Is High-Resolution Audio Irrelevant? - Page 22 - AVS Forum | Home Theater Discussions And Reviews
Baselworld is only a few weeks away. Getting the latest news is easy, Click Here for info on how to join the Watchuseek.com newsletter list. Follow our team for updates featuring event coverage, new product unveilings, watch industry news & more!


Forum Jump: 
 17Likes
 
Thread Tools
Old 06-11-2014, 10:58 PM
AVS Special Member
 
steveting99's Avatar
 
Join Date: Jun 2011
Location: Hong Kong
Posts: 2,914
Mentioned: 19 Post(s)
Tagged: 0 Thread(s)
Quoted: 1563 Post(s)
Liked: 493
^arnyk,

I think you've got much more understanding of limitations brick wall filters employed in the recording process than me. From what you're saying the implementations of these filters would limit the harmonic frequencies from 11kHz upwards? How about IMD from the lower frequencies imposing on the waveforms from below 11kHz?

I note that most human hearing is most sensitive around 2kHz to 6kHz. So this is the range where everything is supposed to be reproduced as clean as possible.

Marantz SR-7009, Oppo BPD-93, MTV 7000D, aTV, Harmony Home Control remote, KEF E301+T101, MiniDSP (2x4)+(10x10HD)+(DDRC-88A), Emotiva 2xXPA-5 (Gen2), Rythmik 2xF12G+2xF8, HiMedia Q16, LG 55EC9300.
steveting99 is offline  
Sponsored Links
Advertisement
 
Old 06-12-2014, 04:25 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,593
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1232
Quote:
Originally Posted by eljr View Post
Quote:Originally Posted by WiWavelength 

Trust me,

AJ


LOL, thought this was an objective science forum?
Figure of speech, no?

The basic statement agrees with objective truth.
arnyk is offline  
Old 06-12-2014, 04:35 AM
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,593
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 851 Post(s)
Liked: 1232
Quote:
Originally Posted by steveting99 View Post
^arnyk,

From what you're saying the implementations of these filters would limit the harmonic frequencies from 11kHz upwards?
Yes.

Quote:
How about IMD from the lower frequencies imposing on the waveforms from below 11kHz?
That, too. The thing about IM is that the original signals stimulating it can be at just about any frequency, audible or not.

Quote:
I note that most human hearing is most sensitive around 2kHz to 6kHz. So this is the range where everything is supposed to be reproduced as clean as possible.

Agreed. The thing about IM is that signals all over the frequency spectrum can create IM products in this range. All you need is two frequencies that are 2 to 6 KHz apart. So they could be 8 KHz and 10 KHz (for a 2 KHz difference) or it could be 50 KHz and 56 Khz (fpor 6 Khz difference).

The rules for creating IM have more opportunities for distortion creation than harmonic distortion. Harmonic distortion is easy, its always integer multiples. The IM products can be at sums, differences, sums of harmonics, differences of harmonics, all sorts of things.
arnyk is offline  
Old 06-14-2014, 02:52 AM
Newbie
 
Join Date: Jun 2014
Posts: 2
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 4
Is HD audio irrelevant? - an experiment

Take a look at the result of this experiment (I used Adobe Audition):

What I did:

1. Pick an HD track that is sampled at, say, 96k/24bit.
2. Convert it to 44.1/16bit PCM (.wav)
3. Upsample the .wav file to the original 96k.
4. Invert one of the waves, say, the converted and upsampled one.
5. Mix the two tracks and listen to the result.
VERY IMPORTANT! You must align the tracks PERFECTLY for the experiment to work.

The results:

First, take a look at the spectrograms:

Original 96k/24bit track

Converted 44.1k/16bit track upsampled to 96k.

Mixed track (What you see is what you cut when you downsample the track. There is certainly a huge amount of data there!)

Now listen to the mixed track (that's exactly what you are losing when you convert this track from HD to CD).
After all we know about the physiology of our sense of hearing, could we have expected a different result?

My conclusion: neither bit depth further than 16bit nor high frequency sampling have anything to offer to improve the audible part of the signal. It's high time we made a clear distinction between data loss and degraded sound quality.

Last edited by gabebcn; 06-14-2014 at 02:57 AM.
gabebcn is offline  
Old 06-14-2014, 04:19 AM
Newbie
 
Join Date: Jun 2014
Posts: 2
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 4
By the way, read this excellent article on the subject, also referred to by Scott Wilkinson.
Steve Crowley likes this.
gabebcn is offline  
Old 06-22-2014, 12:15 PM
Senior Member
 
hogger129's Avatar
 
Join Date: Jan 2013
Location: The Dark Side Of The Moon
Posts: 282
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 86 Post(s)
Liked: 73
I think high-resolution audio is irrelevant for listening. For listening, I don't see any reason to go higher than 16/48. For mastering, that is where I could see using something like 24/176.4 or 24/192, etc.
hogger129 is offline  
Old 07-12-2014, 06:26 AM
 
amirm's Avatar
 
Join Date: Jan 2002
Location: Washington State
Posts: 19,042
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 1334 Post(s)
Liked: 739
Quote:
Originally Posted by gabebcn View Post
My conclusion: neither bit depth further than 16bit nor high frequency sampling have anything to offer to improve the audible part of the signal. It's high time we made a clear distinction between data loss and degraded sound quality.
That is certainly the theory. In practice my recent testing shows otherwise. Arny was kind enough to give us a file with good bit of high frequency information (keys jingling) that he then resampled to 44.1 Khz and 32 Khz. The files were then converted back to the original bit depth and sample rate of 24/96 Khz. Arny's position was that not only 44.1 Khz was transparent but so was the 16 Khz as there is "no useful information" above that. He said he had run both blind and sighted tests and no one could tell the difference.

I performed a double blind using Foobar2000 with its ABX plug-in. Here are the results as I posted here:

Quote:
Originally Posted by amirm View Post
I don't know that I can even hear 14 Khz! Yet these are my results as i post in the other thread: Debate Thread: Scott's Hi-res Audio Test

--------------


Good morning Arny. I was going to say "thank you" for posting these files but after having to listen to jingling keys so many times while our two dogs barked and barked away, not sure I am that thankful .

Here are my results:

32 Khz versus 96 Khz
=================================
foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/09 06:10:07

File A: C:\Users\Amir\Music\Arnys Filter Test\keys jangling band resolution limited 3216 2496.wav
File B: C:\Users\Amir\Music\Arnys Filter Test\keys jangling full band 2496.wav

06:10:07 : Test started.
06:10:38 : 01/01 50.0%
06:10:50 : 02/02 25.0%
06:11:07 : 03/03 12.5%
06:11:23 : 04/04 6.3%
06:11:36 : 05/05 3.1%
06:12:00 : 06/06 1.6%
06:12:14 : 07/07 0.8%
06:12:26 : 08/08 0.4%
06:12:38 : 09/09 0.2%
06:12:49 : 10/10 0.1%
06:13:00 : 11/11 0.0%
06:13:23 : 12/12 0.0%
06:13:42 : 13/13 0.0%
06:13:48 : Test finished.

----------
Total: 13/13 (0.0%)


44.1 versus 96 Khz
---------------------------------

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/09 06:32:02

File A: C:\Users\Amir\Music\Arnys Filter Test\keys jangling band resolution limited 4416 2496.wav
File B: C:\Users\Amir\Music\Arnys Filter Test\keys jangling full band 2496.wav

06:32:02 : Test started.
06:33:07 : 01/01 50.0%
06:33:17 : 02/02 25.0%
06:33:24 : 03/03 12.5%
06:33:36 : 04/04 6.3%
06:33:47 : 05/05 3.1%
06:33:58 : 06/06 1.6%
06:34:12 : 07/07 0.8%
06:34:15 : Test finished.

----------
Total: 7/7 (0.8%)

===============================

I don't know why Foobar stopped all of a sudden at 7 trials on 44.1 vs 96. While I could clearly hear the difference between the files, I would want to run more trials later as I did not expect to be able to tell them apart this easily.

Anyway, how did you do Arny?
As you can see, I was able to differentiate the files completely. The 32 Khz sampling was very distorted. Another poster, Frank, got the same scores as I did in that file. 44.1 Khz was harder to distinguish but I managed to do so with 7 out of 7 trials correctly identified. No one else has post results for that so perhaps this is dependent on one's listening ability.

Scott was kind enough to create another set of files for the same comparison. See: AVS/AIX High-Resolution Audio Test: Take 2

I ran those files double blind with these results:

Quote:
Originally Posted by amirm View Post
Thank you Scott! Much appreciated the effort you have put on this project Scott. For the first time I feel that the forum is moving forward toward better understanding of this topic.

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/10 18:50:44

File A: C:\Users\Amir\Music\AIX AVS Test files\On_The_Street_Where_You_Live_A2.wav
File B: C:\Users\Amir\Music\AIX AVS Test files\On_The_Street_Where_You_Live_B2.wav

18:50:44 : Test started.
18:51:25 : 00/01 100.0%
18:51:38 : 01/02 75.0%
18:51:47 : 02/03 50.0%
18:51:55 : 03/04 31.3%
18:52:05 : 04/05 18.8%
18:52:21 : 05/06 10.9%
18:52:32 : 06/07 6.3%
18:52:43 : 07/08 3.5%
18:52:59 : 08/09 2.0%
18:53:10 : 09/10 1.1%
18:53:19 : 10/11 0.6%
18:53:23 : Test finished.

----------
Total: 10/11 (0.6%)

Quote:
Originally Posted by amirm View Post
The third track was pretty easy. First segment picked was quite revealing:

foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/10 21:01:16

File A: C:\Users\Amir\Music\AIX AVS Test files\Just_My_Imagination_A2.wav
File B: C:\Users\Amir\Music\AIX AVS Test files\Just_My_Imagination_B2.wav

21:01:16 : Test started.
21:02:11 : 01/01 50.0%
21:02:20 : 02/02 25.0%
21:02:28 : 03/03 12.5%
21:02:38 : 04/04 6.3%
21:02:47 : 05/05 3.1%
21:02:56 : 06/06 1.6%
21:03:06 : 07/07 0.8%
21:03:16 : 08/08 0.4%
21:03:26 : 09/09 0.2%
21:03:45 : 10/10 0.1%
21:03:54 : 11/11 0.0%
21:04:11 : 12/12 0.0%
21:04:24 : Test finished.

----------
Total: 12/12 (0.0%)


Quote:
Originally Posted by amirm View Post
foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/11 06:18:47

File A: C:\Users\Amir\Music\AIX AVS Test files\Mosaic_A2.wav
File B: C:\Users\Amir\Music\AIX AVS Test files\Mosaic_B2.wav

06:18:47 : Test started.
06:19:38 : 00/01 100.0%
06:20:15 : 00/02 100.0%
06:20:47 : 01/03 87.5%
06:21:01 : 01/04 93.8%
06:21:20 : 02/05 81.3%
06:21:32 : 03/06 65.6%
06:21:48 : 04/07 50.0%
06:22:01 : 04/08 63.7%
06:22:15 : 05/09 50.0%
06:22:24 : 05/10 62.3%
06:23:15 : 06/11 50.0% <---- difference found reliably. Note the 100% correct votes from here on.
06:23:27 : 07/12 38.7%
06:23:36 : 08/13 29.1%
06:23:49 : 09/14 21.2%
06:24:02 : 10/15 15.1%
06:24:10 : 11/16 10.5%
06:24:20 : 12/17 7.2%
06:24:27 : 13/18 4.8%
06:24:35 : 14/19 3.2%
06:24:40 : 15/20 2.1%
06:24:46 : 16/21 1.3%
06:24:56 : 17/22 0.8%
06:25:04 : 18/23 0.5%
06:25:13 : 19/24 0.3%
06:25:25 : 20/25 0.2%
06:25:32 : 21/26 0.1%
06:25:38 : 22/27 0.1%
06:25:45 : 23/28 0.0%
06:25:51 : 24/29 0.0%
06:25:58 : 25/30 0.0%

06:26:24 : Test finished.

----------
Total: 25/30 (0.0%)


So we now have 3 out of 3 positive detection of differences in Scott's clips.
Summarizing, I managed to consistently tell all three files apart from their downsampled 44.1 Khz/16 bit versions.

I have not seen anyone else post their results. I suspect some have PMed them to Scott. Would be interesting to see where other folks land.

It sure would have been nice to have more people run Arny's files as the case there is simpler than Scott's. Unfortunately no one else other than Frank reported them. Sadly Arny and other vocal members against high-resolution audio would not post any data as to how they did in these tests.

I suspect if we had that data we would find that not everyone has the same listening ability. This is counter to the common argument that "there are no golden ears." I am a believer that when it comes to non-linear distortions, there are people who have far better ability than average.

I will be summarizing the content of the other long thread where the posts came from but for now, I am hoping this information makes people think and gets them motivated to run the listening tests and post them. I suspect we are at the cusp of radically reshaping these conversations in the future given the data we now have.
amirm is offline  
Old 07-12-2014, 12:36 PM
AVS Special Member
 
RayDunzl's Avatar
 
Join Date: Sep 2013
Location: near Tampa, Florida
Posts: 3,263
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 1219 Post(s)
Liked: 1409
"The only reason to record a performance to digital with super hot levels is if you hate the music and you want it ruined"



Now, why is that message coming from this page to my pc speakers?


endlessly...

I'll be back later...


System links::: 1.5RQ > digits from all sources > 1177a > OpenDRC-DI > DEQ2496 > DAC2 > KCT > FPB 350mcx > reQuest + Cheezewoofer Wattless Deluxe > Sweetspot
RayDunzl is online now  
Old 07-14-2014, 08:19 AM
Member
 
centurymantra's Avatar
 
Join Date: Apr 2007
Posts: 72
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 11
Post

Quote:
Originally Posted by arnyk View Post


Actually a large proportion of all professional microphones start rolling off an octave lower - at 10 or 12 KHz.

This is the manufacturer's FR spec for what may be the most widely used professional microphone in the world:

http://cdn.shure.com/specification_s..._specsheet.pdf



it is pretty typical.

I am not chiming in to discuss or argue any of the points made in this conversation, but feel compelled to comment on this factoid that you bring up. For a number of years I was involved in audio engineering as the owner of a small recording studio. While it is true that the Shure SM58 is a widely used microphone - and some engineers are avid fans of the SM57/SM58, it is most common in live PA sound. There is a time and place for this microphone, but I can assure you that very little professional recording is done using the SM58, so I would sway away from using this to reinforce your point.


That being said, it is true that most mics have pretty steep roll-off curves and their own quirky frequency responses. I've even used the equivalent of a Dictaphone mic on a few occasions - with a frequency curve that would make that SM58 graph look like a Neumann in comparison. It is really about fitting the unique signature of a mic to the unique sound of what is being recorded to fit within the context of the track you are recording. I have been known to throw a dozen different mics on a vocalist or instrument to find that fit...and sometimes - yes - it might even be an SM58.
tubetwister likes this.
centurymantra is offline  
Old 02-19-2016, 12:53 PM
Member
 
GXMnow's Avatar
 
Join Date: Jul 2010
Posts: 96
Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 81
Hey Scott, I got stuck due to a delayed flight and was browsing around the web and came across this older thread. I found it very interesting to read through ll the comments for and against high res audio, and it became very interesting how people side for or against it as they do.

I have been working in audio for over 30 years and have dealt with every source you can imagine. To me, a recording medium, format, codec etc. should carry the original signal to the destination as accurately as possible. There is of course a definite point of diminishing returns, but I will take a moderate analog recording over 16 bit 44.1 CD recording for many reasons.

I grew up hearing some great analog sources. High speed large track mag tape and very carefully master albums and even though the noise floor may be higher than a cheap CD player, the nuance of the sound is so much clearer. Analog recording do not hard clip. All sounds in the real world have a certain harmonic characteristic. As sounds are made louder, such as strumming a guitar string with more force, hitting a drum or piano key harder, or blowing harder into a trumpet, the harmonic content increases as the sound level does. Analog recordings on an LP or magnetic tape also will have increasing distortion as the saturation limit is reached, but the increase is usually non symmetrical and gradual. The result is an increase in even harmonics which actually sounds very similar to an acoustic instrument being played louder. This is a big reason why people feel a low power tube amp, especially a single ended one, has such a warm musical sound. Solid state push pull amps as well as a digital audio chain, tends to reach their lowest amount of distortion as the signal level is closets to maximum without actually hitting the clip point. Digital audio certainly has more and more bits in use as it gets louder. But once the clip level is eached, these systems hard clip and flaten off both the top and bottom of the waveform. This symmetrical hard clipping very quickly adds a lot of odd harmonics, which sound very harsh and annoying, and is in no way musical to anyone.

How does all this relate to high resolution recordings???

I just purchased a few more CD's because I really like the music, and a CD at 16 bit 44.1 with no compression is still far superior to MP3 or AAC compressed. If I need some of my music on a portable device, I will do my own conversion to AAC at much higher bit rates to minimize artifacts. So I got my new CD, unwrapped it and cued up a song I have been waiting for. The opening is nice and airy and sounds wonderful, but then it comes to the big corus, and OUCH, that sounds harsh. I had to dig out my old analog oscilloscope and look at the outputs from my Sony CDP-9991 ES CD player. This thing may be old, but it has a wonderful D to A converter and a super clean analog section. The waveform during the beautiful open sounding intro is a smooth flowing wave with clearly visible harmonics moving across the trace. When I heard it get harsh, sure enough, hard clipping was clearly visible on my scope. The 16 bit audio data it slamming into the limits. I tried a few other tracks on the same disc, and what I found was just a crime. The quietest passages during any song were maybe 40 db below clipping, and that was rare. Most of it was less than 20 db down RMS, and the chorus sections were all hitting clip at a shocking rate. When a digital signal hits clip like this, all of the details in the sound go to a flat line until the signal comes back off the rail. I can't believe any recording engineer would sell a music recording like this. I put that disk away and pulled out another one. WHAT!! it was recorded the same way. All of the louder passages were going into flat line hard clipping. These were different artists and different studios making the same horrible recordings. The truly sad pat is, the 16 bit 44.1 medium was not the direct cause of this problem. I have many other CD's that are recorded well and do not suffer any of this, and I pulled a few out to check on my scope. Amazingly, the lowest levels on the other recordings was not even 10 db lower, but the loud chorus sections never touched the clip voltage of the newer recordings. They left a tick of headroom, and it paid off in much clearer sound, and I just had to turn the volum up 4-5 db to have the average of the music sound the same level. Background noise was never an issue, but I can hear the artist inhale before singing once in a while.

Why are the levels slammed up to near and past clipping?? There are a ton of arguments on this, but most seem to stem from wanting to be louder when played on the radio. Had these ben recorded at 24 bit, they most likely would have still slammed them to near clip, but hopefully they would not do that. I have 2 SACD recordings and one DVD audio one, not for a lack of trying though, I would love to get more music in hi res. But all 3 of my high res recordings have plenty of headroom and never come close o the clip limits. With the signal to noise ratio available at 24 bits, it is very common to record with the reference level at 20 db below clipping to give headroom.

I work in high end cinema audio and -20 dbfs is the norm for all of our content. Drop in a cd, and it is 15 db louder if we play it with the same peak full bit level. There is just no reason for this.

SO for me, the 16/20/24 bit argument is in hopes that engineers will leave some room and not clip the track. With just 16 bits, it can be a balancing act to keep the audio in the sweet spot. 24 bits makes it very roomy to make an excellent recoring. I have actually taken analog LP's and digitized them, and using 24 bits makes it so even the hardest crescendo does not clip with the groove noise of the record still modulating a few bits.

On to sample rate....
My hearing is not as good as it used to be, so I can't profes to hear every nuance of a waveform in the top octaves any more, but I can still easilly tell the difference between a sine wave, saw tooth wave, and square wave. All at the same base frequency, the difference is the levels of all the harmonics. At 44.1 khz sampling with a basic anti aliasing filter, the harmonics get tapered off and do certainly effect the wave shpe of non sine frequencies in the upper ranges. Even a 4 khz triangle wave will start to turn back into a sine wave when sampled and restored at 44.1 khz. Even a low frequency, like a drum hit can have some very high harmonics. And then there is the timing. Did the A to D converter sample just before the hit or just after it? We are talking a tiny slice of time, but the hard fast leading edge of some percussive sounds can get blurred at too low of a sample rate.

Higher frequency percussive sounds can be very telling. The clinks of a glockenspiel have been the true acid test for digital recording as they have harmonics that go well out past 20 khz. But non musical sounds are what really showed me that 48 khz is actually missing something. I have a recording of ceramic tiles falling to the ground and breaking. This is not a musical or even pleasant sound, but it does certainly have wide dynamics and frequency range. Hearing the recording at 24 bit 48 khz it sounded very real with a distinctive crack noise, but then they played the same thing recorded at 96 khz sample rate. All I can say is WOW!!! instead of hearing a pair of speakers play the sound of cracking tiles, I now heard ceramic tiles cracking on the floor between the speakers. The difference was far more dramatic than I ever could have thought. Going back to the 48 khz recording sounded almost dull now. The level measured exactly the same, and the frequency response on an FFT analyzer up to 20 khz was also identical within the resolution of the test gear. It was the attack speed of the cracks and the upper harmonics that just brought out the realism.

While it is certainly true that very little music could not be fit into 16 bit 44.1 k digital carrier without losing any audible detail, it is also very true that better recording techniques and higher resolution can absolutely result in a better sonic experience. The medium carrying the recording should not be the limiting factor of what we hear.

If we really wanted to cut to the lowest data rate to truly reproduce music, I am surprised noone ever tried something around 20 bits at 60 khz sample rate. That should cover 99% of everything out in the real world. But 24 96 is just fine with me as is SACD.
billmich and steveting99 like this.
GXMnow is offline  
Old 02-20-2016, 08:08 AM
Newbie
 
dogin's Avatar
 
Join Date: Apr 2010
Posts: 9
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
GXMnow----Thank You for that evaluation
dogin is offline  
Sponsored Links
Advertisement
 
 
taboola here
Thread Tools


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