Imagining a Standard Metric for Music
It might come as surprise but as I pointed out the other day, there is no standard for fidelity or specifications for music like there is for film mixes. Not that the film standard is perfect but at least they’ve made an effort. Of course, it makes sense in the world of film because everyone goes to theaters to see movies and the filmmakers want a uniform experience everywhere. You can thank a number of organizations and companies for injecting some sense into a world of sound that was hit or miss for decades. Dolby, THX, SMPTE, and the MPAA have all contributed in one-way or another. But music is still operating without any standards or certifications.
An individual mixing engineer and producer have no constraints on the final output of their work. The fidelity is theirs to adopt or abandon. The guys at the record companies are looking for uniformity…they want hits…but as long as the record meets the expectations of the band and the A&R people, it’s good to go. That’s what we get.
And then we play it back in the widest possible variety of playback systems. There’s high-end speakers down to ear buds.
If a recording engineer wants to push their mixes 18 dB above reference level, then they can…without consequences. If the resultant sound sucks because of the distortion or the sheer pain that it causes during playback, the people behind it (the label, producer, engineers and artists) can just say, “That’s the way that we like it”. There no recourse. Unfortunately, this situation is the norm rather than the exception.
It’s not unlike the struggle to define high-resolution. If the major labels accept a standard the actually means you and I can count on reasonable fidelity when a project is advertised as HRA, then most of their catalog wouldn’t be available on the new crop of high-resolution digital music sites. Anything and everything goes in the creation or a piece of music. High or Low Resolution doesn’t really matter.
I imagine a certification process that would define the minimum quality of the equipment used to mix and master music projects. Guys like Bob Hodas (room tuning expert to professionals and consumers) would have to come an “pink” each and every studio to ensure that the frequency response was flat from perhaps 20-30 Hz to 40 kHz (for high-resolution projects) or 20 kHz for standard resolution releases. There would be a reference level just like the film guys have 85 dB SPL.
There should be real and measurable dynamic range levels. If a record is hammered and lacks any dynamic range, then it should be identified as such. This is the type of information that will be included in the HRA Database that I’m working on (coming soon, I hope). But why not establish the guidelines for the people doing the work. And then push real hard to get consumers to audition recording with more dynamics and vote with their dollars?
This is the perfect thing for the Producers and Engineers Wing of NARAS (the Grammy People) to get behind. They’ve issued guidelines and papers previously. It makes perfect sense…but they won’t do it. Unfortunately, the Grammy folks are celebrity driven and I’m not talking about the artists. Their membership includes celebrity audio engineers and mixers…talented people that have risen to the top of their craft by delivering the current state of affairs. These very same engineers produce the recordings that you hear on the radio and download from iTunes. They’re content with the status quo.
I’m not.
Of course it could be argued that the complete failure of the old and flawed German DIN 45-500 standard to bring standardization to what actually constitutes a hi-fi system rather suggests that history is not on the side of there ever being a widely recognized internationally recognized method of sorting the wheat from the chaff in hi-res audio recording.
Mark, I am observing your highly commendable efforts to drive this debate forward very closely. Although I think there is probably little or no chance of ever achieving a recognized standard, there is also a “PR War” to be won and I think that creating awareness that a CD rip and/or an analog reissue isn’t true hi-res is kind of the way to go here.
But, to be devils advocate for a second, who’s to say that good old CD isn’t good enough to be called “hi-res” anyway? Compared with any compressed file it can be a total revelation to ears more suited to low bitrate stuff.
My own view is that 16/44 when done properly sounds a whole lot better than any of us gave it credit for when CD technology was dazzling to behold, until one actually played a disc… DAC technology is so impressive now and, aside from the immeasurably superior packaging, I have absolutely no regrets about ditching my vinyl collection this year – somewhat against current fashion.
To complicate things even further, I’m sure many folks here have heard the curious phenomenon of a 16/44 track, normally a 60’s/70’s analog original, sounding better than the same recording upsampled to 24/96 or whatever. So provenance is important and I really have no clue how that can be policed.
In the end, and as ever, it’s a classic case of caveat emptor. At least we know that most recent recordings can demonstrate true hi-res credentials, so there is going to be a growing pool of really great sounding original material. For the historic stuff, well as I say it’s probably more a question of good PR than lobbying for impossible to enforce regulations.
I wonder if Neil Young even realizes how he has contributed negatively to this whole debate by setting the Pono bar so low…
There is the relative assessment that you’ve rightly pointed out. CD are better than they were..and better than lossy compress formats…and then there’s high-resolution recordings which take things every higher. I prefer for define levels of expectation. And yes, CDs can sound terrific.
Interestingly, will Sony/Philips ever guess to swap integer for floating point in their Red Book in order to make notorious dithering rather unnecessary {at least, in case of CD} ?..
Mono instead of stereo would increase CD’s bit depth accordingly, but who cared this triviality ??
Actually, CD format was intended to deliver from magnetic tape, but, ironically, even with the digital format at hand, sound recordists would continue utilizing tape as medium between microphone feed and 44.1/16 recording, thinking so they do better for future digital sound, but in reality they did and maybe still do much much worse for it.
Wavelet Upsampling is the future for 44.1/16’s, but a music material upsampled conventionally to 384 kHz sounds clearly better than the original.
I would have to explore your claims for upsampling…it has not been my experience that the fidelity actually improves. The sound may change but the essentials specs remain the same…except you have a gigantic file.
Fidelity of a binary sound recording is defined by a degree of quantization error, hence a still higher sampling rate should always be beneficial even so.
Generally, the essence of the process behind Upsampling was described in a popular science book dedicated to digital sound.
On Stereophile site there was publicized a claim saying that Upsampled 44/16 sounds very like 192/24.
Also, one guy after some listening concluded that Upsampling is superior to oversampling and that it lets a CD track sound just like a high-definition audio track.
dCS may be familiar with Upsampling.
Is the DIN so very flawed? It’s considered superior to the “standard” RCA phono, which was really only designed as a cheap connector. Some high-end manufacturers still use them.
I too have experienced CD-quality versions from old analog tape masters sounding better than their 96/24 counterparts: in my opinion, the higher resolution can reveal tape noise and distortion in too much detail if it’s not well managed.
Regarding provenance: As I learnt from a document made available by Mark, certain definitions have been agreed upon by major labels…
MQ-P
From a PCM master source 48 kHz/20 bit or higher; (typically 96/24 or 192/24 content)
MQ-A
From an analog master source
MQ-C
From a CD master source (44.1 kHz/16 bit content)
MQ-D
From a DSD/DSF master source (typically 2.8 or 5.6 MHz content)
Surely these need to be legally supported, so that anyone advertising ‘high definition’ or ‘high resolution’ audio is required to state the applicable code for each downloadable album or file, Blu-ray disc, or whatever.
Ponomusic included—I must say that their statement,
“…any PonoMusic purchases by Kickstarter player backers will be upgraded for free should a higher resolution version of that music become available at any time in the future.”, is somewhat vague, in that there’s no pledge that they’ll actually seek out or create these versions, or specify their provenance.
Chris, you are absolutely correct. CD has been vastly over-maligned; if the full capacity of the medium is carefully utilized, a very satisfying listen can be and is had. And yes, if the “worst sound quality” anyone could obtain was good 16/44, the a national holiday should be declared! mark makes his points, but it’s a bit over the top. It should never have been called ‘high-res-‘, since that has proven to promote hair-splitting and possibly de-rail the whole thing. It should have simply been called “First Gen Sound,” with the basic requirement being a faithful copy of the master tape that preserves master tape grade audio quality. SACD often suggests ‘master tape’, sonically, as does all the high-rate PCM formats and DSD done well. Neither CD nor vinyl can clone a master tape; with the known limitations why then do folks attack media which have given them decades of pleasure? Best, Craig
I completely agree…CDs can provide a really terrific sonic experience.
Yes, wouldn’t it be great to have the equivalent of THX certified for the music industry. While standards aren’t perfect they do at least ensure a certain level of quality.
For people who are interested in listening to better quality music it is very hit and miss at the moment. Unfortunately, the well recorded music tends to be esoteric things that I am not all that interested in listening to.
Perhaps Mark, you should draft that paper for UHRA standards. At least, there would be a reference for some audiophiles. Personally, I do not mind mixes with little dynamics for cars and ear buds on the run, as long as an audiophile mix does exist. … and of course a surround mix.
I’m truing to write a book to be called High-Resolution Audio: Demsytified…we’ll see.
If you do get round to writing the book, please try to keep it as all encompassing with regard to the demographic and not cater to the lowest common denominator.
It will be an information report more than an advocacy text.
A well-known LA sound engineer that would support your efforts is Bill Schnee.
Bill Schnee: http://bravurarecords.com/a_schnee.html
I met Bill at the recent AES convention and am aware of his work in the past and with Bravura records…thanks.
The new BS 1770-2 / R128 broadcasting standards could be adopted without modification for recorded audio products, and would bring a huge change straight away. After all, how do you make a product sound loud and impactful when the average level is -23 dB LUFS? By making it dynamic!
I’ll investigate the broadcasting standards. However, my experience tells me that broadcasting is as bad as music when it comes to fidelity.
I think you will find them pretty exciting in the face of the impact of the loudness war.
Maybe start here, Mark: https://tech.ebu.ch/docs/techreview/trev_2010-Q3_loudness_Camerer.pdf
Imagine if audio recordings were all produced to the same target level (-23 LUFS)? It should have been organised and happening by now; start pushing!
Wouldn’t a large part of THX specs make for a great outline of a wonderful sounding HiFi?
It could but music people won’t go there because they trade on their uniqueness.
Mark wrote:
“It might come as surprise but as I pointed out the other day, there is no standard for fidelity or specifications for music like there is for film mixes. Not that the film standard is perfect but at least they’ve made an effort.”
Interestingly this these was discussed in the latest show on Home Theater Geeks – Scott Wilkingson with his guest Brian McCarty.
If you (from experience) didn’t already know how bad the ‘audio-situation’ in cinemas is, you can learn about it here:
http://twit.tv/show/home-theater-geeks/229
Not only that the standards are old and ‘false’, they aren’t either followed in in dubbing rooms and cinemas.
There is even a ‘loudness war’ going on there too……
Brian McCarty descibes the situation as ‘chaotic’!
Much work to be done there for the future.
Some kind of specifications in the ‘pure’ audio world would be appropriate too, I guess.
So keep up your effords, Mark – please.
But don’t expect any wonders.
I’ll listen to the show…thanks.
Regarding an audio recording quality database, I assume you’re aware of this:
http://dr.loudness-war.info/
Measurements are made using the foobar2000 software.
I am aware of the database…and I think is reasonably well done. But it just doesn’t go far enough. The plan for the HRADB is more information, reviews, opinion, technical facts etc.
Hello Mark, Thank you for all your hard work on this topic and others. I know we’re a niche market but at least we are a market. It would be sad to spend a lot of money on a great sound system only to discover there is nothing worthy to play on it.
A couple of questions. I love movie soundtracks. I consider them the symphonies and operas of our day. There was a brief moment when soundtracks were released in multi channel surround on SACD and DVD-A. That has disappeared.
Question 1: If music is recorded digitally, could not an engineer “rescue” that music and re-engineer it to a much higher standard? I know it may not be a commercially viable effort in many cases. But, could it be done?
Question 2: Blu ray movies are generally released with DTS HD Master Audio or Dolby True HD. But, these are just containers. How would I know the recording quality of the music inside these containers?
Thanks again.
Thanks…for the comments. I love a well done score as well but after listening to a bunch of classical symphonies and works this morning in my DIgital Media and the Arts course, I wouldn’t put a John Williams score on a par with the great masterpieces of Beethoven, Mozart, Berlioz and Brahms.
The specifications of a digital file determines the fidelity of that file AND nothing can be done to improve the core fidelity. Not oversampling, not upsampling, not converting it to DSD, period. You can change the sound of the files…but that’s another thing.
You can’t know the original format of a soundtrack. It’s likely a 48 kHz/24-bit .WAV file that’s been mastered…but it could be analog tape or even three stripe mag.
And that’s what dCS claims: Upsampling somehow lets reveal in a digital recording all the details of master tape. The same effect as when a sound is being recorded with very high sample rate. Is this a right change ?
Moreover, file specs don’t contain info on whether is the track noise-shaped and/or dithered, and how heavily, either signed or float, which all is quite important when it comes to assessment of fidelity.
> “rescue” that music and re-engineer it to a much higher standard
I do it prior to each listening session & the end result justifies itself completely .