Help Determine if High Definition is better than CD

[LHC]

Hi Everyone. For your awareness Dr Mark Waldrep has developed a 'blind' test for people to compare a true HD music file mastered in 96 kHz/24-bit PCM, and various carefully down-sampled versions of this file. One could download these music files from this website https://www.realhd-audio.com/?p=6713 to conduct your own listening test and comparison. 

 

If you wish to report the results of your own listening test and have it count towards Dr Waldrep's survey of results, please contact him before the END of MARCH. That is when the listening challenge will finally close. He said he has received 200 result already, but are withholding announcing them until the end of March. 

 

Please do not write about your comparison result here to make it fair for others who wish to try it on their own. We can certainly discuss the outcome and interpretation once all results become available later. 

 

 

[Decky]

Of course they are better - because they are High Definition (HD) compared to CD which is Crappy Definition. 

Edited March 6, 2020 by Decky

[Steffen]

You'd need some blind A/B switching software for this to be useful. Someone mentioned a command line tool for this on SNA not long ago, does anybody remember what it was?

[Guest]

Irrelevant. 
 

Everyone knows that mastering rules. 

Edited March 7, 2020 by Sime

[LHC]

26 minutes ago, Sime said:

Irrelevant. 
 

Everyone knows that’s mastering rules. 

 

Dr Waldrep knows better than anyone about the importance of recording and mastering. He operates his own studio for years, makes his own recordings, and lectured much on their importance. He made his views known clearly in this heart-felt article https://audiophilereview.com/audiophile-news/the-truth-about-high-resolution-audio-facts-fiction-and-findings.html

 

I quote: "My enthusiasm for high-resolution has diminished in recent years. After reading numerous studies and articles on the topic, I'm inclined to agree with those that believe Redbook CDs are sufficient to capture all of the fidelity we need when listening to recorded music. I do believe that audio engineers reap tremendous rewards by using high-resolution specifications during recording sessions but that doesn't mean that delivering 96 kHz/24-bit WAV or FLAC files to consumers makes any perceptible difference in the final listening at home. I understand that this position goes against the industry position. In fact, it rejects 20 years of my own thinking. I've spent millions of dollars and thousands of hours producing and releasing over 75 real high-resolution albums. Did I waste all that money and time?

 

What could possibly bring me to reverse my own strongly held position? It was a study - a survey that is happening right now. As part of a sabbatical I was granted during the fall 2019 semester from my college professor gig, I'm conducting a survey open to all music listeners. Those who sign up are provided access to 40 full length, downloadable music files."

Edited March 7, 2020 by LHC

[metal beat]

try complete definition - vinyl 

[muon*]

6 minutes ago, LHC said:

 

Dr Waldrep knows better than anyone about the importance of recording and mastering. He operates his own studio for years, makes his own recordings, and lectured much on their importance. He made his views known clearly in this heart-felt article https://audiophilereview.com/audiophile-news/the-truth-about-high-resolution-audio-facts-fiction-and-findings.html

 

I quote: "My enthusiasm for high-resolution has diminished in recent years. After reading numerous studies and articles on the topic, I'm inclined to agree with those that believe Redbook CDs are sufficient to capture all of the fidelity we need when listening to recorded music. I do believe that audio engineers reap tremendous rewards by using high-resolution specifications during recording sessions but that doesn't mean that delivering 96 kHz/24-bit WAV or FLAC files to consumers makes any perceptible difference in the final listening at home. I understand that this position goes against the industry position. In fact, it rejects 20 years of my own thinking. I've spent millions of dollars and thousands of hours producing and releasing over 75 real high-resolution albums. Did I waste all that money and time?

 

What could possibly bring me to reverse my own strongly held position? It was a study - a survey that is happening right now. As part of a sabbatical I was granted during the fall 2019 semester from my college professor gig, I'm conducting a survey open to all music listeners. Those who sign up are provided access to 40 full length, downloadable music files."

Who would have thought I'd agree with the  Waldrep on something.

[Satanica]

32 minutes ago, Sime said:

Everyone knows that’s mastering rules. 

If only.

[b0dhi]

A/B testing of audio close to perceptual limits is pointless for many reasons, but mainly because consciously recallable auditory memory is far too short lived and information-sparse, and A/B tests have no resistance whatsoever to Type 2 error. Blinded long term listening would work better, but is also very impractical. On top of that, the way recorded audio is currently done is fairly primitive compared to how a human being's hearing actually works in reality (in-situ). So, a "challenge" like this is like "challenging" someone to tell the difference between 2 similar paintings flashed before your eyes for half a second each through a brown muddy window.

[LHC]

6 hours ago, b0dhi said:

A/B testing of audio close to perceptual limits is pointless for many reasons, but mainly because consciously recallable auditory memory is far too short lived and information-sparse, and A/B tests have no resistance whatsoever to Type 2 error. Blinded long term listening would work better, but is also very impractical. On top of that, the way recorded audio is currently done is fairly primitive compared to how a human being's hearing actually works in reality (in-situ). So, a "challenge" like this is like "challenging" someone to tell the difference between 2 similar paintings flashed before your eyes for half a second each through a brown muddy window.

Your concerns are valid. Dr Waldrep does not specify how one should go about doing the comparison, whether one choose to use rapid A/B test, or take one's time to listen and compare over a long time, it is up to the individual. The only thing required is that one do not attempt to analyse the files (to break the 'blind' aspect), and to ensure one's system does not down-convert the files during playback. He is aware of the limitations and deficiency in rigour of this challenge. IMO I think in this case some evidence is better than no evidence, and some dirty data is better than no data. 

[muon*]

As my system is 16bit 44.1kHz non oversampled (It's all I need) I likely can't replay the high res files............ so won't be taking part

 

Edit: dirty data better than no data? I disagree, but that's just my view.

Edited March 9, 2020 by muon*

[lemarquis]

On 07/03/2020 at 12:16 PM, LHC said:

I'm inclined to agree with those that believe Redbook CDs are sufficient to capture all of the fidelity we need when listening to recorded music.

+1 for this. Although this presumes playback on a good enough CD player. It has certainly been my experience that physical CDs on a very good playback system can sound as good and often better than hi-res files. Some of that also has to do with poor (re)mastering of some of the said hi-res files.

 

After employing some tweaks to the Halcro system my CDs have never sounded so analogue. To the point that I regularly swap between CD sessions and DSD sessions with equal enjoyment.

 

Might try the test anyway

[MattyW]

On 07/03/2020 at 10:49 AM, Sime said:

Irrelevant. 
 

Everyone knows that mastering rules. 

 

I'm in complete agreement.  I did try high res briefly though have since found 16 bit 44khz superior through a well implemented Phillips TDA1541 DAC.

[eltech]

I registered to do the test. I downloaded the files.

I honestly performed the test.

I got my results today.

I got 60% correct.

And 40% wrong. 

Edited May 2, 2020 by eltech

[LHC]

The preliminary results are in! You can read about it in this link: http://www.realhd-audio.com/?p=6993

 

"Hi-Res Audio or HD-Audio provides no perceptible fidelity improvement over a standard-resolution CD or file." 

 

That is just a short summary and take-away; please read the whole article for the full picture. 

 

 

[Steffen]

Seems like a bad time to be out of popcorn...

[Guest DrSK]

Is there any more detail beyond the link? 

 

On average his results are probably correct. What would be interesting is whether any individuals reliably kept getting it right for reasons other than chance. 

 

Conclusions on averages can be misleading without full control and understanding of all independent variables. Often data outliers and the tails of distributions hint at other independent variables not accounted for. 

 

Economists get stuffed up by this all the time. I've seen this cost $millions in dud advice, but they don't like listening to engineers. 

 

So is some of the recent dyslexia research that concludes some long standing assistive aids provide no benefit. They didn't on average but for a percentage, mainly high functioning dyslexics, they make enough of a difference to fully function at work vs not functioning. Yet the current professional medical opinion is it doesn't make a difference based on research reporting the average result. As a result some people that could benefit don't use them. 

 

One economist did a paid assessment concluding that adding cyclists to a road where traffic averaged 25km/h would have no impact as bikes travelled at 25km/h. However the traffic average was made up from a top speed of 50km/h and time spent doing 0km/h at traffic lights. In reality adding bikes would cap maximum speed of motorists to 25km/h dropping their average to 12.5km/h or there abouts.

 

Same with State funded solar feed in tariff cost estimates. Economist took average solar panel solar generation and average home consumption and took the difference. Not considering that the peak solar generation and peak home consumption occur at different times. 

 

Averages can be a dangerous thing to believe. 

Edited July 8, 2020 by DrSK

[Steffen]

9 hours ago, DrSK said:

Averages can be a dangerous thing to believe. 

On average, that’s true.

[davewantsmoore]

10 hours ago, DrSK said:

Averages can be a dangerous thing to believe. 

Indeed.  Statstics can be very very misleading.

 

They need to be partnered with the WHY behind it.   eg. statistics from the experiment will indicate whether or not people can hear/perceive things (eg. "hires" vs CD) .....  much more interesting than this is the explanation about how it works.    Did we expect them not to be able to hear it?  Why can't they hear it?  or vice-versa ... What would make them show a different result? ..... etc.

 

For example, measuring the force of gravity at different locations, is one thing....   being able to explain why it changes, and what we should expect, is the real important bits.

 

 

Very few people in the audio industry want to explain to consumers that 44khz and 16bits is "enough" ..... because the "why" is complicated and has a lot of caveats  (ie. you can make a mess of 44/16) .... and it kills the marketing of "hi res".

[Irek]

I believe in A/B test with certain rules and those rules are there are no rules. I was listening to 20 secounds fragments multiple times for 1 hour. PC, laptops are the best for testing , me think.

 

[jeromelang]

 

Comparisons are best done through a stand-alone streaming device. Not a PC based player.

Main issue is memory retention. On PC this is a serious problem affecting playback sound quality.

On a stand-alone streamer device, the problem is much less acute if users prepare the following:

 

(1) Place the 2 files being compared in 2 separate folders.

Placing them in 1 single folder means that the 2 files will be played in sequence, one after another.

The subsequent file will always sound closed-in, dulled, dynamically compressed.

This is the adverse effect of memory retention by the streamer device.

But, placing the 2 files in 2 separate folders means that users will have to exit that current folder, going into another folder to select and play the other file being compared to.

This procedure apparently has an effect of refreshing playback memory and allow newly selected files to play with unrestrained vitality. 

In this condition, users will be able to hear more sonic differences between files of the same recording/mastering with differing resolution.

 

(2) Refreshing the external dac

Users using an external dac with stand-alone stream needs to be aware that if they are currently on another input on the dac, and when they are switching over to the input to receive audio from a streamer, they need to power down/power up the dac again, so that the dac "sees" the incoming data stream from the stand-alone freshly for the first time. Do this when the streamer has finished playing the last song on an album/folder and has stopped. Sounds better this way.

 

Step by step:

 

The dac is currently on another input (coaxial, optical, etc)

You switch over to USB (or coastal or opt) to receive audio from the streamer.

You select a song (not the song you actually want to play) to play on the streamer.

Usually streamers somehow sound a bit lethargic and closed in.

You choose the last song on this album. You let it play till that last song ends and the streamer stops operation. (For this purpose I always choose an album that has a very short last track.)

You shut down the dac.

Then power up the dac again.

The dac powers up and settles on the same input and freshly "sees" the incoming stream for the first time

You play your desired song (from another album/folder).

Always choose another song from another album/folder so that the streamer erases memory of what has been played before and starts with a clean slate. It will sound fresh, dynamic, and spacious.

Coupled with the power refreshing of the dac, you will hear the most detailed, the most dynamic and spacious reproduction of what your streamer and dac are capable of.

 

And if the user is powering up both the streamer and the external dac for the first time (knowing that the DAC was last powered off while switched to the input receiving the streamer) there is always a proper sequence - first the streamer, then the dac.   

 

 

 

 

Edited July 9, 2020 by jeromelang

[davewantsmoore]

7 hours ago, Irek said:

I believe in A/B test

It depends on what question you are trying to answer, as to what sort of test, and how it should be structured.

 

People often (mistakenly) use a  "can you hear a difference between A and B" type of test ....  to answer a "which do you like better A or B" type of question..... and then they actually report the results as "A IS BETTER THAN B!!!  It must be true, I tested it".      

[davewantsmoore]

5 hours ago, jeromelang said:

 

Comparisons are best done through a stand-alone streaming device. Not a PC based player.

Main issue is memory retention. On PC this is a serious problem affecting playback sound quality.

On a stand-alone streamer device, the problem is much less acute if users prepare the following:

 

(1) Place the 2 files being compared in 2 separate folders.

Placing them in 1 single folder means that the 2 files will be played in sequence, one after another.

The subsequent file will always sound closed-in, dulled, dynamically compressed.

This is the adverse effect of memory retention by the streamer device.

But, placing the 2 files in 2 separate folders means that users will have to exit that current folder, going into another folder to select and play the other file being compared to.

This procedure apparently has an effect of refreshing playback memory and allow newly selected files to play with unrestrained vitality. 

This is such a load of utter nonsense.

 

 

 

[Ittaku]

21 minutes ago, davewantsmoore said:

This is such a load of utter nonsense.

 

 

 

You really should try judicious use of the ignore feature.

[Irek]

1 hour ago, davewantsmoore said:

It depends on what question you are trying to answer, as to what sort of test, and how it should be structured.

 

People often (mistakenly) use a  "can you hear a difference between A and B" type of test ....  to answer a "which do you like better A or B" type of question..... and then they actually report the results as "A IS BETTER THAN B!!!  It must be true, I tested it".      

100% correct. 

Very often it's hard to distinguish the difference between personal preferences and "better".