Comparison: PCM DXD DSD (Sound Liaison High Res Format Comparison)

[amirm]

Yes, another older high-resolution music format comparison, this time with a different twist than the previous ones:

The tool, musicscope seems to be offered for free now although there is stipulation that you have to be an existing customer (?).

Edit: company responded and has redone the content to fix the issues in the video!!! https://www.audiosciencereview.com/...pare-formats-sampler-a-new-2-0-version.23274/

Much kudos to them.

 

[martijn86]

Thank you for educating us! I feel it is much needed in the enthousiasts HiFi community.

 

[MCH]

hello Amirm, all,
thank you for these very informative videos. I have a couple questions, from the point of view of an average music consumer:

Are the tracks you analize cherry picked or in your experience, these "irregular" quality is a general thing?

these were a few years ago, have you observed things have changed recently?

last one: ok, we can see things have not always been done properly, but what is the potential or the advantage of hi-res music? do you think that if artists with resources get involved and do it right, we consumers are going to have a better experince provided we have the right equipment?

thank you for your thouhts

 

[martijn86]

what is the potential or the advantage of hi-res music

I don't know if I'm correct but I personally feel like hi-res is a better source if you use DSP corrected hardware. I draw a comparison to photography* where a raw image and a JPG that is processed in camera look practically identical but once you start to push it in certain areas during processing you start to decrease the effective dynamic range. Although when hi-res means that you start to throw ultrasonic noise into the DSP bag, who knows what that will do to the audible range?

These are all assumptions I'm making that I think, sound logical at least. With a whole fresh generation of active DSP corrected speakers on the market today and no limitations of physical media. Why wouldn't producers delivier 24bit, 48kHz/88.2Khz as a default to file-based- and streaming services?

*bits in audio would be color depth in photography. Sampling frequency would allow for a greater spectrum of sound/lights but as with ultraviolet, there would be no point to capturing ultrasonics.

 

[voodooless]

*bits in audio would be color depth in photography. Sampling frequency would allow for a greater spectrum of sound/lights but as with ultraviolet, there would be no point to capturing ultrasonics.

Your conflating a few things here. Bit depth is just the amount of discrete values you can assign. For audio that equates to signal level, for photography it’s the amount of light hitting the sensor. Sampling frequency equates to the amount of samples per time interval. In photography it’s more or less the resolution of the image. Colors are not even part of this, because there is really no direct comparison for it in the audio world. Color happens just with a filter in front of the photo sensor. And there are just three of them (with rare exceptions). Image based comparison rarely make for good analogies to audio.

 

[Francis Vaughan]

I don't know if I'm correct but I personally feel like hi-res is a better source if you use DSP corrected hardware. I draw a comparison to photography* where a raw image and a JPG

If the audio is uncompressed it is already raw. If the only source is a lower bit rate mpeg, then there is potentially an issue if you wanted to use the audio to create say a new mix or remaster the audio. But if all you are doing is adding a bit of frequency shaping or reverb it almost certainly won’t matter even then.
Hires audio would be like using an ultra high resolution sensor where the lens is not capable of delivering the resolution the sensor can resolve (for sample rate) and a very deep bit depth sensor to photograph a scene with intrinsically lower contrast then the sensor’s range (for bit depth).
In both cases the sensor is delivering redundant information. As Amir has shown, it is actually worse than this. It seems so much high rez audio is like getting a photograph where not only does the sensor’s resolution exceed the optics, but the sensor has lots of glitches and errors but they are only visible when you go pixel peeping.
There is a common attitude that it is always better to have more. But it is really better to get what is needed. High rez just gets you a zoomed in view of empty or erroneous information. It’s presence can only lead to downsides.

 

[frankh]

Amirm, thank you for showing all of us how the Hi Res audio companies are ripping off their customers that are paying so much of their hard earned cash for garbage. You are really doing all of us here a big huge favor by telling the truth. Don't worry about the excuses they come up with to protect their business. Truth is truth and it takes a lot of courage and integrity for one to speak the truth. Keep up the fantastic reports and reviews. You're the best Amirm

 

[sarumbear]

Yes, another older high-resolution music format comparison, this time with a different twist than the previous ones:

Thank you for this eye opener, @amirm

It is blatantly obvious that mastering of these so called hi-resolution tracks are not correct. I don't see this as the label's fault. We have mastering engineers for a reason. I'm 100% sure that all those tracks and versions are processed by one or more mastering engineer. It is them who should stand up, took the blame and correct their mistakes.

I wonder if there are any mastering engineers in this forum? I like to hear their comments on this.

 

[Hippocamp]

With this series of videos, I'm interesting in downsampling some hi-res content to test out for myself whether there are any audible changes (I'm assuming there won't be). I've always used the SoX resampler in Foobar with the default settings (best quality, passband 95%, aliasing/imaging unchecked, phase response 50% linear). But I have Adobe Audition and Audacity available, too. Any differences in how these programs implement resampling, other software options, a "best practice" set of recommendations?

Also, I use Qobuz for streaming. I've always assumed that they do a competent job downsampling their content, but I saw recently that there have been problems with Tidal downsampling...

 

[Hippocamp]

It's fun to pull my hi-res files (mostly 96K and 192K) in to Audition and see what's going on. A lot of them are fine, with a smooth, exponential dropoff and basically everything correlated with music. Some have lots of HF noise...

But here's one that has me puzzled. It's a 192K file of the third movement of Beethoven 5th produced by Reference Recordings. Most of the file looks pretty much OK. Here's a snapshot of the frequency content early in the track.

BUT, there are small sections of the track where the HF noise goes absolutely nuts. They are obvious in the bottom-left spectrograms.

Any idea what might be going on? Could they be splices from multiple recording sessions with different engineering?

 

[sarumbear]

I was searching guidelines for the recording industry for high-resolution audio and I find this paper from the Recording Academy (Grammy) dated September 2018. Emphasis mine.

"Legitimate hi-res audio files should contain content that extends well above 20 kHz. However, files that have been upsampled from CD-quality audio will display a brick wall filter at 20 kHz because of the low-pass filter applied minimize problematic aliasing. Consumers are willing to pay extra for the highest quality audio but they deserve to get what they pay for. When a consumer purchases or streams a file labeled “Hi-Res,” that file must conform to the definition of hi-res, meaning that the file originated as hi-res and that conversions do not obscure or misrepresent the quality of the original source files.

"Consumer confidence depends on reputable practices that deliver all audio files at the stated resolution. This is a serious concern for producers and engineers who might be tempted to upsample files that aren’t legitimately hi-resolution. It is extremely important to be aware that hires distributors and labels—and many technically adept consumers—use software to analyze the files they are sent. If the files are not truly hi-res, it will be discovered quickly. The distributor/label who has delivery requirements that specify hi-res audio will reject the files and could refuse payment to the producer and/or engineer."

There is nothing in the paper that anyone can disagree. However, it is obvious that mastering engineers have either not read it, or simply ignoring it. What I said earlier is echoed by none other than the Grammy: If the file is not prepared correctly ask your money back. Force their hands.

 

[Francis Vaughan]

Any idea what might be going on?
Could they be splices from multiple recording sessions with different engineering?

Certainly possible. Some conductors are known to obsessive in the extreme and demand that a note perfect final artefact be produced, even to the extent of splicing out tiny blemishes. Taking bits from different sessions and patching the bits is pretty much how it goes.

The notes from the Reference Recordings web page suggest this is exactly what has happened:

Soundmirror have made this recording, comprised of three live concerts from December 2014, possible.

So yes, looks as if there was a nasty problem in at least one of the concerts. It is pretty damning that these sorts of artefacts appear and nobody notices. The idea that there is some sort of magic in the silly high frequencies is pretty much refuted by this example.

 

[FrantzM]

THANK YOU AMir!

I had you to scream it.

Unfortunately, those who want to pay will continue to pay. Companies being in the business to make a profit, then .. profit from the beliefs of their market. I don't see it as "ripping-off", as proof, you just need to go to one of those subjectivist websites: The forum participants, will rip your head off if you dare to ask for a sliver of objective proof.

Again thanks!

 

[Pluto]

It is blatantly obvious that mastering of these so called hi-resolution tracks are not correct

Now let's keep this all in perspective. Using Amir's example, he was admiring the aesthetic beauty of the high frequency decay up to the point where the vocal mic was opened and the aforementioned beauty went tits up. Nothing obvious changed about the sound as the vocal mic was opened, although good recording practice dictates that the mic should have been open from the beginning of the number to avoid any audible change in the acoustic perspective that might be down to the additional microphone being added midstream, as it were. Unless, that is, the singer was so well screened acoustically that next to no spill from the band was arriving at the vocal mic (or the vocal recording was an overdub, recorded after the band performance).

We observe (NB observe, not hear) noise getting added at frequencies so far beyond the range of human hearing as to be inconsequential before we even consider that it's at levels that would be barely audible even if it were within the normal human range, which it is not.

So are mastering engineers now being required to produce records that are more designed for forensic examination than listening??? Please note that I am not condoning the up-conversion and sale of a Red Book recording as “hi-rez” (unless this fact is declared). The entire recording chain was not designed to produce clean noise floors at ultrasonic frequencies. If consumers really are dim enough to want to buy into technologies designed to reproduce 200kHz bandwidth (over ten times that required for homo sapiens listening) then they shouldn't be put off by the small amount of noise delivered as part of that package.

 

[phoenixsong]

Awesome work, time to reselect my music test files for comparing audio equipment with! Any idea how I can easily and accurately look for artifacts in my music on a Windows 10 laptop? Preferably with no/low cost

 

[sarumbear]

So are mastering engineers now being required to produce records that are more designed for forensic examination than listening???

Yes. It’s part of their job!

The entire recording chain was not designed to produce clean noise floors at ultrasonic frequencies.

You are wrong. There are many recordings systems that are perfectly suitable for high resolution/frequency recording.

If consumers really are dim enough to want to buy into technologies designed to reproduce 200kHz bandwidth (over ten times that required for homo sapiens listening) then they shouldn't be put off by the small amount of noise delivered as part of that package.

So you do not believe in consumer protection. Fine, it’s your choice. You are most likely an American and believe in the Wild West principal. I’m a European and I expect the consumer to be protected.

Selling the following as high resolution is blatantly lying. you cannot argue with a fact. The seller should be shamed. I personally want them to be fined even!

 

[Francis Vaughan]

So you do not believe in consumer protection. Fine, it’s your choice. You are most likely an American and believe in the Wild West principal. I’m a European and I expect the consumer to be protected.

You might note that Pluto's abode is listed as in the UK, like you, and not exactly far away. Which is a clue. You might re-read his post with your standard British sarcasm and dry humour filters turned up a couple more clicks. I'm just from the colonies, but I think I got his message loud and clear.

 

[sarumbear]

I forgot to address this on my previous post.

We observe (NB observe, not hear) noise getting added at frequencies so far beyond the range of human hearing as to be inconsequential before we even consider that it's at levels that would be barely audible even if it were within the normal human range, which it is not.

As @amirm has demonstrated almost all amplifiers has enough bandwidth to amplify the entire high resolution file's range. Look at this example. It has enough energy to potentially burn your tweeter if you like listening music loud.

 

[sarumbear]

I'm just from the colonies, but I think I got his message loud and clear.

My sarcasm maybe misplaced but my replies are not. Tell me if they are. I especially have a problem with the following attitude and I made that known.

If consumers really are dim enough to want to buy into technologies designed to reproduce 200kHz bandwidth (over ten times that required for homo sapiens listening) then they shouldn't be put off by the small amount of noise delivered as part of that package.

 

[GWolfman]

Do you have any DSD files that show proper use/conversion? It's good to now know how to recognize poor DSD releases.