• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Mark Waldrep's listening tests

Prep74

Member
Joined
Mar 22, 2020
Messages
59
Likes
137
The results of Mark Waldrep's "HD Audio Challenge" are live.

Mark Waldrep has always had respect, both in terms of the quality of his recordings and his generally "no nonsense", factual approach to marketing. In the years past he did believe that hi res did offer a slight audible benefit over red book CD. As he states, " I was convinced that high-resolution recording — real HD-Audio — would be perceptible." and "I was among the strongest advocates for this new and exciting “upgrade” to audio reproduction." However, a few years ago he (reliably/controlled) tested HD vs CD and kudos to him, he actually publicly admitted he couldn't distinguish them.

Now he's gone a step further, tested hundreds of others and states: "The hundreds of people that have participated in the second round of the HD-Audio Survey, have confirmed the results of the previous project. It is no longer possible to claim that “hi-res audio” is an important next step in the evolution of audio. HD-Audio is completely unnecessary for the reproduction of hi-fidelity."

Amazing considering his business model is based on hi res recordings. However, it should demonstrate that it is the actual quality of his recordings and production which makes his material sound great, ie skills and experience rather than choice of media.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
I agree with his conclusions. It is slightly possible hi-res would for a few people on a very few recordings be different. But at best that difference is small and not very significant. And for most people there is no difference. If the differences were as large as claimed we'd have figured that out by now. The fact it is even arguable either way means hires is somewhere between meaningless and the tiniest sliver of any difference for anyone. Not like the hires version of something will greatly enhance how its sounds.
 
OP
P

Prep74

Member
Joined
Mar 22, 2020
Messages
59
Likes
137
Yep, if it was conclusive we would not still be debating it 30 odd years on since hi res first appeared as SACDs. The most salient point he makes is that most of what passes as hi res came from analog tape or CD masters which are both 'standard res'. Transferring these recordings to a larger bucket doesn't change their provenance. The thing I like about the test he ran is that the files were true hi res from recording to production and yet the overwhelming majority could either not pick a difference or were just guessing.
 

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
Yep, if it was conclusive we would not still be debating it 30 odd years on since hi res first appeared as SACDs. The most salient point he makes is that most of what passes as hi res came from analog tape or CD masters which are both 'standard res'. Transferring these recordings to a larger bucket doesn't change their provenance. The thing I like about the test he ran is that the files were true hi res from recording to production and yet the overwhelming majority could either not pick a difference or were just guessing.
This whole thing about Hi-res is confused, often distorted by different mastering in a high res vs 'normal' version. The worst thing about high-res is that it is a more 'accurate' version of the same noticeably mis-mastered materai.

I have run some tests for some friends, where I give them a 44.1k/16 version, and it can be better/cleaner than the high-res version. This isn't because of my 'brilliant' creation of a special 44.1k/16 version, but instead it is a difference in mastering.

The biggest bang-for-the-buck is proper mastering first, then lets talk about resolution next. There is typically very significant damage in the very audible regions of the spectrum -- fix that first, then we can talk about parts of the signal above 15kHz or artifacts below -85dB once the mastering is fixed.

John
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
This whole thing about Hi-res is confused, often distorted by different mastering in a high res vs 'normal' version. The worst thing about high-res is that it is a more 'accurate' version of the same noticeably mis-mastered materai.

I have run some tests for some friends, where I give them a 44.1k/16 version, and it can be better/cleaner than the high-res version. This isn't because of my 'brilliant' creation of a special 44.1k/16 version, but instead it is a difference in mastering.

The biggest bang-for-the-buck is proper mastering first, then lets talk about resolution next. There is typically very significant damage in the very audible regions of the spectrum -- fix that first, then we can talk about parts of the signal above 15kHz or artifacts below -85dB once the mastering is fixed.

John
I would agree although I think one of the most misunderstood parts of pro audio is the signal chain, whether digital or analog or mixed. Setting it up for maximum SNR and resolution from beginning to end is confused with editing for sound. I don't think there will be much progress until those are made distinct.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
Yep, if it was conclusive we would not still be debating it 30 odd years on since hi res first appeared as SACDs. The most salient point he makes is that most of what passes as hi res came from analog tape or CD masters which are both 'standard res'. Transferring these recordings to a larger bucket doesn't change their provenance. The thing I like about the test he ran is that the files were true hi res from recording to production and yet the overwhelming majority could either not pick a difference or were just guessing.
Note that some people still debate on vinyl sounding better than CD though ;-)
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
I agree with his conclusions. It is slightly possible hi-res would for a few people on a very few recordings be different. But at best that difference is small and not very significant. And for most people there is no difference. If the differences were as large as claimed we'd have figured that out by now. The fact it is even arguable either way means hires is somewhere between meaningless and the tiniest sliver of any difference for anyone. Not like the hires version of something will greatly enhance how its sounds.
The only thing I would add to your comment is that there are good archival reasons (taking audibility aside, in the same way you would preserve an obscure book in a library) for recording and rendering at 96/24 PCM. Full capture of HF transients and full DR without truncation.

I don't see any benefit to go beyond that. The current proliferation of formats and streaming harms the archival side the most IMO.
 

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
Note that some people still debate on vinyl sounding better than CD though ;-)
A lot of people still don't realize that the original 'nasty' CD sound was a mastering difference, in fact, a very pervasive, systemic mastering difference. As far as a non-technical person might reason, that difference in sound is due to digital vs. vinyl, which it is NOT. Except for minor defects, given ZERO mastering differences, vinyl doesn't really sound all that different from digital/CD. The average person just doens't get an opportunity to control the comparison. As far as the consumer goes, a carpenters 'A Song for you' is the same whether on CD or vinyl, but that is CERTAINLY not true. We were not sophisticated enough at the time to fully research EXACTLY the differences, so the intellectual property owners of the recordings don't need to distribute high quality versions of their 'family jewels', and the general llistening public is none-the-wiser.

The mastering has been the big variable all along. Back in the late 1980s/early '90s, I just gave up, couldn't get a good sounding CD (by my standards.) All along, they had been adding a special layer of 'love' onto almost every CD sold. BTW, I am no 'golden ears', but simply a Bell Labs engineer that realized something perverse was going on -- now we know EXACTLY the troubles.

John
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,495
Agree with everyone here. The case for 24 or 32 bits of depth and 88.2k or 96k sample rates seems quite obvious from the point of view of recording, processing, mixing, mastering, and archiving. But for the end product, properly produced 16/44.1k redbook is perfectly adequate.

One could argue that 24/48k would be nice for the end product as it's a little more fool-proof: the 24 bit depth means you don't need to worry about what kind of dither (if any) was applied during final processing, as even undithered 24-bit will get you a S/N ratio beneath the threshold of human hearing. And a 48kHz sample rate gives you about twice as much wiggle room between the upper limit of human hearing and the Nyquist limit, so if your playback system has a somewhat slow/leaky reconstruction filter you still won't get any aliasing in the audible range.

But if we're talking properly produced digital sources and well-engineered equipment, redbook should be indistinguishable from higher-res formats.

One final note is that potential (albeit very small) differences between different resolution versions of the identical source can come from the format conversion itself. To compare a 24/96 digital master with the redbook version of the same master, you have to do sample-rate conversion, and you have to reduce the bit depth, usually with dither. If you load up both versions in an audio editor, invert one, and then run a null test on them, the "difference file" will be small, but the difference will sound a lot more like music than one would guess.

In that case there are real differences, regardless of whether or not they're consistently audible in a blind test. But the key issue is that while they are different there is no objective or experimentally repeatable basis upon which to claim that one is better or worse than the other. Technically one could say the 24/96 version is "higher fidelity" but that's pedantic: it's higher fidelity to the source only because it is the source, and you have to change the data in order to convert it to 16/44.1k.
 
Last edited:

Midwest Blade

Senior Member
Joined
May 8, 2019
Messages
401
Likes
535
Note that some people still debate on vinyl sounding better than CD though ;-)

Yep...I had one of those moments at Axpona a few years ago when a few turntables came out with USB connections for digital recording. The representative played the album and then a hi-res recording of it...clicks and all. I thought it was hilarious as I exited and headed to the next room. Vinyl...so hot! :rolleyes:
 

CDMC

Major Contributor
Forum Donor
Joined
Sep 4, 2019
Messages
1,172
Likes
2,321
I would gladly trade all of my redbook and hi rez for 320mp3 or Apple AAC well mixed and produced.
 

GXAlan

Major Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
3,868
Likes
5,954
As a counter point though, if you were running a Marantz AVR with the HF roll off due to the fixed slow digital filter, high resolution audio would give you a better response by eliminating the effect of that poor digital filter. Many people no longer have hearing that extends that high, but I suspect that is one reason where high resolution audio is more consistent.

Second, once mastered in high resolution, the down sampling to 44.1 can leverage dithering or Sony SBM or JVC K2 type algorithms, which would be superior to recording at 44.1 from the get go with a brick wall filter.
 

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
I would gladly trade all of my redbook and hi rez for 320mp3 or Apple AAC well mixed and produced.

I have rather completely, reverse-engineered the DRC compression on almost every CD made (except a few really good esoteric ones, and some MFSL, etc.)

Much of the time, on older CDs, the specific kind of damage is approx this (in sox lanuage):

1) Start with DolbyA tape (compressed in mid/side instead of left/right).
2) Apply this eq: gain -6 bass 3 1k 0.50q treble 10 1k 0.50q treble -10 3k 0.50q treble -6 3k 0.50q treble 6 9k 0.50q

The result will sound similar to the original input to the DolbyA, but will be multi-band compressed.
This is VERY VERY similar to the compression done to CDs since the middle 1980s.

The actual EQ used on CDs is a little different than the simple depiction above, and the DolbyA is usually used where
the ch1=mid and ch2=2*side The actual EQ is more based on 1st order EQ filters instead of 2nd order.

My free-for-consumers software decoding SW is super accurate & capable of decoding DolbyA tapes in commercial mode, but the consumer mode has the EQ built-in (optionally enabled) that can reverse this ugly compression done on ALMOST every CD made. It really works, and I just FINALLY reverse engineered the EQ rather precisely. (The decoding of raw DolbyA tapes is astonishingly good -- recordings revealed the first time since being recorded onto tape.)

Some recordings actually have two passes done to them (Al Stewart's Year of the Cat), and some three passes (ABBA Dreamworld.) The decoder is stable/accurate enough to do a good job at two passes, but three passes are still a challenge. I doubt that a true DolbyA could even do two passes reasonably well.

MOST recordings are only processed in one pass, and the results can be VERY engaging.
The sw is available for downloading, is available for Linux and WIndows. It is primitive command line, but the internal algorithms are far far from primitive, for sure!!!

John
 

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
I have rather completely, reverse-engineered the DRC compression on almost every CD made (except a few really good esoteric ones, and some MFSL, etc.)

Much of the time, on older CDs, the specific kind of damage is approx this (in sox lanuage):

1) Start with DolbyA tape (compressed in mid/side instead of left/right).
2) Apply this eq: gain -6 bass 3 1k 0.50q treble 10 1k 0.50q treble -10 3k 0.50q treble -6 3k 0.50q treble 6 9k 0.50q

The result will sound similar to the original input to the DolbyA, but will be multi-band compressed.
This is VERY VERY similar to the compression done to CDs since the middle 1980s.

The actual EQ used on CDs is a little different than the simple depiction above, and the DolbyA is usually used where
the ch1=mid and ch2=2*side The actual EQ is more based on 1st order EQ filters instead of 2nd order.

My free-for-consumers software decoding SW is super accurate & capable of decoding DolbyA tapes in commercial mode, but the consumer mode has the EQ built-in (optionally enabled) that can reverse this ugly compression done on ALMOST every CD made. It really works, and I just FINALLY reverse engineered the EQ rather precisely. (The decoding of raw DolbyA tapes is astonishingly good -- recordings revealed the first time since being recorded onto tape.)

Some recordings actually have two passes done to them (Al Stewart's Year of the Cat), and some three passes (ABBA Dreamworld.) The decoder is stable/accurate enough to do a good job at two passes, but three passes are still a challenge. I doubt that a true DolbyA could even do two passes reasonably well.

MOST recordings are only processed in one pass, and the results can be VERY engaging.
The sw is available for downloading, is available for Linux and WIndows. It is primitive command line, but the internal algorithms are far far from primitive, for sure!!!

John
After a lot of testing, and detailed improvement in the DHNRDS decoder, I have found that many recordings (perhaps most) ARE 2 passes!!!

A group of friends and I have been playing with some Telarc disks -- LOTS of dynamic range, in fact -- TOO MUCH. However, when decoding with the normal gain calculation (gain * signal) like a true DolbyA would, the results are kind of sloppy. With the anti-MD enabled, the recording is at least mostly restored.

To do two passes at audiophile standards, (the decoder) required a lot of improvement -- the dynamics have to match PERFECTLY, and some of my lazy techniques had to go away. Both the Hilbert detectors and the modulation distortion reduction have been massively improved. Bottom line, two passes is a b*tch.

The decoder can now do the decodes at high quality, but two explict passes is too ugly to use, so I am working on doing the two passes built-in as an option (so, it will support 2 pass FA, in addition to the already existent 1 pass FA, and the now further improved DA.)

John
 
OP
P

Prep74

Member
Joined
Mar 22, 2020
Messages
59
Likes
137
Note that some people still debate on vinyl sounding better than CD though ;-)
Yes but that is subjective to an individual and subjectively vinyl sometimes can sound better than a CD or hi res version of a particular album due to mastering and production choices. Objectively though, CD is a higher fidelity format than vinyl.
 
Top Bottom