Canuck Audio Mart Hifi and Audio ForumCanuck Audio Mart Hifi and Audio Forum
It is currently Sat Dec 16, 2017 12:53 pm

All times are UTC - 8 hours




Post new topic Reply to topic  [ 24 posts ]  Go to page 1, 2  Next
Author Message
PostPosted: Tue Nov 21, 2017 5:00 pm 
Offline

Joined: Tue Dec 17, 2013 7:36 pm
Posts: 128
Location: toronto, ON, CA
With all the ongoing discussion about high res digital files sounding superior and many audiophiles and manufacturers jumping on the high res bandwagon, I decided to make a little experiment and hear for myself the quality of high res.

I found a good source of free/public-domain high res files. See this link: http://www.2l.no/hires/index.html for description of the performance and the recordings.

From the above source I took the piano concerto: Ola Gjeilo: Ubi Caritas - piano improvisation Ola Gjeilo - 24bit/96Khz version. Direct link below

http://www.lindberg.no/hires/test/2L-082_stereo-96kHz_01.flac

I took the first 60 seconds and re-sampled them to 16bit/11025Khz - the sample rate is 4x smaller than redbook 44.1khz.

For the benefit of the community, you can hear the differences at the link below:

https://goo.gl/WJPT7Y

Let me know which one you prefer!


Top
 Profile  
Reply with quote  
PostPosted: Tue Nov 21, 2017 6:10 pm 
Offline

Joined: Mon Apr 11, 2011 9:42 pm
Posts: 3
Location: NL, CA
Why 11khz and what did you resample with?


Top
 Profile  
Reply with quote  
PostPosted: Tue Nov 21, 2017 6:44 pm 
Offline

Joined: Tue Dec 17, 2013 7:36 pm
Posts: 128
Location: toronto, ON, CA
FloconsDeMais wrote:
Why 11khz and what did you resample with?


11KHz was chosen for the following reasons.

1 - It is 1/4 the sampling rate of redbook and 1/8 of the high res file - any superiority of the high res file should be night and day obvious
2 - It should be enough digitize piano music.

Despite point 2, I have read a ton of reviews of high res audio where piano is specifically touted as the prime example of the need for high resolution - i.e. note decay, blackness, etc,etc....

All digital processing was done with audacity on linux.


Last edited by zsss on Tue Nov 21, 2017 6:59 pm, edited 1 time in total.

Top
 Profile  
Reply with quote  
PostPosted: Tue Nov 21, 2017 6:50 pm 
Offline
User avatar

Joined: Fri Apr 01, 2005 7:15 am
Posts: 4599
Location: Newmarket, ON, CA
a 44.1 sample would be nice, just saying


Top
 Profile  
Reply with quote  
PostPosted: Tue Nov 21, 2017 7:05 pm 
Offline
User avatar

Joined: Mon Oct 17, 2011 6:55 pm
Posts: 1946
Location: Toronto, ON, CA
zsss wrote:
FloconsDeMais wrote:
Why 11khz and what did you resample with?


11KHz was chosen for the following reasons.

1 - It is 1/4 the sampling rate of redbook and 1/8 of the high res file - any superiority of the high res file should be night and day obvious
2 - It should be enough digitize piano music.

Despite point 2, I have read a ton of reviews of high res audio where piano is specifically touted as the prime example of the need for high resolution - i.e. note decay, blackness, etc,etc....

All digital processing was done with audacity on linux.


That is assuming information was lost during the resample. No information was lost with your experiment. Invert one file then mix and render the 2 files together. The result should be the difference between the files.


Top
 Profile  
Reply with quote  
PostPosted: Tue Nov 21, 2017 9:52 pm 
Offline

Joined: Tue Sep 25, 2012 1:54 pm
Posts: 363
Location: Edmonton, AB, CA
It all depends on the master...


Top
 Profile  
Reply with quote  
PostPosted: Tue Nov 21, 2017 9:59 pm 
Offline

Joined: Mon Apr 11, 2011 9:42 pm
Posts: 3
Location: NL, CA
zsss wrote:
FloconsDeMais wrote:
Why 11khz and what did you resample with?


11KHz was chosen for the following reasons.

1 - It is 1/4 the sampling rate of redbook and 1/8 of the high res file - any superiority of the high res file should be night and day obvious
2 - It should be enough digitize piano music.

Despite point 2, I have read a ton of reviews of high res audio where piano is specifically touted as the prime example of the need for high resolution - i.e. note decay, blackness, etc,etc....

All digital processing was done with audacity on linux.



Comparing 24/96 to 16/11 is bordering on pointing out the obvious. It's like comparing flac with 96kbps MP3. A comparision with 16/44.1 would be more salient and reflective of the choices in mainstream music reproduction and distribution, and would be more of a challenge to distinguish.

Also, SOX or SSRC would be better choices for downsampling.


Top
 Profile  
Reply with quote  
PostPosted: Tue Nov 21, 2017 10:12 pm 
Offline

Joined: Tue Jul 24, 2007 10:54 am
Posts: 214
Location: Vancouver, BC, CA
msommers wrote:
It all depends on the master...


I believe OP is using same master.


Top
 Profile  
Reply with quote  
PostPosted: Wed Nov 22, 2017 5:07 am 
Offline

Joined: Thu Jan 13, 2011 11:29 am
Posts: 501
Location: St.Catharines, ON, CA
Did you use dither?

Gary


Top
 Profile  
Reply with quote  
PostPosted: Wed Nov 22, 2017 5:43 am 
Offline

Joined: Tue Dec 17, 2013 7:36 pm
Posts: 128
Location: toronto, ON, CA
FloconsDeMais wrote:
zsss wrote:
FloconsDeMais wrote:
Why 11khz and what did you resample with?


11KHz was chosen for the following reasons.

1 - It is 1/4 the sampling rate of redbook and 1/8 of the high res file - any superiority of the high res file should be night and day obvious
2 - It should be enough digitize piano music.

Despite point 2, I have read a ton of reviews of high res audio where piano is specifically touted as the prime example of the need for high resolution - i.e. note decay, blackness, etc,etc....

All digital processing was done with audacity on linux.



Comparing 24/96 to 16/11 is bordering on pointing out the obvious. It's like comparing flac with 96kbps MP3. A comparision with 16/44.1 would be more salient and reflective of the choices in mainstream music reproduction and distribution, and would be more of a challenge to distinguish.

Also, SOX or SSRC would be better choices for downsampling.


Trying to find the differences between 24/96 to 16/44 would be impossible as there would be no audible difference. 16/11 was chosen deliberately as en extreme case. I know there is some audible difference between the files and information was definitely lost in the 16/11 conversion. However, it is not as obvious as redbook vs 96kbs.. It is too early to see any trends in the survey results.


Top
 Profile  
Reply with quote  
PostPosted: Wed Nov 22, 2017 6:47 am 
Offline
User avatar

Joined: Mon Oct 17, 2011 6:55 pm
Posts: 1946
Location: Toronto, ON, CA
I stand corrected. Not sure what I measured last night but mixed and rendered the 2 files again. It is clear that one sample is missing information between about 5k and 15k....although not a lot of information was removed. And despite I am suffering from a cold, I can hear a difference between the 2 files.

The problem with tests such as this is if the subject is not familiar with the material, he/she may never be able to detect the differences. As quoted from Wiki,

Quote:
Prior to formal grading, subjects must be allowed to become thoroughly familiar with the test facilities, the test environment, the grading process, the grading scales and the methods of their use. Subjects should also become thoroughly familiar with the artefacts under study. For the most sensitive tests they should be exposed to all the material they will be grading later in the formal grading sessions.


These tests have been beaten to death – MP3 vs WAV vs Hi Res vs…… Testing a single sample (or even a small handful of samples) does not mean redbook will sound just as good as 24/96 for all cases. Or that one is unable to distinguish between MP3 vs WAV.


Top
 Profile  
Reply with quote  
PostPosted: Wed Nov 22, 2017 7:57 am 
Offline

Joined: Tue Dec 17, 2013 7:36 pm
Posts: 128
Location: toronto, ON, CA
sthomas1049 wrote:
I stand corrected. Not sure what I measured last night but mixed and rendered the 2 files again. It is clear that one sample is missing information between about 5k and 15k....although not a lot of information was removed. And despite I am suffering from a cold, I can hear a difference between the 2 files.

The problem with tests such as this is if the subject is not familiar with the material, he/she may never be able to detect the differences. As quoted from Wiki,

Quote:
Prior to formal grading, subjects must be allowed to become thoroughly familiar with the test facilities, the test environment, the grading process, the grading scales and the methods of their use. Subjects should also become thoroughly familiar with the artefacts under study. For the most sensitive tests they should be exposed to all the material they will be grading later in the formal grading sessions.


These tests have been beaten to death – MP3 vs WAV vs Hi Res vs…… Testing a single sample (or even a small handful of samples) does not mean redbook will sound just as good as 24/96 for all cases. Or that one is unable to distinguish between MP3 vs WAV.


Everything is beaten to death multiple times on this forum. 99.9% of the time it is anecdotal evidence of I told you so so it must be true or your system is not resolving enough.

I encourage everybody to download all the high res samples from the site quoted in the first post and do their own analysis of what is there. You can draw your own conclusions on marketing vs reality.


Top
 Profile  
Reply with quote  
PostPosted: Wed Nov 22, 2017 8:59 am 
Offline
User avatar

Joined: Fri Apr 01, 2005 7:15 am
Posts: 4599
Location: Newmarket, ON, CA
My DAC comes with an explicit warning against inputs at less than 44.1. Paraphrasing, from the admittedly broken English, it went something like ...

DAC may lose consentration [yeah I can't translate that] with too much time on hands, maybe start play around, play backward, gamble on internet, consequentials [I like that!] sound may be little unease.

... and on like that, but basically ... don't do that.

Great DAC for listening to CDs - Love it!


Top
 Profile  
Reply with quote  
PostPosted: Wed Nov 22, 2017 11:35 am 
Offline

Joined: Mon Apr 11, 2011 9:42 pm
Posts: 3
Location: NL, CA
sthomas1049 wrote:
That is assuming information was lost during the resample. No information was lost with your experiment. Invert one file then mix and render the 2 files together. The result should be the difference between the files.


He resampled one to 11khz. That's a 5.5khz lowpass - that is, everything over 5.5khz was lost. Not even LAME sets a lowpass that low for even its lowest bitrate.

-- 22 Nov 2017 19:45 --

zsss wrote:
Trying to find the differences between 24/96 to 16/44 would be impossible as there would be no audible difference. 16/11 was chosen deliberately as en extreme case.


That is the current debate - whether 24/96 offers any audible advantage over 16/44. It should be obvious that chopping off 3/4 of the audible range of a sample would be audible to most people. That's what makes this test pointless. :lol:


Top
 Profile  
Reply with quote  
PostPosted: Wed Nov 22, 2017 11:47 am 
Offline

Joined: Tue Dec 17, 2013 7:36 pm
Posts: 128
Location: toronto, ON, CA
FloconsDeMais wrote:
sthomas1049 wrote:
That is assuming information was lost during the resample. No information was lost with your experiment. Invert one file then mix and render the 2 files together. The result should be the difference between the files.


He resampled one to 11khz - that's a 5.5khz lowpass... that is, everything over 5.5khz was lost. Not even LAME sets a lowpass that low for even its lowest bitrate.


Yes, 5.5Khz low pass is very low and nothing uses such a low pass in general. However, the highest frequency of piano is 5587.65Hz (https://en.wikipedia.org/wiki/Piano_key_frequencies) so 11Khz sampling rate should be just enough to capture the recording.


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 24 posts ]  Go to page 1, 2  Next

All times are UTC - 8 hours


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group