The Continuing Myth of Hi-Res: Streaming Upgrades
The news this week included the announcement by Spotify is upgrading the fidelity of their music streams to “CD-quality.” I watched a video by John Darko, read the press release about Spotify Hi-Fi and another article on CEPro by Robert Archer, and was so thoroughly disgusted at the inaccuracies included in a piece titled “High-resolution audio: everything you need to know” written by Verity Burns and Becky Roberts on WhatHiFi that I had to post a lengthy comment pointing out a few of the issues. Why is it so hard for “professional” writers to get a grip on real facts when they write about the latest marketing initiative pushed the hardware and software companies (actually it’s been around for almost 20 years but only recently seems to be getting some traction). Don’t they realize that the information they are disseminating is almost word for word the spin put forward by the companies? I know everyone wants to have something to say but when authors simply repackage a company’s press release, I have to wonder about the future of audio journalism. Do your research, talk to real experts, and challenge the hyperbolic claims made in those releases.
While I generally dismiss anything the John Darko says, at least he acknowledged that virtually all of the so-called “hi-res” content on Qobuz, Spotify, and the other services isn’t really high-resolution. He claimed 90% of the 70 million tracks available are limited to CD-quality. And he’s right. Robert Archer, the author of the piece at CEPro, is confused about the term “CD quality.” I’ve talked about this before. He wrote, “Spotify’s newly announced Spotify Hi-Fi is an upgrade that will be available to users shortly, and the popular streaming audio company says the service will provide listeners with a CD quality (16-bit/44kHz, 1,411kps bitrate) level listening experience. Here’s what he — and many other writers — fail to understand. There is the Red Book or CD-Audio specification, which provides all of the physical, logical, and electrical details about the compact disc. The CD-Audio specification delivers 2-channels of 44.1 kHz (not 44 kHz)/16-bit linear PCM encoded audio at a bitrate of 1411.2 kbps (44100*16*2). However, “CD-quality” audio is defined as audio that is perceived as equal in fidelity to the Red Book specification. It’s not required that it actually be CD-Audio spec. If a group of listeners hears a 320 kbps org vorbis audio steam and is unable to reliably identify it as a lossy encode then it is deemed “CD-quality.” The download and streaming companies have been using this descriptive — and deceptive — term for years. It is not what the Red Book describes.
The WhatHiFi article makes an attempt to define high-resolution audio. They reference the June 2014 press release issued by the DEG, CEA, Recording Academy, and the labels, which establish hi-res audio as anything “better” than CD-quality (there’s the same erroneous definition again!). The authors continue with a simple explanation of sampling frequency and word length. They wrote, “The more bits there are, the more accurately the signal can be measured in the first instance, so going 16bit to 24bit can deliver a noticeable leap in quality.” I would challenge the authors to take the HD-Audio Challenge II, the survey I created in 2019 to test whether hi-res is perceptible vs. CD-Audio spec audio. As most of you already know, the results showed that even experienced audiophiles and professional audio engineers using state-of-the art, expensive equipment did no better than a random coin toss at picking the native hi-res audio files. It seems to me that writing there is a “noticeable leap in quality” is at best wishful thinking and at worse outright lying. Finally, they listed a number of the “main file formats” associated with Hi-Res Audio including MQA. They wrote that MQA is, “a lossless compression format that efficiently packages hi-res files with more emphasis on the time domain.” Missed it again. It is not lossless unless you think slicing off 7 bits from the actual data words to make room for their “origami” magic qualifies as “lossless.”
I know I’m always harping on the BS that seem to dominate audio articles, FB posts, and Youtube videos, but I have a very hard time accepting that this is the state of information in our hobby.
Here are the facts you should know:
- All download and streaming services get the same masters from the labels and make them available to subscribers at varying quality levels according to their own business model. Once they all arrive at CD-Audio spec, we’re getting everything “the artist intended.”
- Virtually all of the masters offered by these companies are not bona fide hi-res audio productions. They are “hi-res transfers” of older non hi-res masters. Does anyone think a 192 kHz/24-bit transfer of a recording made in 1932 is hi-res? What about an analog tape track from the 60s?
- CD-quality is not the same thing as Red Book standard – CD-Audio specification audio. Anything less than 1411.2 kbps is NOT CD-Audio and will require a codec — lossy or lossless — to deliver it. Most CD-quality audio is encoded at 320 kbps.
- Qobuz, TIDAL, Deezer, Amazon Music HD, Apple Music, and now Spotify Hi-Fi are delivering standard-resolution audio NOT HD! And it’s OK because no one can tell the difference anyway — remember the HD-Audio Challenge II. Just because they label it HD doesn’t make it so. Amazon Music HD shifted all CD-spec audio to their HD category. Why? Because they want us to think it’s better.
- The fidelity of any audio reproduction is established at the time of the original recording and the master is delivered to the label NOT by the platform that ultimately delivers it to you. They can only make it worse. Sure, TIDAL, Apple and the others want us to believe we’re at the dawn of a new listening experience a high-resolution experience, but it’s all marketing nonsense. A well made recording at Red Book specification can sound astounding and is sufficient to meet even the most discerning audiophile. Just think of all of the bandwidth and storage we’re saving.
- Avoid MQA, DXD, DSD and stick with good old PCM. It’s been the standard in the industry since the 70s and will continue to outclass all of the newcomers.
The move to CD-Audio specification streaming is to be applauded. We aren’t stuck with Lo-Fi MP3, AAC, or Org Vorbis files. But moving beyond Red Book audio is NOT necessary. The artists, engineers, and producers know that incredible recordings can be delivered at CD-Audio spec. It’s the marketing departments and the uninformed that continue to push for formats and standards that don’t move the needle.
I’m a Private Pilot!
I started flying gliders in the summer of 2018 with the goal of becoming a private pilot with a glider certification. I had to take 2019 for medical reasons but started soaring again last summer. I’m thrilled to tell you that on Friday, February 26, 2021 I passed my oral exam and check ride and became a private pilot. I’ve worked really hard on this over the past few years. There’s a lot of detailed information to learn and flying skills to be mastered. My father was a bomber pilot during WWII and became a corporate and ultimately a check pilot for GM before he succumbed to cancer at the age of 43. And my aunt received the Congressional Gold Medal in 2010, 60 years after her service as a WASP during the war. My half brother’s widow informed me after my posting on FB that he and his sister were also trained as pilots. I’m so happy that I’ve reached this first milestone on the way to more time learning to fly.
25 thoughts on “The Continuing Myth of Hi-Res: Streaming Upgrades”
Wow great job on the pilot’s license!!!! Was your Dad a WWI or WWII pilot?
Enjoy the flying! It is really awesome to do.
Why in the heck does the mainstream tech media not understand your argument about hi-res!!!!
Your description is so clear and precise!!! I think your efforts will eventually result in common understanding about the technology but holy cow the marketing push is so idiotic… May the truth eventually be known! LOL!
My Dad was born in 1924 and died in 1968. He was a very young Captain in the Air Force and flew B-17 bombers.
As a long time Spotify user I am pleased to notice that Spotify is not repeating the Hi-Res BS, that the other services continue to do.
As I have understood it, we will be able to play the files at 16/44.1 quality – details about container, bitrate etc are not known yet.
Thus I can only applaud it.
Furthermore they do not talk about prizing – just about being able to ‘upgrade the Premium service’. What that exactly means, time will tell.
What I am much more concerned about is, whether it will be possible to play the files gapless (important for concept-/concert-/classic music-/albums as you know) or not.
I will definitely try it out…..even if I doubt that I will be able to tell much difference to the now 320-OggVorbis product.
In case of ‘no gapless playback’ I might stick to the Premium quality – just for that reason.
Note: Not all online journalist are telling us just what the industry wants us to believe.
Here is an example:
Keep up the good work 🙂
Mark, congratulations on getting your private pilot license!
Thank you for everything that you do for the industry! I’m not the only person you’ve saved tons of money by not “investing” in the Hi-Res / MQA marketing hype. 🙂
Congrats Mark. You must be thrilled. Agree on the PCM. I only grab DSD now if that is the only way to get it and convert it to PCM for playback. Works perfectly.
Mark, I am a longtime admirer of your work, but for once I think you were being a little hard on the What Hi-Fi authors when you castigate them for equating “CD-quality” and Red Book standard audio. CD quality to most of us is uncompressed PCM audio with a 1411.2kbps bitrate, as you have stated, and I’m sure that’s what they meant.
The unanswered question here is whether Spotify Hifi will actually be streaming at 1411.2 kbps. (What they have stated publicly is only “CD-quality, lossless audio format,” which implies the full bitrate, but who knows?) And if they do, will the results actually be audibly distinguishable from the present 320 kbps Ogg Vorbis stream?
The confusion over the terminology of “CD-Audio spec” and “CD Quality” is widespread. Heck, even the Recording Academy got it wrong in their press release. The authors at WhatHiFi should do more research. But that’s not what irked me about their article. To blatantly accept that more bits and a higher sampler rate equates to better sound is parroting the measuring of the marketers. I know it’s common practice but it still needs to be pointed out.
‘CD Quality’ means absolutely nothing if you stop to think about it. ‘CD specification’ when you understand it has meaning but still is no guarantee of the ‘quality’ of the sound. The bottom line is being told the original quality of the finished mixed master and then basing your choices on this.
Thanks for yet another “hitting the nail on the head” post Mark!
Keep them coming as many of us are out there trying to get the fibbers to tell the truth with little success.
The uneducated Joe Sixpack has no way of knowing how badly they are being lied to without posts like yours.. Sadly there’s little to no oversight of industries snake-oil peddlers, if this was medicen many of them would be in jail.
Congrats on getting your licence, as we bikers say, keep the rubber side down and the shiny side up. 😉
Is there any software able to distinguish downloaded “HiRes” file from Red Book standard ?
I am afraid they could be “tweaked”, however.
But is it possible ?
Fly high 😉
You can use analytical tools like Sound Mirror or Audition. Looking at a spectral display can be very illuminating.
Hi. I have your excellent book and know you are working on a new edition. I was wondering what device you would recommend for playing surround-sound recordings over an Ethernet network using Roon. If it could also handle DSD and serve as a 4K video receiver, that would be a bonus. Looks like Oppo has stopped selling its products. Some audiophile products seem ridiculously expensive.
If someone already has surround recordings in DSD, I am guessing that the easiest solution is to convert them to FLAC, though I suppose it would somewhat degrade the sound.
The more I looked into the technology of the Digital Compact Disc produced to RBCD standards the more in awe of its engineering I became. RBCD is much more than 2 x 16 x 44.1 Khz. That’s only a small part of what is actually on the disc. There’s a lot more information and there are error checking systems and error correction systems that can only be described as belt and suspenders. Each track is scanned a minimum of six times. If it’s wrong it continues to scan it until it gets it right. This is why a scratched CD might not play certain tracks. The pits are embedded below the surface coating. The scratch changes the optical reflectivity of the disc at that point. It is often possible to rub the scratch out with a mild abrasive that will restore the player’s ability to correctly track the recording. Try that with a scratch on a phonograph record. So just because a stream has 2 x 16 x 44.1 khz doesn’t mean it is CD quality as defined by RBCD standards.
As for DSD, one company PS Audio builds D/A converters that convert all digital formats to DSD before it becomes an analog signal. Its designer Ted Smith uses FPGA technology (Field Programmable Gate Arrays) where the programming is periodically upgraded and download without cost to PS Audio Customers. Ted Smith claims to be able to hear “digital jitter” of two parts in ten billion. His test instrument by his own claim is a toe tap meter. (I’m not making this up.)
After watching several tube videos and reading audiophile blogs by Paul McGowan of PS Audio, my conclusion was that these people freely lie and deceive as they feel in order to make lots of money off gullible (mostly solvent I guess) Hi-Fi hobbyists aka audiophiles. Lots of bullshit going on there. There are many other ways how one can make good money by delivering honest, real, tangible values in goods or services they provide. Not just a perception of solutions to falsely created problems that in reality do not exist at all.
There are very few sources of accurate, unbiased information regarding audio equipment, formats, and content.
I doubt that I could hear any difference between DSD and PCM, though a Japanese study from around 2014 supposedly found a preference for the former. It’s just that I have a bunch of SACDs already in that format, some in surround.
For all the manufacturers, including PS Audio, that criticize listening comparisons, has anyone ever challenged them to propose a double-blind study to test whether listeners can discern a difference? They are great at criticizing this or that study, so why don’t they specify the criteria that they would find acceptable for a study? They raise all these claims that are not falsifiable. So why should anyone believe them? I would like to see them take their own challenge. If they can’t pass it, then they have a lot of ‘splainin’ to do.
There are a number of studies comparing formats and specifications within categories. The DSD debate is moot for me. The methodology and limitations of producing audio in 1-bit DSD makes it a non-starter for me. I dedicated an entire chapter to the format and included a detailed explanation why it is highly flawed. But the magazines and other companies jumped in with nonsensical reviews akin to the hoax of MQA.
As for companies or publications backing rigorous testing, they will never do it because it would diminish profits and limit growth.
You included Apple among those who “want us to believe we’re at the dawn of a new listening experience”. I don’t recall hearing anything from Apple about HD. Did I miss it?
Apple has been requiring high-resolution sources for their platforms for many years. You’re right that they haven’t made a big deal over HD like Amazon and others. But they have touted new formats and surround delivery.
More fuel to the MQA fire.
See the new debunk video of MQA by goldenSound.
Keep the good work for real HD Music.
So why does my TV’s high end sound better than the high end I get from CDs through my lossless player? My lossless player isn’t the problem as far as I can tell, because CDs seem to limit its sound quality. I play both through the same receiver and speakers, so they’re not causing the difference. Might the dither used on CDs be a factor, or is dither completely transparent?
To me, cymbals are the acid test for digital recordings and playback devices. Although cymbals sound more realistic on TV in general than on CDs in general, there are quite a few CDs with decent-sounding cymbals. Perhaps I could improve them up by replacing my lossless player with a Modi 3+.
My previous concern that dither might explain the slight incoherence of some cymbals was erroneous. It might actually be a result of substandard digital mixing.
Niacin deficiency is common, and can cause loss of high-end acuity, and minor problems such as cancer. I recommend visiting DoctorYourself.com and searching for niacin. I take a gram (1/2 tsp) of niacin powder at a time (which requires starting out with small doses and gradually increasing them, to avoid the intense itchiness of an intense niacin “flush,” which can last for about an hour), by mixing it with half as much baking soda in about 1/4″ of water to neutralize its strong acidity and to improve its solubility. I assume that mixing it with baking soda is OK, but it was my idea based it on mixing ascorbic acid with baking soda, which is approved by experts.
Is there a site or article that explains what to look for? I am an engineer and understand what a spectrogram (e.g. from Audacity) is showing me, and I get simple things like Nyquist rate sampling and low-pass filtering. But I don’t know the signatures of other audio processing artifacts like bit-rate compression, upsampling, dithering, etc. Is there a good tutorial on how to spot these things in a spectrogram?
I remember reading years ago that digital recordings made with DSD or PCM samples the source at 2.8 million bits per second in the case of DSD and some similar equivalent in PCM. Where 1/4” analog magnetic tape running at 15 ips yields a sampling rate of 80 million particles per second. I admit some cd’s sound pretty darn good the days, but I think a good vinyl pressing played back on a quality turntable still sounds more faithful to the original. Tape to vinyl is a physical representation, it’s not chopped up & reassembled. This would work if the sampling rate were high enough. If you look at an old photograph, it’s a continuous tone print. The newer digital photos can come pretty close and look better as the sampling rate goes up. Does this make any sense?
Thank you, Michael Ruggieri
The fidelity specifications of the best analog tape machines maxes out at around 72 dB (with noise reduction up to 85 dB), which is far short of the fidelity delivered by PCM audio at 44.1 kHz/16-bits. The notion that digital is “chopped up and reassembled” misses the point. The reconstruction of the analog waveform without the noise and distortions inherent in ALL analog methods makes it more faithful to the original.