-08-2014, 05:22 AM
Member
 
morphemes's Avatar
 
Join Date: Mar 2014
Posts: 36
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10

I keep seeing this graph comparing "S/PDIF" with  "HDMI"

Is this comparing  the same "S/PDIF" bitstream over both HDMI & TOSLINK?

Because HDMI isn't an audio signal, it's a cable & S/PDIF isn't a cable, it is a signal.
the HDMI plot looks kind of... well.. analog?
The "distortions" look like... I don't know the right technical terms...  probably something to do with the compression/decompression of the signal being sent over the HDMI (likely the exact cause)  That is my guess.  It's too uniform...  too fractal...  the signal would have to be analog for it to display that kind of pattern & it to have been caused by the signal path...
 

morphemes is offline  
post #104 of 144 Old 05-08-2014, 07:20 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by morphemes View Post

I keep seeing this graph comparing "S/PDIF" with  "HDMI"

Is this comparing  the same "S/PDIF" bitstream over both HDMI & TOSLINK? Because HDMI isn't an audio signal, it's a cable & S/PDIF isn't a cable, it is a signal.

the HDMI plot looks kind of... well.. analog? The "distortions" look like... I don't know the right technical terms...  probably something to do with the compression/decompression of the signal being sent over the HDMI (likely the exact cause)  That is my guess.  It's too uniform...  too fractal...  the signal would have to be analog for it to display that kind of pattern & it to have been caused by the signal path...
Sorry about not explaining the background better. The graphs are from an article I wrote for Widescreen Review Magazine that I have not yet had time to put online (if you are a subscriber you can download it from their site).

Briefly the measurements are from the "line out" RCA jacks of the AVR. In other words, it is the analog signal that would have been amplified to go to your speakers. The source signal that is being played is identical digital file in both cases. The only difference is that the source is either connected using HDMI or S/PDIF.

In other words, we are not looking at the waveforms that exist over the wire. We are looking at the final outcome that comes out of the receiver. The reason we do it this way is that the AVR has filtering that cleans up some of the distortions that occur on the wire. So looking at what is on the wire would be misleading and would show things to be worse that they really are.

In summary, yes, you are looking at analog measurements and hence you are going to see lots of variations, noise, etc. in the waveforms as you would when operating the AVR yourself. There is no compression/decompression as the original content was uncompressed being transmitted over two digital interfaces to the AVR unchanged. In other words, the same digital samples were received by the AVR and told to convert them to analog. The analog differences that you see are due to a number of reasons such as the DAC clock or reference voltage and bleeding of system noise intos its analog outputs, e.g. through power supply rails (connections).

The differences show that just because we are using a digital interface, it does not mean that we get the same analog output from the DAC. The audio samples remain the same but the way the DAC converts them into analog varies.
morphemes likes this.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #105 of 144 Old 05-08-2014, 07:28 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by JHAz View Post

the spectrum shown is for line level signals. The top of the peak, not shown, would just touch zero line, presumably. Lopping off the top third of the chart makes it easier to see what's happening at low levels but also has the unfortunte byproduct of making the low level stuff look relatively bigger than it is.

Anyway, how loud that particular tone would be in your room depends on the output of the device reproducing it and one's volume setting.

But assuming no compression, the spurious sounds will come from the device at 90 dB below the main tone. Assuming linear amplification and preamplification (no clipping, no compression) the signal keeps its spectral content ll the way through the amplification chain, so the spurious tones are 90 dB below whatever level the main tone is in your room (assuming no compression in the speaker). So if the min tone is 105 dB in your room, the spurious tones are at 15 dB, well below the noise floor in my house.
Good explanation except the last sentence. The noise in your room has a spectrum. It is not a flat spectrum. Higher frequencies get filtered much more easily from intruding into your room than low frequencies. Here is an example graph from an article I wrote on actual noise and dynamic range in listening spaces, http://www.madronadigital.com/Library/RoomDynamicRange.html:

FielderRoomNoiseFloor.gif

This is a survey of many listening spaces and shows the min, maximum and average of those rooms. The best rooms go below the threshold of listening in mid-frequencies. Even the average dedicated rooms do pretty well there.

Importantly we can hear signals below the noise floor. So even if you did have noise at the same level as the distortion levels, you can easily hear it. I don't recall if I cited that in my article but if you read the references, we see an interesting phenomena in addition to that whereby the brain is able to hear tones much better because they arrive from a specific speaker compared to noise that is there from all directions.

So we really can't go there in determining audibility. Even if we could, I can show you AVR measurements that are much worse in amplitude than this one.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #106 of 144 Old 05-08-2014, 07:37 AM
Member
 
Riffmeister's Avatar
 
Join Date: Jul 2003
Location: Indianapolis
Posts: 194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 34
Quote:
Originally Posted by morphemes View Post

I keep seeing this graph comparing "S/PDIF" with  "HDMI"


Is this comparing  the same "S/PDIF" bitstream over both HDMI & TOSLINK?


Because HDMI isn't an audio signal, it's a cable & S/PDIF isn't a cable, it is a signal.

S/PDIF is not a signal, it is a connection format, as is HDMI. In each, there are standards for both the physical connection and the signals that pass through them. Both S/PDIF and HDMI pass PCM and compressed multichannel audio. HDMI has the added ability to pass "lossless" multichannel audio. On paper, HDMI should work as well as S/PDIF, in reality, it is otten not implemented as well as S/DIF, as you can see in the graphs.
amirm likes this.
Riffmeister is online now  
post #107 of 144 Old 05-08-2014, 07:45 AM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 3,958
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 95 Post(s)
Liked: 154
Quote:
Originally Posted by amirm View Post

Good explanation except the last sentence. The noise in your room has a spectrum. It is not a flat spectrum. Higher frequencies get filtered much more easily from intruding into your room than low frequencies. Here is an example graph from an article I wrote on actual noise and dynamic range in listening spaces, http://www.madronadigital.com/Library/RoomDynamicRange.html:

FielderRoomNoiseFloor.gif

This is a survey of many listening spaces and shows the min, maximum and average of those rooms. The best rooms go below the threshold of listening in mid-frequencies. Even the average dedicated rooms do pretty well there.

Importantly we can hear signals below the noise floor. So even if you did have noise at the same level as the distortion levels, you can easily hear it. I don't recall if I cited that in my article but if you read the references, we see an interesting phenomena in addition to that whereby the brain is able to hear tones much better because they arrive from a specific speaker compared to noise that is there from all directions.

So we really can't go there in determining audibility. Even if we could, I can show you AVR measurements that are much worse in amplitude than this one.

wait until you connect it to a transducer.
JHAz is offline  
post #108 of 144 Old 05-08-2014, 07:53 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by JHAz View Post

wait until you connect it to a transducer.
??? The measurements of how well we hear below noise floor was done with the "transducer" (speaker). Here is the quote from Fielder's follow on paper that I referenced in my article:

In the author’s original study on dynamic-range requirements, it was determined that the listener was able to perceive audio system noises as much as 15 db below that of the listening-room noise for monophonic noise sources.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #109 of 144 Old 05-08-2014, 09:04 AM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 3,958
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 95 Post(s)
Liked: 154
Quote:
Originally Posted by amirm View Post

??? The measurements of how well we hear below noise floor was done with the "transducer" (speaker). Here is the quote from Fielder's follow on paper that I referenced in my article:

In the author’s original study on dynamic-range requirements, it was determined that the listener was able to perceive audio system noises as much as 15 db below that of the listening-room noise for monophonic noise sources.

i would have thought you were aware that other factors affect audibility. I didn't think I needed to put up every freaking caveat to every freaking point every time I type on here. I'll call you bosso, junior.

The THD of a -90dB spurious signal is .0031623 percent. Even the hyper -subjectivists at Stereophile use 1 percent distortion as the threshold of "clipping," for electronics except when they relax it to 3% for some tube amps to give their manufacturers a little room to save face. I am sure a person could design a signal, say with the fundamental at 100 Hz and all the distortion in the 1000-4000Hz range, and play it loudly enough to perhaps, perhaps make -90dB audible. But in the real world? Especially given that whatever your electronics are doing the transducers to which I referred are cranking out what 10 to 100 times the distortion? Masking effects, etc. etc. etc.

I do not care if I COULD HAVE MAYBE heard the spurious frequencies if only the fundamental was not present or was reduced by 40 or 50 or 60 or 90 dB. I care if I can hear the distortion in a real context. The rest is of great theoretical interest to me but impacts my enjoyment precisely zero.

Sorry I did not make that, which I thought self-evident to folks who care about and have studied so clear that a moderately articulate hobo could comprehend that it was part of the "thang"

And while I remain somewhat subjectively open to the whole IM distortion and addition of distortion stuff (I am a guitar player after all, and that's where I live t least 50% of the time - preamp distortion + power amp distortion and maybe a little somethin'-somethin' from a distortion pedal to put it just a touch farther over the top) in the end with music reproduction and the distortion levels we're talking about, it just doesn't hold water.

Here's why. If the IM or additive distortion caused by low level (ie probably inaudible in the context of the signal) added up to audible distortion, then get this: MP3 REMOVES all those low level things that are otherwise masked. So high bitrate MP3s should, by all rights, sound BETTER than the original because distortions that would have accumulated via the interaction of the otherwise inaudible parts of the signal are gone, onaccounta the inaudible parts of the signal have been scraped off by the lossy compression. AFAIK, NOBODY in their right mind says MP3s sound better than the original. At best, at BEST, they are indistinguishable (which is fine). Of course, to caveat the heck out of it so you don't raise spurious signals, as you reduce bitrate, the encoding system introduces distortions (not THD, but changes) that certainly will become audible at some point. Rolling off the highs. Taking the high frequencies to mono. Those are, IMO, the easy-to-understand and not difficult to hear "atrocities" of too much data compression. I'm not talking about that. I'm saying if you remove low low level signals that otherwise (one imagines) should add to each other and create distortions, but you cannot hear the distortions, then the removal of those low low level signals didn't reduce distortion. IMO, if it ain't audible in a proper double blind test it ain't audible And IMO there's way too much money at stake for it to be the case that nobody has gone looking for those circumstances where, in fact, audible differences at a statistically significant can be found among some segment of the population in properly conducted double blind tests. The non-existence of such published data suggests the non-existence of the phenomenon, because the research surely would be funded. Again, I don't care if I might be able to hear a sound, reproduced by itself, pristinely, at 5 dB in my room, if the sound, in the context in which it occurs is inaudible because there's other stuff going on.

Rock on.
JHAz is offline  
post #110 of 144 Old 05-08-2014, 11:04 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 JHAz View Post

the spectrum shown is for line level signals. The top of the peak, not shown, would just touch zero line, presumably. Lopping off the top third of the chart makes it easier to see what's happening at low levels but also has the unfortunte byproduct of making the low level stuff look relatively bigger than it is.

Anyway, how loud that particular tone would be in your room depends on the output of the device reproducing it and one's volume setting.

But assuming no compression, the spurious sounds will come from the device at 90 dB below the main tone. Assuming linear amplification and preamplification (no clipping, no compression) the signal keeps its spectral content ll the way through the amplification chain, so the spurious tones are 90 dB below whatever level the main tone is in your room (assuming no compression in the speaker). So if the min tone is 105 dB in your room, the spurious tones are at 15 dB, well below the noise floor in my house.

Thats what I suspected. Thanks for explanation. At reference volume level, 105db, the ugliest distortion generated by the hdmi conversion results in 15dB.

So to put this into real world context using the SPL thermometer posted below, I should ask myself: "can I hear someone whisper while simultaneously operating a chainsaw?" Seems like the noise of the chainsaw would mask the sound of a whisper; or is my reasoning out of line?



http://www.dangerousdecibels.org/virtualexhibit/6measuringsound.html
UndersAVS is offline  
post #111 of 144 Old 05-08-2014, 12:03 PM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 3,958
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 95 Post(s)
Liked: 154
Quote:
Originally Posted by UndersAVS View Post

Thats what I suspected. Thanks for explanation. At reference volume level, 105db, the ugliest distortion generated by the hdmi conversion results in 15dB.

So to put this into real world context using the SPL thermometer posted below, I should ask myself: "can I hear someone whisper while simultaneously operating a chainsaw?" Seems like the noise of the chainsaw would mask the sound of a whisper; or is my reasoning out of line?



http://www.dangerousdecibels.org/virtualexhibit/6measuringsound.html

that's kind of how I see it. Theoretically, at least, if the thing making the chainsaw-loud noise has little or no sound in the octave or couple of octaves (I'm no expert on masking) around the "distortion" the distortion may be audible at far lower levels than it would in a more normal context . . . . I'm not saying it's impossible to hear such distortion. It may be. But I'm far closer to pretty dang sure that you'd never her such distortion sidebands within the same octave as the primary signal, or (if it were typical distortion of amps and/or speakers) in the context of sounds that already contain not only much louder fundamentals, but likely contain harmonics in the general range of the distortion that are at least equal to, if not 10-40 dB above the distortion (assuming distortion at the edge of the harmonic envelope). Play a low E on an electric bass. The fundamental at (about) 40 Hz is somewhere round 10 dB LOWER, coming off the pickups, than the second harmonic t 80 Hz. Harmonic series continues, (mostly) reducing in SPL, at least to say 640 Hz (higher harmonics are at far lower levels, although I'm pretty sure you'd miss them if you rolled them off . . . . see http://www.stereophile.com/reference/338/index.html. Even clear out at about 1000 Hz, the harmonic is at about 60 dB below the fundamental, or 50 dB below H2. If it were distortion, that harmonic alone would be 0.1% of the fundamental, or just over .3% of the second harmonic. So 300 percent higher [edit YIKES! 300 times, not 300 percent. That would be 300,000 percent) "distortion" than the -90dB sidebands . . .There was a really cool thread on a site for bass players, talkbass.com, where many different players got broadly similar results from a wide range of instruments. (To get the fundamental close to the second harmonic requires an upright bass, an electric upright bass, or a really esoteric design like the few that have an extra "arm" connecting the body of the bass to the headstock . . . esoteric materials just don't get ya there . . .)
JHAz is offline  
post #112 of 144 Old 05-08-2014, 12:27 PM
Member
 
morphemes's Avatar
 
Join Date: Mar 2014
Posts: 36
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
Quote:
Originally Posted by amirm View Post


Sorry about not explaining the background better. The graphs are from an article I wrote for Widescreen Review Magazine that I have not yet had time to put online (if you are a subscriber you can download it from their site).

Briefly the measurements are from the "line out" RCA jacks of the AVR. In other words, it is the analog signal that would have been amplified to go to your speakers. The source signal that is being played is identical digital file in both cases. The only difference is that the source is either connected using HDMI or S/PDIF.

In other words, we are not looking at the waveforms that exist over the wire. We are looking at the final outcome that comes out of the receiver. The reason we do it this way is that the AVR has filtering that cleans up some of the distortions that occur on the wire. So looking at what is on the wire would be misleading and would show things to be worse that they really are.

In summary, yes, you are looking at analog measurements and hence you are going to see lots of variations, noise, etc. in the waveforms as you would when operating the AVR yourself. There is no compression/decompression as the original content was uncompressed being transmitted over two digital interfaces to the AVR unchanged. In other words, the same digital samples were received by the AVR and told to convert them to analog. The analog differences that you see are due to a number of reasons such as the DAC clock or reference voltage and bleeding of system noise intos its analog outputs, e.g. through power supply rails (connections).

The differences show that just because we are using a digital interface, it does not mean that we get the same analog output from the DAC. The audio samples remain the same but the way the DAC converts them into analog varies.

okay, cool...  still...  I can't help but wonder if the HDMI "distortions" are actually the resonance caused by the harmonic distribution of the sound waves themselves in the space which they are occupying...   i.e. the sound wave from the test tone is crossing itself, creating multiples/divisions of the source waveform..... which is what it looks like.

On the odd chance that the HDMI distortion is actually harmonic resonance...this would be very interesting indeed.  Contrary to how I assume you are depicting this data, my observation would indicate that the HDMI is actually producing a significantly more accurate signal.  This observation would be based on the notion that the waveform sourced from the fiber optic connection is so distorted that the tone it produces is unable to harmonize with itself by any significant measure.

It kind of sounds nuts...  I know.  Then again, one begins to question the structural integrity of a glass thread.  Just something to consider.

Do it again in a different room.:D 

morphemes is offline  
post #113 of 144 Old 05-09-2014, 08:02 AM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by JHAz View Post

i would have thought you were aware that other factors affect audibility. I didn't think I needed to put up every freaking caveat to every freaking point every time I type on here. I'll call you bosso, junior.
???
Quote:
The THD of a -90dB spurious signal is .0031623 percent. Even the hyper -subjectivists at Stereophile use 1 percent distortion as the threshold of "clipping," for electronics except when they relax it to 3% for some tube amps to give their manufacturers a little room to save face. I am sure a person could design a signal, say with the fundamental at 100 Hz and all the distortion in the 1000-4000Hz range, and play it loudly enough to perhaps, perhaps make -90dB audible. But in the real world? Especially given that whatever your electronics are doing the transducers to which I referred are cranking out what 10 to 100 times the distortion? Masking effects, etc. etc. etc.
Thanks for the math. THD is an obsolete measure. It was designed long time ago to be something that is easy to measure. It is totally devoid of psychoacoustics science. To understand why, we need to first look at what THD is supposed to be. Mathematically it is the square root of the sum of the harmonics combined:

d546b2fa7653a4e0f9669f0adc19bee7.png

Focusing on the numerator, we see that there is equal weighting to each harmonic. That violates our understanding of how our hearing works. "Masking" theory says that the earlier harmonics (multiples of our original source signal) are less audible than the later ones. The math for that is complicated since it depends on which source frequency we are talking about. Computer analysis is required to arrive at perceptually accurate value which no one performs. The result is a number that is inflated and psychoacoustically blind. If THD were valid, MP3 would never work!

To use THD to dismiss the audibility of DAC distortions would be the proverbial jumping from the frying pan into the fire. It is a lot easier to shoot holes in THD than DAC distortions.

Another reason to not use THD is that it has no bearing to the type of distortions we are talking about here. If I have a 1 Khz tone, THD measures harmonic distortions created at multiples of 1 Khz, e.g. 2Khz, 3 Khz, etc. DAC distortions are "aharmonic" meaning that they are generated with no regards to the frequency of our source tone. If I modulate the clock of the DAC by 200 Hz, I get +- 200 Hz sidebands no matter what my source frequency is. At 1 Khz, we would get distortion spikes at 800 Hz and 1.2 Khz. At 5 Khz, we would get distortions at 4.8 Khz and 5.2 Khz. Lack of similarity to the original signal means their audibility is very different than simple harmonic distortions.

As to values being small, you are going by the one sample measurement that I put out there for another reason. There is no rhyme or reason on what spectrum of distortion these mass market AVRs produce over HDMI. Here is an example with much elevated distortion spikes:

i-7wBmTvV-XL.png

Now the distortion spikes are in -70 db range, not -90. Who knows then what distortion your AVR or OP has. Maybe it is better, maybe worse. The only thing we know with confidence is that it is higher than S/PDIF No data has been presented to dispute this. As such, if you have an option of using either interface for music purposes my recommendation, backed by a lot of measurements is S/PDIF over HDMI. There is just no reason to put up with more distortion and grasp sinking life rafts like THD to say it doesn't matter.
Quote:
I do not care if I COULD HAVE MAYBE heard the spurious frequencies if only the fundamental was not present or was reduced by 40 or 50 or 60 or 90 dB. I care if I can hear the distortion in a real context. The rest is of great theoretical interest to me but impacts my enjoyment precisely zero.
I can only present the data. You draw the conclusions you want. As far as I am concerned, every time we post saying these distortions don't matter we are doing a disservice to the rest of us. Currently no one measures these distortions in any review. As such, the performance and level of distortion remains unknown and not dealt with. Look at my signature. It is not my genes to sanction poor engineering. It is not like good HDMI implementations are impossible. Here is the ancient Mark Levinson 502 Processor compared to the Anthem with both being tested using HDMI:

i-5Btksr7-XL.png

Look at how it outperforms the Anthem across the board -- both as far as noise level and distortions. Once we have this level of performance then you can use either interface and be fine.

To the extent you can't convince your doctor that being overweight by 10 pounds is OK, you can't convince me that having so much distortion over HDMI is OK. I am going to encourage people to use alternate interfaces which will cost you either nothing or a few dollars for a cable. No amount of stomping one's feat is going to change my recommendation. I see no reason to put up with elevated distortions when we have an alternative digital interface devoid of it.
Quote:
Here's why. If the IM or additive distortion caused by low level (ie probably inaudible in the context of the signal) added up to audible distortion, then get this: MP3 REMOVES all those low level things that are otherwise masked.
I find it odd that you use the example of MP3 and talk about masking yet you presented THD to make your case. How often have you heard of THD measurements of MP3 encoders? Never, right? How about IM distortion? Never again, right? These old measurements have little reason to exist in the context of audibility discussions.
Quote:
So high bitrate MP3s should, by all rights, sound BETTER than the original because distortions that would have accumulated via the interaction of the otherwise inaudible parts of the signal are gone,
That is not how MP3 works. It is not removing things in the way you are imagining.
Quote:
I'm saying if you remove low low level signals that otherwise (one imagines) should add to each other and create distortions, but you cannot hear the distortions, then the removal of those low low level signals didn't reduce distortion. IMO, if it ain't audible in a proper double blind test it ain't audible
Again, that is not how MP3 works. But yes, double blind tests would be good but none has been presented on difference between HDMI and S/PDIF. We don't get to imply that if one existed it would have X results. That is not "scientific." So let's not use that tactic lest we want to look like we are spinning the science.
Quote:
And IMO there's way too much money at stake for it to be the case that nobody has gone looking for those circumstances where, in fact, audible differences at a statistically significant can be found among some segment of the population in properly conducted double blind tests. The non-existence of such published data suggests the non-existence of the phenomenon, because the research surely would be funded. Again, I don't care if I might be able to hear a sound, reproduced by itself, pristinely, at 5 dB in my room, if the sound, in the context in which it occurs is inaudible because there's other stuff going on.

Rock on.
Sorry can't follow your logic here. Arny said at the beginning of this thread that he could run such a test in a "heartbeat." Now you say funding needs to be there to run such a listening test? Who funded me performing these measurements? Nobody. I was curious to have first hand data on the performance of HDMI over S/PDIF and set out weeks of my free time to create this data. That is the *only* data we have. Everything else is speculation. If you guys believe so much in blind tests to settle the score, then you should get together and run them. Complaining about the measurement data with arguments that unclear and irrelevant don't help our cause.
Riffmeister likes this.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #114 of 144 Old 05-09-2014, 08:15 AM
 
mc2ed's Avatar
 
Join Date: Apr 2014
Posts: 198
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 20

I agree with Amir!

 

Let's organize some regional AVS Summits!

mc2ed is offline  
post #115 of 144 Old 05-09-2014, 12:30 PM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 3,958
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 95 Post(s)
Liked: 154
Quote:
Originally Posted by amirm View Post

???
Thanks for the math. THD is an obsolete measure. It was designed long time ago to be something that is easy to measure. It is totally devoid of psychoacoustics science. To understand why, we need to first look at what THD is supposed to be. Mathematically it is the square root of the sum of the harmonics combined:

d546b2fa7653a4e0f9669f0adc19bee7.png

Focusing on the numerator, we see that there is equal weighting to each harmonic. That violates our understanding of how our hearing works. "Masking" theory says that the earlier harmonics (multiples of our original source signal) are less audible than the later ones. The math for that is complicated since it depends on which source frequency we are talking about. Computer analysis is required to arrive at perceptually accurate value which no one performs. The result is a number that is inflated and psychoacoustically blind. If THD were valid, MP3 would never work!

To use THD to dismiss the audibility of DAC distortions would be the proverbial jumping from the frying pan into the fire. It is a lot easier to shoot holes in THD than DAC distortions.

Another reason to not use THD is that it has no bearing to the type of distortions we are talking about here. If I have a 1 Khz tone, THD measures harmonic distortions created at multiples of 1 Khz, e.g. 2Khz, 3 Khz, etc. DAC distortions are "aharmonic" meaning that they are generated with no regards to the frequency of our source tone. If I modulate the clock of the DAC by 200 Hz, I get +- 200 Hz sidebands no matter what my source frequency is. At 1 Khz, we would get distortion spikes at 800 Hz and 1.2 Khz. At 5 Khz, we would get distortions at 4.8 Khz and 5.2 Khz. Lack of similarity to the original signal means their audibility is very different than simple harmonic distortions.

As to values being small, you are going by the one sample measurement that I put out there for another reason. There is no rhyme or reason on what spectrum of distortion these mass market AVRs produce over HDMI. Here is an example with much elevated distortion spikes:

i-7wBmTvV-XL.png

Now the distortion spikes are in -70 db range, not -90. Who knows then what distortion your AVR or OP has. Maybe it is better, maybe worse. The only thing we know with confidence is that it is higher than S/PDIF No data has been presented to dispute this. As such, if you have an option of using either interface for music purposes my recommendation, backed by a lot of measurements is S/PDIF over HDMI. There is just no reason to put up with more distortion and grasp sinking life rafts like THD to say it doesn't matter.
I can only present the data. You draw the conclusions you want. As far as I am concerned, every time we post saying these distortions don't matter we are doing a disservice to the rest of us. Currently no one measures these distortions in any review. As such, the performance and level of distortion remains unknown and not dealt with. Look at my signature. It is not my genes to sanction poor engineering. It is not like good HDMI implementations are impossible. Here is the ancient Mark Levinson 502 Processor compared to the Anthem with both being tested using HDMI:

i-5Btksr7-XL.png

Look at how it outperforms the Anthem across the board -- both as far as noise level and distortions. Once we have this level of performance then you can use either interface and be fine.

To the extent you can't convince your doctor that being overweight by 10 pounds is OK, you can't convince me that having so much distortion over HDMI is OK. I am going to encourage people to use alternate interfaces which will cost you either nothing or a few dollars for a cable. No amount of stomping one's feat is going to change my recommendation. I see no reason to put up with elevated distortions when we have an alternative digital interface devoid of it.
I find it odd that you use the example of MP3 and talk about masking yet you presented THD to make your case. How often have you heard of THD measurements of MP3 encoders? Never, right? How about IM distortion? Never again, right? These old measurements have little reason to exist in the context of audibility discussions.
That is not how MP3 works. It is not removing things in the way you are imagining.
Again, that is not how MP3 works. But yes, double blind tests would be good but none has been presented on difference between HDMI and S/PDIF. We don't get to imply that if one existed it would have X results. That is not "scientific." So let's not use that tactic lest we want to look like we are spinning the science.
Sorry can't follow your logic here. Arny said at the beginning of this thread that he could run such a test in a "heartbeat." Now you say funding needs to be there to run such a listening test? Who funded me performing these measurements? Nobody. I was curious to have first hand data on the performance of HDMI over S/PDIF and set out weeks of my free time to create this data. That is the *only* data we have. Everything else is speculation. If you guys believe so much in blind tests to settle the score, then you should get together and run them. Complaining about the measurement data with arguments that unclear and irrelevant don't help our cause.

I am not Arny. SO what he says doesn't affect what I say. Is that too complex a concept. Should I just go away, since I don't much care about all this weenie wagging anyway?

Inaudible distortion levels are inaudible. I'd love to have amps that have amazingly great THD and IM, but cannot afford them. I am perfectly aware that my desire is intellectual (or perhaps ego) based, and that if it ain't audible it ain't audible. Reducing inaudible distortion to less inaudible (or more inaudible depending on which side of the curve you're looking at) doesn't make the inaudible distortion audible. That which makes no audible difference makes no audible difference. In general, it seems to me, DBTs demonstrate that a lot of measurable differences are inaudible, that one percent distortion is, at least in the vast majority of cases, inaudible, so that distortion levels at a tenth or a hundredth or less of that level HAVE to be inaudible, unless logic ceases to apply. "I cannot hear the added tone when it's 50 dB below the signal, but if it's 90 dB below the signal, I can" just is so utterly nonsensical, at least to me, as to not require further discussion. Saying you want less distortion is wonderful, unless you're telling people that they MUST attend to levels of distortion that are, according to what I understand, clearly demonstrated to be inaudible. Or if you prefer, not demonstrated to be audible. I'd like to sing all my parts to within a half cent of perfect pitch, but (a) nobody can tell if I am 3 cents off, especially with backing instrumentation, and, (more importantly but not relevant to this discussion) re-singing every phrase until the pitch is perfect is how you end up with emotionally dead music. It's why classical artists want to see their performances complete. Little differenes (that may sound like errors to, say Hilary Hahn) inform how they play the next phrase, or a related phrase in the next movement. It all goes together, musically, and the more we take it apart, the less soul remains. Whether you are Hilary or Henley. But again, not strictly relevant because we are talking about reproduction, not production.

I stand by my main point, though. If distortions at -90dB actually could be proven to impact the listening experience through scientifically reproducible testing (or if Conrad-Johnson could prove their baked in set of frequency and harmonic distortions are massively preferred, just as an example) it would be worth millions of dollars over time. IME, science and psuedo science (in non-audio arenas, at least) that is worth even hundreds of thousands gets done, and gets published. While golden eared superhumans may view themselves as above science, their economics work exactly like everybody else's. If I prove my widget is better, I'll sell more widgets. I'll make more money. So if it were falling-off-a-log easy to present the proof, I'd'a done it decades ago. Or maybe everybody gets into audio manufacturing to avoid making too much money. Somehow I doubt it . . .
JHAz is offline  
post #116 of 144 Old 05-09-2014, 01:38 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by JHAz View Post

I am not Arny. SO what he says doesn't affect what I say. Is that too complex a concept. Should I just go away, since I don't much care about all this weenie wagging anyway?
I know you are not Arny. I actually made the opposite point that your statement that you need funding to run double blind bests on this topic is at odds with what he said. You two need to sort that out. You need to explain what you think is expensive to do and have him set you straight or accept he is wrong. Without that discussion and documentation it leaves the situation as we don't even know what is involved in the blind tests that we keep saying have saved hungry children in Africa.
Quote:
Inaudible distortion levels are inaudible. I'd love to have amps that have amazingly great THD and IM, but cannot afford them.
I just got done explaining why THD and IM are useless metrics and you continue to use them in your arguments. Once more, we cannot tel what is audible or inaudible using those metrics.
Quote:
I am perfectly aware that my desire is intellectual (or perhaps ego) based, and that if it ain't audible it ain't audible. Reducing inaudible distortion to less inaudible (or more inaudible depending on which side of the curve you're looking at) doesn't make the inaudible distortion audible. That which makes no audible difference makes no audible difference. In general, it seems to me, DBTs demonstrate that a lot of measurable differences are inaudible, that one percent distortion is,
They do. Problem we have once again is that no double blind tests have been cited for this topic. I repeat again: the only data presented shows that HDMI consistently has much elevated distortions relative to S/PDIF in the same AVR. No argument has been made why, when I have a choice of both interfaces at zero or a few dollar costs, I should use the interface with more distortion.
Quote:
...at least in the vast majority of cases, inaudible, so that distortion levels at a tenth or a hundredth or less of that level HAVE to be inaudible, unless logic ceases to apply. "I cannot hear the added tone when it's 50 dB below the signal, but if it's 90 dB below the signal, I can" just is so utterly nonsensical, at least to me, as to not require further discussion. Saying you want less distortion is wonderful, unless you're telling people that they MUST attend to levels of distortion that are, according to what I understand, clearly demonstrated to be inaudible.
You are not consistent in this paragraph. You first say "vast majority of cases" but then finish with "clearly demonstrated to be inaudible." And how has it been clearly demonstrated to be inaudible? Demonstrated how? Again, no listening test results are shown for DAC distortions. Inferences from THD and such don't count and even there, no specific test or data has been presented.

On the "vast majority of cases" I can say that about MP3. In "vast majority of cases" MP3 at 128 kbps is transparent to its source. I can get you 100 people and have 95 of them say they can't tell the difference. Does that mean that we have "clearly demonstrated" the MP3 distortions to be inaudible at 128 kbps?

You have to keep your audience in mind when you make these arguments. I know all of these tactics we use to convince the other side. The key phrase is "the other side." I am on your side of the argument. What differs between us is that I don't try to use persuasion, inference, vague references, etc. to make an argument for inaudibility. You can't do that and still say you are on the objective side of things.
Quote:
Or if you prefer, not demonstrated to be audible.
I have demonstrated clearly with data and measurements that HDMI has vastly more distortion than S/PDIF. We are talking about orders of magnitude higher. As much as we use that argument of "you show me the difference is audible and if you can't it must be inaudible," it is the wrong tactic. You have real data in front of you showing HDMI to be the dirtier interface on the same machine. And machine after machine. Based on that I made a recommendation to use S/PDIF over HDMI for music applications. Why all the protests? What do you lose if the person uses S/PDIF over HDMI? What do we all gain by papering over distortions on an interface that is 20 years newer to be so much worse? You are advocating mediocracy.

You say that you love to have an amplifier that has less THD and IMD. Well, S/PDIF has less distortion than HDMI. And it doesn't cost you a dime to get there. How come it doesn't sit well with you? Doesn't sound like you really wanted a lower distortion product to keep arguing this way.
Quote:
I stand by my main point, though. If distortions at -90dB actually could be proven to impact the listening experience through scientifically reproducible testing (or if Conrad-Johnson could prove their baked in set of frequency and harmonic distortions are massively preferred, just as an example) it would be worth millions of dollars over time. IME, science and psuedo science (in non-audio arenas, at least) that is worth even hundreds of thousands gets done, and gets published. While golden eared superhumans may view themselves as above science, their economics work exactly like everybody else's. If I prove my widget is better, I'll sell more widgets. I'll make more money. So if it were falling-off-a-log easy to present the proof, I'd'a done it decades ago. Or maybe everybody gets into audio manufacturing to avoid making too much money. Somehow I doubt it . . .
A lot of inferences free of real data and explanation of science. And off-topic. OP asked which interface is better. Do you have data and listening tests that back your point that the two are equal? If so post it. Please don't keep using these layman arguments with me. I know them all and then some. Remember, I am not here telling you I listened to two speaker wires and one sounded better. I showed you detailed measurements using $20,000 pieces of equipment. This is not what subjectivists do. Please don't lump me in there and then use the camp talking points on me. Provide data, or proper explanation of science and then I listen.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #117 of 144 Old 05-09-2014, 01:49 PM
Senior Member
 
koturban's Avatar
 
Join Date: Jan 2007
Posts: 356
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 133 Post(s)
Liked: 96
20k pieces of equipment? Wow. That convinces me.
koturban is offline  
post #118 of 144 Old 05-09-2014, 04:19 PM
AVS Addicted Member
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 18,114
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 768 Post(s)
Liked: 448
Quote:
Originally Posted by koturban View Post

20k pieces of equipment? Wow. That convinces me.
Yeh, you tell him. Oh wait, you are shooting at me! biggrin.gif

Take a look at this measurement:

i-bs5DqCC-X2.png

This is a special measurement that shows how well the DAC filters incoming jitter. The input to the DAC always has jitter. The jitter can be from the source, the cable, or inside the box. Here we measure how well the DAC filters jitter that would manifests itself in the audio band. As I have indicated on the graph, the lower the graph, the better the filtering.

Note that this is using S/PDIF interface. Despite being the better implemented input on the AVRs, they all do exceedingly poor here. Some even amplify the incoming jitter! Now we see why HDMI performance is poor in these products. When we use HDMI tons of circuits inside the AVR become active to deal with video and rich set of features in that interface. Since the DACs in AVRs have poor rejection/filtering of incoming jitter, all of that activity bleeds into them resulting in distortion in the output of their digital to audio converter (DAC).

The units that do better are all "high-end" products. First is the Lexicon 12B, a now obsolete processor costing $12K. Notice how it beats all the mass market (premium) AVRs that were released later than it.

Lower down yet and by a good margin is the Mark Levinson No 502 which retailed for $30K. Another obsolete product that predated these AVRs. Yet is performs superbly in rejecting jitter. No wonder then that its HDMI performance is also leaps and bounds better than mass market AVRs.

Way, way lower at the bottom is my dedicated Mark Levinson DAC No 36S. This thing eats jitter for breakfast biggrin.gif.I bought this i think in year 2000. So it was 12 years old at the time of this test. Yet, it shows superlative performance in how it rejects incoming jitter. Excellent engineering comes right through the measurements despite it being more than a decade older than the AVRs tested.

Note that the lower the graph, the less the source device can impact the performance of the downstream DAC. Provide no filtering as in the case of AVRs and your performance will be source dependant. Therefore, to get good performance out of these AVRs, you better have low jitter output sources.

Now you see why a $20,000 instrument is helpful in this regard. It enables one to gain key insight as to why the measurements were as bad as they were when using HDMI in AVRs. Without such a system you would have no idea what is going on. With it, we can objectively categorize products and learn to separate good engineering from sloppy.

Amir
Founder,
To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

"Insist on Quality Engineering"

amirm is offline  
post #119 of 144 Old 05-09-2014, 08:41 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 amirm View Post

Yeh, you tell him. Oh wait, you are shooting at me! biggrin.gif

LOL!!! Some of those obscure, snarky comments can be confusing.
Quote:
Take a look at this measurement:

i-bs5DqCC-X2.png

This is a special measurement that shows how well the DAC filters incoming jitter. The input to the DAC always has jitter.

While it might be true that the pricier components reduce spdif INPUT jitter, I'm not convinced that input jitter is even worthy of mention. I queried reviews of the Oppo blu-ray players from a trusted website, Audioholics. The lesser endowed unit costs $499. This is what was reported:

"We ran both 192kHz/24 bit 6 channel and 96kHz/ 24 bit 8 channel Dolby TrueHD signals into both players and both produced similarly ruler flat frequency performance from 20Hz all the way up to the Audio Precision test gear bandwidth limitation (80kHz).

We also ran Bit Error Rate (BER) tests on both players using Dolby TrueHD and DTS HD test signals and both players produced a 0% BER which means both players delivered audio via HDMI 100% error free. Early HDMI products (ver 1.0 to 1.2) reportedly suffered from jitter related issues but HDMI ver 1.3 and above has completely eliminated jitter related issues thus ensuring bit for bit exact signal transfer from the player to the A/V receiver or processor."

http://www.audioholics.com/blu-ray-and-dvd-player-reviews/copy_of_oppo-bdp-93/oppo-bdp-93-bdp-95-on-the-bench
Quote:
Note that this is using S/PDIF interface. Despite being the better implemented input on the AVRs, they all do exceedingly poor here. Some even amplify the incoming jitter! Now we see why HDMI performance is poor in these products.

Seems there is no need for "great" performance with jitter input as long as a player costing $499 is within financial means. Clearly the performance of the Oppo players must be some form of enigma, the 8th wonder of the world. How about a $40 portable mp3 player? While not measuring hdmi or spdif output, and the source not as trustworthy as Audioholics, here is what I found:

http://nwavguy.blogspot.com/2011/02/sansa-clip-measured.html



The inexpensive, little Sansa Clip gadget measures less than -100dB. Though lacking the architectural beauty of the Shwedagon Pagoda the Clip might just register as the 9th Wonder of the World.
Quote:
Now you see why a $20,000 instrument is helpful in this regard. It enables one to gain key insight as to why the measurements were as bad as they were when using HDMI in AVRs. Without such a system you would have no idea what is going on. With it, we can objectively categorize products and learn to separate good engineering from sloppy.

I wouldn't call it sloppy engineering. I'll call it intelligent, cost-effective design. Why engineer a solution for a problem that doesn't exist?
UndersAVS is offline  
post #120 of 144 Old 05-09-2014, 09:40 PM
 
mc2ed's Avatar
 
Join Date: Apr 2014
Posts: 198
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 20
Quote:
Originally Posted by UndersAVS View Post


LOL!!! Some of those obscure, snarky comments can be confusing.
While it might be true that the pricier components reduce spdif INPUT jitter, I'm not convinced that input jitter is even worthy of mention. I queried reviews of the Oppo blu-ray players from a trusted website, Audioholics. The lesser endowed unit costs $499. This is what was reported:

"We ran both 192kHz/24 bit 6 channel and 96kHz/ 24 bit 8 channel Dolby TrueHD signals into both players and both produced similarly ruler flat frequency performance from 20Hz all the way up to the Audio Precision test gear bandwidth limitation (80kHz).

We also ran Bit Error Rate (BER) tests on both players using Dolby TrueHD and DTS HD test signals and both players produced a 0% BER which means both players delivered audio via HDMI 100% error free. Early HDMI products (ver 1.0 to 1.2) reportedly suffered from jitter related issues but HDMI ver 1.3 and above has completely eliminated jitter related issues thus ensuring bit for bit exact signal transfer from the player to the A/V receiver or processor."

http://www.audioholics.com/blu-ray-and-dvd-player-reviews/copy_of_oppo-bdp-93/oppo-bdp-93-bdp-95-on-the-bench
Seems there is no need for "great" performance with jitter input as long as a player costing $499 is within financial means. Clearly the performance of the Oppo players must be some form of enigma, the 8th wonder of the world. How about a $40 portable mp3 player? While not measuring hdmi or spdif output, and the source not as trustworthy as Audioholics, here is what I found:

http://nwavguy.blogspot.com/2011/02/sansa-clip-measured.html



The inexpensive, little Sansa Clip gadget measures less than -100dB. Though lacking the architectural beauty of the Shwedagon Pagoda the Clip might just register as the 9th Wonder of the World.
I wouldn't call it sloppy engineering. I'll call it intelligent, cost-effective design. Why engineer a solution for a problem that doesn't exist?


I just found a picture of me online, from the 90's... I found it intriguing...

 

However, this last post is much more intriguing to me...

 

I am going to buy this product this weekend and measure it. And yes, I will post my measurement's and thoughts...

 

Sorry guys, not to take away from your effort in here, but you both know this crowd... If my finding mirror yours, we'll have the beginnings of a small up-rise in our r hands... what's new! LOL!

 

I trust your individual integrities. PM if you desire...

 

Objective brother, signing out!

mc2ed is offline  
Closed Thread Audio theory, Setup and Chat

User Tag List

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