Hi-Res audio files: the real thing or marketing crap?

In this link from Sony there is an example to “feel” the difference between a compressed file and a high-res file

https://helpguide.sony.net/high-res/sample1/v1/en/index.html

It would be nice to perform the same comparison with a 44.1kHz/16bit FLAC and the 96kHz/24bit version of the same file

This this this. And when artists care about mixing and mastering you get outstanding quality especially at Hi Res and 44/16 both of which are easily discernible from their compressed versions.

Here is a good example, listen to Stevie Windwood 44/16 Back in the HIgh Lfe and it sounds dull and flat. Now go to any Steely Dan Aja song or Gaucho and notice the tremendous sonic difference. That is mixing and mastering.

And anyone that does Apple Music should have noticed that everything just sounds better now that they went fulllossless.

All Steely Dan albums sound wonderfully, I love “Two Against Nature” too, wonderful arrangements, mix and master. It’s impressive on a good HiFi system. Acoustic perfection.

About Apple Music: I have some doubts about the discernable differences between different album versions or between the same albums played on different online platforms.
I don’t like to make assumptions without evidence, but I don’t feel I can completely rule out that they could slightly modify the sound in some way with DSP, or that the proprietary acoustic algorithm could give different “colors” to the sound.
I know, it seems I’m talking about how many angels can dance on the head of a pin, but in my experience, when I tried to make comparison between a 16/44,100 file and, for example, a 24/192 one - both not streamed but played locally on my HDD - I couldn’t spot any tangible differences.

The ability to hear accurately does of course differ between people, but what I have found in A-B testing is that most people can hear the difference between CD-quality versus records on a high end stereo system. More recently, I decided to develop a digital collection recorded from my records. I started out with 384 kHz/16-bit files, and I thought that was the best I could hear. I was wrong. I purchased an RME ADI-2 Pro converter capable of 768 kHz/32-bit. I was expecting and hoping to not hear any difference between 384 kHz and 768 kHz. I was amazed that I immediately heard more details: space between instruments, clearer sounds in general, more specific instrument locations (better sound stage). That was with my Koss electrostatic headphones. I did not hear the difference with my Sennheiser HD600 headphones. I have since upgraded cables and power supplies, and the sound has improved more. I find that a stereo system is like a chain: the chain is no stronger than its weakest link, and a stereo system is limited by it’s weakest (noisiest, slowest, etc.) component. So my answer to HD sound is, you can only hear the difference if your total system is capable of reproducing the detail available in the recording, and your ears are capable of discerning it . . . like a chain. In my case, I am lucky (or unlucky) enough to be able to hear the details. In that sense, my hearing has cost me a lot of money over the years.
As for “science”, people generally ignore the fact that humans are capable of perceiving high frequency harmonics way beyond what they can hear as a clear tone. I can’t seem to hear a 10 kHz tone, but I can easily identify details (high frequency harmonics) exposed at 300 kHz . . . but only with electrostatic drivers that accurately reproduce those frequencies. So bottom line: I think a person has to match the digital quality with the chain: their complete stereo system plus their hearing.
As for my source, I have about 800 records, some of which are really good recordings using pure analog mastering. Most of my records from the 60’s and 70’s are analog, and they benefit from super high-res digitization. In the last few weeks, I added a Synology NAS with 30TB of space, because I’ve been recording more than I had planned to. As I said, good hearing can be expensive, especially if you enjoy the natural sound of instruments as much as I do: records such as Solti/London playing The Planets from Original Master Recordings. I use CD-quality music strictly for background music. I’m not an elitist. I just love music.

Just to be clear on what you mean: are you talking about digital rips of vinyl records?
Well, I’m not an expert at all on this practice, I’ve never tried to digitalize my records.
It’s possible that by doing the conversion at a higher resolution (like the impressive 768 kHz/32-bit you mentioned), you’ll hear more details in soundstage, maybe beacuse the conversion process benefits of that higher resolution.
However, I think a vinyl can offer the best of itself when… played as a vinyl :smiley: It’s self-evident.
It’s impossible that the sound recorded between the grooves of a vinyl can be improved when digitalized. You cannot improve the native (limited) resolution of a vinyl.
If a digitalized vinyl sounds better than the vinyl itself, it means that your digital audio chain in your system is better than the analogue one.

In any case, what matters to me with this thread, is to understand with all of you if there are discernible differences between native (not converted from an analogue support) digital masters of the same album when you play it at a: A) “standard” cd like resolution (44/16) vs B) hi-res (e.g 96/24).

When you talk about the audio chain, where everything should be up to par, including your ears, I agree.
Although, I must say, I am rather skeptical of the improvements that components such as power supplies and cables can offer. Floyd Toole, one of the world’s leading hi-fi experts, is also skeptical (to say the least) about it.

1 Like

Yes, digital rips of vinyl records. The reason is that I no longer have high end speakers, and digital is more practical for me now. No, of course the sound cannot be improved by digitizing. I just want to come as close as my hearing can discern to the ideal sound from the records . . . which I think I have achieved.
As to other sources, DSD128 and DSD256 are available on the Internet. If your stereo system includes a DAC such as the ADI-2 that can reproduce high frequency harmonics, and everything in the chain is consistent, those are much better than 44 or 96 kHz. Incidentally, I also have a Chord Qutest DAC. The Chord DACs are intentionally and explicitly not capable of reproducing the high frequency harmonics. Disappointing, but true.
As to power supplies and cables, there is a lot of debate about it, but my experience has been that they make a significant difference. For example, RME claims that the ADI-2 Pro can use pretty much any power supply with no audible difference. Just for the fun of it, I tried a linear power supply on it, and then I had a sound stage with depth instead of flat. I thought a wall wart power supply should be fine for a computer that sends the DSD to the DAC, but a linear PSU helps. Have you noticed light flickering sometimes? When the power fluctuates either in voltage or current, it affects the supported component. For a mid-level stereo chain, that is not noticeable, especially when playing at 44 kHz. But if you want sound reproduction comparable to a front-row-center experience, CD’s just don’t cut it.
People say that USB cables don’t matter because “it’s just 0’s and 1’s”. I was using a cheap USB cable, and wondered why the sound wasn’t very good. I switched to a slightly better cable, and the sound improved a lot. The reason is simple: when you are using high frequency digital like DSD256, for example, the timing of the 0’s and 1’s is critical. I eventually settled on an Audioquest Diamond USB cable. But you really need electrostatic drivers (speakers or headphones) to get the best detail. Other drivers are slower, so they can sound really good, but they can’t give you the higher frequency harmonics. The WBC cables are quite good, and much less expensive. I recommend them. But if your hearing and budget demand it, the top Audioquest cables are probably unbeatable. I had been using the original cable between my turntable and the preamp. 30 years ago, I thought the lack of base in my high end stereo was due to the speakers not reproducing frequencies below 40 Hz. A month ago, I switched to an Audioquest Leopard cable, and wow! Now I get base from my records like never before. I don’t know why the “experts” think cables don’t make any audible difference. Most of those same “experts” claim a human can’t hear any difference between 96 kHz and higher sample rates. I have done A-B comparisons, with half a dozen people who didn’t have good stereo systems, and they all agreed with me, so I think the “experts” are not basing their opinions on human experience.

1 Like

New productions will be produced and mixed in high resolution, mainly due to the advantages regarding noise accumulation which will be less dominant with a higher headroom. But I am convinced that you are not able to hear any difference when listening to either a CD-quality or high resolution version of that same production.

But I also believe, that when record labels sometimes release ‘new’ high resolution productions from older releases aimed at audiophiles, they sometimes use a different source. Maybe a source without the annoying mastering, that alters the sound dramatically. Or – at least less mastering. That can be very audible.

And I have noticed difference in quality from different streaming sources. I have CD quality FLAC files of my own CD-collection, and a few high resolution FLAC files also, and use Tidal for streaming. In my opinion Tidal has the best sound for streaming. Some other sources seem to add ‘something’, maybe like a bit of mastering, to make the sound more impressive in a poor system or headphones, which the majority of end users tend to use, but make it sound bloated on a better sound system.

Mastering (and re-mastering) typically involves returning to the original multi-track recording and revisiting the mix levels, fades, effects processing, and other dynamic adjustments made during the mixdown. Since many older recordings were mastered with the target of vinyl at the time, remastering of those recordings can result in dramatically improved tracks. I agree that just taking an original master and upsampling to higher resolution audio formats will not yield improvements. The processes used by TIDAL are clearly beneficial as far as my system is concerned, in many cases the streams sound better than my CD quality FLAC collection.

Many people report that listening through Tidal result in a discernable better quality.
It would be interesting to understand why this happens.
If I make a perfect and lossless 16/44 rip of a CD I own and play it through a quality system like Volumio, why should the same version (not hi-res, but still 16/44) be supposed to sound better on a streaming platform?

This goes back to what I wrote earlier about my assumption that the audio could be somehow artificially processed by these platforms (DSP? The debate exists).

I feel like there is a big difference as long as the artist that is doing the recording at high quality too because you can notice it when using headphone when something is just loud but very little detail separation when its clean with clear instrument separation that doesn’t hurt the ears.

1 Like

I think for most people the limiting factor will be the quality of their playback system not the limits of the 16/44 format. In my experience over the years upgrading cd players, streamers, amps and even cables to better ones results in improved sound from the same cd or file. This suggests to me that the limits of 16/44 as a format have not been reached by the hifi system and that the limitations of playback equipment should not be mistaken for the limitations of the format being played. To my mind mastering of the music has a far greater impact on the sound quality than the limits of 16/44 ever will; there are great sounding CDs and not so great ones. In terms of playback equipment, for the vast majority of us it seems you can always improve what you can pull from the original file, a whole high end hifi industry has evolved to do just that.

I think for most people the limiting factor will be the quality of their playback system not the limits of the 16/44 format. In my experience over the years upgrading cd players, streamers, amps and even cables to better ones results in improved sound from the same cd or file. This suggests to me that the limits of 16/44 as a format have not been reached by the hifi system and that the limitations of playback equipment should not be mistaken for the limitations of the format being played.

True. But there is another limit to keep in mind: what our ears are able to hear. You may also have the best performing system in the world, but, talking about frequencies, there’ll always be the natural limit of our hearing system (20 - 20,000 hz at best, when you are a child).
Anyway, I agree with your general point of view: a better playback system improves everything you hear on it, even lossy files.

To my mind mastering of the music has a far greater impact on the sound quality than the limits of 16/44 ever will; there are great sounding CDs and not so great ones.

This is the real thing! Loudness war and the logic for which music should be listened to “loud” even on smartphones, ruines a flood of great albums.
I listen mainly to music from the last 30 years and this is a problem that I suffer a lot.
9 out of 10 modern pop and rock productions have masters compressed as hell, sometimes clipped. it’s a shame.
Sometimes it happens that for some strange reason a more dynamic master of an album originally released with a high compressed master, come out. The difference is incredible, no matter the format, it can also be an mp3. It is the dinamyc range that makes all the difference in the world.

2 Likes

99% of “hi-res” music is upsampled … i.e. it was originally recorded at a lower bit rate/sampling rate … so no possibility to add additional information just bogus samples based on arithmetic.

Dr Mark Waldrep of AIX records (a label dedicated to hi-res original recordings) conducted an extensive survey and himself concluded that it is not possible to tell in blind listening between hi-res and red-book (16 bit/44.1khz) see HD-Audio Challenge II: Preliminary Results – Real HD-Audio

This guy dedicated his whole career to hi-res recording exclusively in 24/96khz but has accepted that there is no audible benefit … as others have noted the recording quality has little do with hi-res. It is simply marketing hype.

320kpbs ogg (spotify) is all you need - anything more is just wasted bandwidth

Yes, Mark Waldrep knows his job.
But did he compare his masterings or badly undersampled files ?
Looks like 16/44 mastering is very often deliberately worse than the same recording published in Hires. Some good MP3s could be better done !
IMHO it is a marketing trick to make money with selling Hires files.
But beware of the contents, at least have a look with “Spek” or anythig similar : you could be surprised :frowning:
Regards,

1 Like

What you mention is, in my opinion, a phenomenon quite popular with people who are able and willing to spend a lot of resources into building their “perfect” music-playback system. You spend months and years building your system, checking for any weaknesses, trying this, trying that and end up with a decent high-end reproduction chain. Since all of us are more or less very interested in the technical aspects of music reproduction, this lures us into believing that sampling with 768 kHz has to be better than sampling with 384 kHz. And we hear what we believe - it’s a self fulfilling prophecy. The only remedy would be a blind test.
You state, that you are barely able to hear a 10 kHz tone. Yes, the same is true for me (age 60). But no way, that you are able to hear anything at 300 kHz (doesn’t matter if its harmonics or not). What your ear doesn’t hear, doesn’t have any influence.
What definitely influences the quality we perceive when listening to any recording, be it analog or digital, is the recording and mixing process. In my opinion, quality and excellence are created here. And there is a huge amount of music out there with hilarious quality due to low recording and mixing quality. But i seriously doubt that the frequency used in the digital recording, if it is at least CD-quality, does have any influence in the result.

Yes, music is very important to me. In the 80’s, I spent a significant portion of my income on my stereo, and I am now. I understand if it is not that important to you. As to a self-fulfilling prophecy, no. I expected and wanted 384 kHz to be beyond my ability to hear the difference. I made 4 recordings of the same high quality record at 384 kHz/32-bit, DSD128, DSD256 and 768 kHz/32-bit. I expected all of them to sound the same to me, and was shocked when the higher res recordings revealed more detail . . . but only with electrostatic headphones. I thought I made that clear earlier. I gather that you are unaware of how high frequency harmonics play a part in what humans hear. It has very little to do with clear single tones. But I doubt that anything I write will affect your beliefs. For others who may be more open minded, yes, to hear the difference requires a considerable investment. If you have the means and the inclination, I can describe what I’ve done to create a front-row-center experience with digital, if you’re interested. And by the way, I am a scientist with a PhD in Computer Science, if anybody cares.

1 Like

Your point is clear, @jack. But you are talking about something else. As I wrote before, probably in a vinyl rip a high resolution brings audible benefits because it reconstructs the signal in a better way (I don’t know how, I’ve never tried to rip vinyl).
But if we are talking about a native digital signal (same album, same master) played at different resolutions, these differences are not noticeable. Max Waldrep, Floyd E. Toole, Ethan Winer, archimago’s blog: all of them experts agreeing on this.
Anyway guys, I wrote an article on the subject, when it will be published, if you like, I will share it (even if in Italian, you could use Google Translate).

4 Likes

I look forward to your article. Please share the link when it is published.

1 Like

You can’t put back what was never recorded, so of course, if the master is at 48 kHz, that’s all you get. But some albums are still mastered at ultra high res, including DSD256+ and PCM 768+ kHz. If the master is as good as my rips to DSD256, I have found that my system shows the difference between different down-samples. I tried converting my DSD256 (and PCM 768) to lower resolutions (including MP3 and max res ogg) to save space, and the difference was clear, so I bought a NAS to hold my music. (I was expecting the max MP3 and/or max ogg formats to be as good as my ears can hear, and was disappointed at what I heard.)
DSD256 and DSD512 files are available from NativeDSD.com. I have some music from them, and it’s definitely way better than CD quality. So if you have a system like mine, and your ears can hear well enough, some music is available on the Internet.
I’m not saying that every file you get from NativeDSD is going to be great. The quality of the recording is affected by many factors, such as the quality of the instruments, acoustics of the hall/room where it was recorded, quality of the musician, the mastering process, etc. Music from electronic instruments might as well be recorded at 48 kHz. I mainly listen to classical played on the best instruments by the best musicians in the best halls, analog mastered by Germans or British. I do have some really good acoustic pop/rock records recorded before 1980, before they decided that 48 kHz digital is good enough.
On an average or slightly better stereo system, I have no doubt that your statements are true. And for most people, CD quality is all they care about. But if you go to classical concerts and have tears for the beauty of what you hear, and you have the money you are willing to spend to get that experience at home, what I’m saying is that it is feasible, and the resolution makes a big difference in those rare cases.

With my 70yo ears and my ESL63 panels, I still can perceive that Hires 24/96 or 24/88 files sound better than 16/44.
My hypothesis is that, at least on recent classical music, the downsampling to redbook is badly done, be it on bandpass AND dynamics.
Is it deliberate or not, that is the question…
But, as you write, more important are the sound take and the mixing/mastering.
Regards,