George Massenburg Toronto AES Section September 25 2012
Rob DiVito welcomed everyone. He started by providing some background on Mr. Massenburg:"He is currently Associate (not Adjunct) Professor of Recording Art and Sciences at McGill University in Montreal, Quebec, Canada and visiting lecturer at UCLA and USC in Los Angeles, California; Berklee College of Music in Boston, Massachusetts; and MTSU in Tennessee.
He has been working to qualify extended resolution and bandwidth as a goal of modern professional digital recording standards work, and has worked unceasingly to improve analog-digital-analog analysis and conversion methods. He and GML, Inc. are currently researching extended automated work-surfaces, high resolution graphical interfaces, extensible network automation for audio production environments, and automation data interchange standards".
Rob directed the audience to look over their pdf's for tonight's meeting to see George's extensive industry awards
Rob read some "online regrets" from friends and colleagues unable to attend tonight's meeting, to demonstrate George's wide influence.
George began his presentation by asking the audience to lower their expectations 20%! The subject of disruptive innovation vs. sustaining innovation is different for him because he's been researching other people's work in the field, and seeing where it can be picked up and made use of in making better recorded music. He thanked David Neal, Bob Ludwig, Bill Stewart and Stanley Chu for their help.
He started playing Phil Collins "In The Air Tonight" - a version containing the original dynamic range. He mentioned that he almost drove off road hearing it on the radio for the first time "I couldn't get to a record store fast enough!". He then played a recent 'typically mastered' mp3 version. He had some choice colourful comments for that version - a comment repeated later in the evening, too!
He followed with Metallica's "That Was Just Your Life" - the Guitar Hero version. Then he played the released CD version — displaying the waveforms for the audience: they were practically square waves. "That was a low point we haven't quite come back from yet. We still have some work to do".
Then he played the CD version matched to the level of the Guitar Hero version. Referring to the Guitar Hero version: "It's just BETTER!! Why would anybody make that other choice?!
It's our responsibility as engineers to end this".
He gave a 'visual' history of loudness: from about 1979 to 2009 levels increased 16 dB!
Before playing his next example - Etta James "Don't Cry Baby" - he said "when we used to do mixing(!), we used to move the faders! Not only did we move faders, we heard mic breakup, mic pre's breakup, tape breakup - but it all works because somebody's listening and responding".
Next he played an associate's recent work to represent the modern approach to mixing and recording live. Eric Clapton's "Something You Got". "It's great....It's not clipped....And it's live".
He explained the terms disruptive technology vs sustaining innovation.
Disruptive technology upsets the existing order of things in an industry - citing electro-mechanical calculators and bias-belted tires as examples of industries that collapsed. It usually doesn't attempt to bring better products - sometimes it does, sometimes it doesn't. It redefines the competition. It almost always brings in a better price. Existing companies don't react to disruptive technology until its too late. Here he cited traditional record labels in 1996.
Sustaining innovation targets demanding, high end customers with better performance than what was previously available. Some are gradual improvements and others are breakthrough products. Established competitors almost always win the battle. The strategy entails making a better product at a higher price sold to their best customers. He produced the iPad as an example of this.
Examples of disruptive technology: the Berliner flat disc vs. the Edison cylinder; magnetic recording; the 8-track player; Sony Walkman; Compact Disc; DAT recorders.
In 2011, the 3 (now 2) large record companies announced ceasing production of CD's.
He displayed yearly price charts for the recorded music business ("They're all depressing"): Number of units is going up, but revenue is going down, digital is not replacing the old merchandise model. Subscriptions are flat.
As an aside, he described as appalling, Pandora petitioning congress to lower license rates of Internet radio to that of Pandora's.
He displayed David Neil's chart tracking U.S. recorded music showing recorded music revenue per capita in adjusted 2011 dollars from 1973 to 2009. In dollars per capita buying in 1979 was $63, in 1998 $71, but in 2009 only $26. "This doesn't work."
The modern Fraunhofer HD AAC is brilliant: it's a seamless integration of low bit rates ("which we need sometimes") all the way to 192k/24 7.1 surround in any format. It's too expensive - no one wants to buy it. Adaptive bit rate streaming avoids payments to Fraunhofer.
One of the problems is Audio is always being told 'we don't have enough bandwidth'. But adaptive streaming adapts as the technology changes, and the audio does not drop out which is "important for the user experience". One great benefit is that it has no additional Fraunhofer license fees.
He next discussed Apple encouraging the submission of files in 96/24 format. He disputed claims against using 96/24 formats: hard discs are getting bigger and cheaper, faster; except for North America, networks are capable of distributing high res digital files.
"The worst kind of lie" is 'no one can hear the difference between AAC and hi res files' - because it's "stopping us from making better recordings". George noted there is a growing market for downloadable high res files. There's good reasons for putting high up quality, high res files for the future. He cited the Beatles 'Love' show as an example of an incredible high res, hi-fi experience. "That would never work off an mp3 born master".
He discussed iTunes match, where for a yearly fee, iTunes scans your files and streams substitutions with high res files if they have them on their cloud servers. The flaw is if one has illegal mp3's, those will get matched to the cloud as well.
What it means to us is Apple has moved from disruptive innovation to sustaining technology. They have a lot of cash, are incredibly profitable, they service the high-end sector of the business, and "it's the high end customer that wants incrementally better performance".
This moves us to how we can address loudness; how to adapt to what's coming and use it as a tool to get people to not 'crush' their music.
What's coming is the new loudness constraints. He briefly reviewed the CALM act.
The key is understanding what the terms mean, such as program loudness, loudness range, the distribution of loudness and the maximum true peak level. He mentioned TC Electronics' Thomas Lund's presentation regarding reconstruction filters here contributing to distortion, downstream, (of files with maxed out levels).
George next discussed Rudolph Ortner's masters thesis analyzing over 10,000 of the most popular tunes sold. While maximum true peak and program loudness have gone up in the last 14 years, the loudness range has been rather constant.
It starts here by having a better dialogue with artists to have them understand they're not getting anything more by turning it up. Having level matched comparisons is a necessity in this regard. There's no advantage to having loud masters, because broadcasters will now (or soon) take tracks (like Metallica's) and bring them down to an average level, negating any attempts to be louder.
What else can we do as engineers/mixers to prepare for the high res future?:
In addition to better understanding the new loudness terms, we can migrate workflow to 96/24; consider the channels your work might populate before delivering mixes; postpone decisions on dynamic range until mastering; and print uncompressed versions of your mixes.
As producers, we can (as artists ourselves) put in the "10,000 hours that it takes to get insanely great; to write, perform, and produce much better music".
Here everyone took a break.
From the perspective of disruptive innovation, Rob DiVito asked George regarding Apple's motives (with the iPod for example), if they weren't so much based on a quality of life but more on commerce.
George quoted from Steve Jobs' book stating his motivation was making "insanely great products". To demonstrate this he discussed the iPad — when the product was finished it had square corners and couldn't be picked up with only one hand. This was corrected before its introduction with rounded corners to make it easier to pick up. Not many companies would make such a major change on a developed product. George believed Steve didn't do things only for commercial value.
Mr. Massenburg continued the second half of the meeting talking about some of the projects he was doing. He said he doesn't like doing rock and roll anymore - "You can guys can do it!"
He played a video passage with choir director Eleanor Stubly conducting; followed by a Jazz One excerpt. His favourite example was the McGill student production of La Boheme. He proposed that student (camera) shooters are better than those that come out of news and sports. Essentially they're more involved and passionate with the actual music score. The student performers are closer to the actual age of the opera's characters. George "loves" this much more than rock and roll.
Next, he played a Bon Iver track. He liked it a lot but said there's no way he can make this sound; he's ready to pass the torch to people with new ideas.
He demonstrated the TC Loudness Meters with an Amy Mann track. He pointed out the BS 1770 terms: loudness range and program loudness as key figures. He noted, even though the waveforms are clearly flatlined, it still has a sense of dynamic. He said it's 1 dB back from 'ridiculously too far', referring to Bob Ludwig's mastering presentation whereby one tries to incrementally achieve an extra dB of gain: "there's a point where it's too far".
He played Eric Clapton's piece again through the loudness meters explaining that the loudness range keeps getting bigger and bigger, 'so it's kind of working'. He noted that even though there's compression everywhere, it's not smashed and it feels great.
Playing a Jennifer Warrens, he said the meter was lying. Someone mentioned meter is made for broadcast. Shouldn't if be for broadcast music too?
We can do better by providing masters with more dynamic range.
Closing his presentation, he said world wide music revenue is higher than it ever was — it's just not getting back to the producers and musicians. It's not the same system. "The idea of a label was always a license to steal from artists". It's time to get serious with revenue distribution. In the meantime, acts are making money on the road by selling CD's at their shows. Artists running their own sites is presently the best way to make an income.
Questions and discussions:
One person asked: "How do we know people want 96/24 files?"
George responded that the "postage-stamp audio" does less to convey what's important in music, which is: music identifies who we are in a culture; listening to music we make a connection, as young listeners, that grows more mature as we grow out of adolescence, and we start to have a sense of what music is and how it identifies our culture. What we've done with mp3-type distribution is "squeezed the detail out of it" so that we don't have the "lace and filigree" that we had that we know worked - "we've taken something essential away - we've got to get it back...this (96/24) is the way we're gonna connect with the world".
Another member wondered if kids won't need more than 'high-fidelity' sound; since most of the time they're listening on ear-buds through iPhones?
George felt it wasn't about high-fidelity sound. Stating he mixes sometimes with his Etymotic system, you have to be informed what that environment (ie: ear-buds) is.
Another member stated a problem with his younger clients is that they perceive clean, well-recorded audio as in-authentic, because much of their experience comes from making music from a lo-fi, low tech medium. "Grungy, crappy sound resonates to them as something they and their friends can make".
A retired educator offered that he had students who had never been exposed to decent sound reproduction, so he built them a simple high quality two-way system and had them play their favourite CD's. He stated when they heard music on a decent system, he had converts left, right and centre. They were all out to build better things.
This statement clicked with George.
"That's it!" If we can bring that experience to kids, it would get their attention, and they'd know there was something better.
George did agree with another comment that it has to be the kid's motivation; "the kid has to own it".
"You've got to play stuff to them. We have to take the responsibility". He did not agree with a statement that the artist has to take responsibility because most of them are not going to take that call. "You can't trust artists. We know what sounds good; we've got to do it one on one". He said he doesn't know if there's any other way of doing it.
Someone stated that content providers have done a dis-service because they've underestimated people's ability to appreciate quality when it's given to them. George replied marketing isn't about that. The audience member clarified that there needs to be advocacy at that level to make sure they (ie: content providers) are not breaking that part of the chain (high quality delivery). George related back that marketing people won't let companies put out product they can't make profitably.
The next discussion concerned codecs and George's views on them. Someone wondered where people can download something that's lossless because it's not something that's readily available. George replied that initially producers weren't providing hi-res files to the services that had deals worked out - the licensing has to be worked out to put up hi-res tracks.
Continuing his discussion, users can use the Sonnox Pro-Codec plugin, or do a 'null test' in order to audition the artifacts, and determine which codecs are "particularly heinous" for a file. George cited variable-bit-rate coding as an example of one of those codecs as it tends to shift the artifact in time, so it's not synchronous anymore. The new AAC is "pretty good": 256 AAC is more than 6 dB better than 128k mp3. But it still suffers in "high entropy" or "high complexity - it just crashes".
Asked if there wouldn't be an advantage if there was an industry standard, George replied there is: 256 AAC. "We didn't have any say in it! It's like having Dolby AC-3 on every...DVD because it's the industry standard."
Asked about production credits being included in downloads, George said this is a whole other subject falling under "boring" digital meta-data. It's hard work "and you got to do it". It starts with having good, share-able meta-data. It's all about coming up with a new scheme for a database and the Library of Congress is central in the end.
"Has anyone ever thought of imitating IMDB?" (Internet Movie Data Base). George mentioned AllMusic but it has "a lot of mistakes"; and Gracenote. George felt for the time being artists running their own web sites was the answer. He cited Imogen Heap as a brilliant example.
Finally, someone mentioned the Howard Soroka project which allows both hi and lo res downloads for one price. George reminisced briefly about his associations with Howard.
George closed the evening saying he really appreciated everyone's patience, kindness, great questions and great thoughts; and to "get out there and play some really cool" stuff!
Referring to this as a monumental occasion for the Toronto AES, Rob DiVito presented George with a Toronto AES Certificate — its first re-instatement for many years — and an AES mug.
George was greatly appreciative and thanked everyone again.
