Is this whole music policy report the result of a calculator slip?

<![CDATA[Today my attention was drawn to a widely-covered (link, link, link) MusicTank energy report with an alarming headline statistic – Streaming media could have larger carbon footprint than plastic discs – and, specifically, streaming an album 27 times could use more energy than producing and shipping the physical product.

Wow, sounds alarming. Could that be true?

Digging into the actual report I found the quote on p14: “…streaming or downloading 12 tracks, without compression, just 27 times by one user would, in energy terms, equate to the production and shipping of one physical 12-track CD album.”

Ok, so there’s a major caveat in there – “without compression” – which almost never happens in the real world and overstates the case seven-fold. But even so, downloading the same MP3 album ‘just’ 189 times (27 x 7) supposedly uses as much energy as physical production and shipping… still hugely surprising.

Where did these figures come from? Page 13 of the report shows some comparisons in the form of equivalent “light bulb hours”. Apparently a CD takes 38 “light bulb” hours worth of energy to produce and ship, while an uncompressed 12 track WAV download consumes 88 “light bulb” seconds.

Converting those to the same units gives (38hrs x 60 mins x 60 secs =) 136,800 seconds for the CD, and 88 seconds for the uncompressed 12 track download.

Wait a minute, wasn’t the CD supposed to consume just 27 times the energy of an uncompressed album download? From the reports own figures it looks like the CD uses 1,555 times more energy (136800 / 88)!

Is the whole report based on a calculator slip? If the uncompressed WAVs instead consumed 88 light bulb minutes their figures would have been roughly right (38hrs x 60 = 2280 mins… divided by 88 mins =) 26 times bigger…. very close to the 27 times of the report. But that would be equating minutes to seconds, putting the whole comparison out by a huge magnitude.

I’m still ignoring the potentially confounding flaw in the comparison – the source of the energy cost figures isn’t clear but it looks like they’re based on data from 2010 at best. By the reports own admission, the energy required to transmit 1GB of data halved between 2008 and 2010.

Since data transmission rates historically improve on an exponential curve against broadly flat energy usage, it’s fair to assume that the energy cost to transmit an album will be much lower today. And even lower tomorrow.

The twisted conclusion that the report is trying to reach is that everyone having “all the worlds music” stored on a memory card may soon be more efficient than transmitting that music over the air.

Apart from being based on figures which are just plain wrong, this conveniently ignores several points:

1. The majority of music purchased is new/current music, which would have to be downloaded to the memory card in the first place. The report suggests that all new music could be downloaded to each user ‘only once’, but this would be even more wasteful – by the reports own admission we’re talking about roughly 342 years of non-stop music to date, and there are years-worth of new tunes being recorded every week.

2. It would be more efficient to stream music constantly from the users birth to death than it would be to transmit hundreds of years worth of listening updates that a particular listener will never consume.

3. As data networks improve, more and more mobile listening is to ‘live’ channels – 1,000’s of global radio stations, Podcasts, etc. This type of listening doesn’t fit the download once, listen forever model.

4. The ‘store everything’ model of the report assumes AAC compression. Using the same level of compression on the downloaded versions (as almost every online music download currently does) would yield a seven-fold improvement – i.e. downloads being 10,000 times more energy efficient than CDs.

I could go on, but hopefully that’s a start. MusicTank – would you care to comment?

Update (14 Sept): Following my blog post, MusicTank have withdrawn the original report and replaced it with a version with 88 seconds changed to minutes. This makes the numbers square but, as noted in my comment below, it still puts their figures out of step with other comparable figures by more than an order of magnitude.

3 thoughts on “Is this whole music policy report the result of a calculator slip?

  1. Hi Grant.

    Thanks for taking the time to go through this. You have picked up a typo in the report which we are in the process of amending. You are absolutely correct, an uncompressed 12 track WAV download consumes 88 “light bulb” minutes, not seconds.

    One of our team will be in touch again with full references for the numbers used, and an answer to your other points. Please bear with us on this, but for the time being be assured the comparison is correct. Thanks very much!

    Best

    Sam
    Associate Director, MusicTank

    Like

  2. Hi Sam,

    Interesting – thanks for the feedback and note about the typo. The numbers do square better at 88 minutes, although it does perhaps leave a question over the general “light bulb” measurement. It’ll be interesting to see the sources for those figures.

    For example, the quoted time of 38 light bulb hours to manufacture and ship a CD suggests an energy requirement of 380 Wh, which seems low.

    The closest comparison I could find is Google’s figure that it takes 3KG of CO2 to produce and ship a DVD. Turning their figure back into watts using the DECC standard of 0.542 kg CO2 per kWh we get 5.5 kWh.

    I only mention that because, turning their figure back into light bulbs, it would suggest 550 hours to produce their DVD against 38 hours to produce the MusicTank CD. Perhaps you have a more efficient disc press or distribution network?

    Either way, it puts the difference between physical and virtual distribution back into the 300-400 times less efficient range, rather than in the tens.

    Cheers,
    Grant

    Like

  3. Hi Grant,

    Thanks for bearing with us – the data on which these calculations are based is correct, with the factor of ’27-times’ valid. The figures in the infographic have been corrected.

    The report’s primary purpose is to focus discussion and (hopefully) stimulate much-needed and more extensive research into how we might meet industry forecasts of accelerating data traffic which, some say, will be exponential. It also considers future new consumer storage solutions and the challenges this might give the industry.

    Regardless of increasing technological efficiency, the report questions whether newer energy efficient servers and transmission technologies can forever compensate predicted data traffic growth which many sources predict will become exponential – therefore it cannot be assumed that the increase in data traffic for the next 10 years can be compensated by more energy efficient networks.

    The main issue – the energy/lightbulb hour calculation – is based on energy data published in the ‘Shipping To Streaming’ paper, duly referenced in our own report. This is one of the most authoritative studies available and serves as a reliable source of baseline data.

    You can access that report here (http://dl.acm.org/citation.cfm?id=1851304&bnc=1); the download energy calculation used data in Table 3, p64 (Energy Costs: Non-Energy Optimized Streaming Method), the comparison with physical (CD) used data from Table 2 p63 Energy Costs: DVD Shipping Method.

    Our report also references the fact that the energy consumption figures (figure 3, p13) show an energy consumption rate 10 times lower than those found in a previous and again widely acknowledged study conducted some years earlier (Koomey in 2004 and updated again in 2008 by Taylor & Koomey). In that respect, our report’s energy consumption figures are truly conservative and supports your view that technology is becoming more efficient every year.

    Inclusion of uncompressed files in our analysis is, in our view, essential – as the report notes on p14, advances in network infrastructure speeds in excess of 100 mbps (wireless) mean that audio files would no longer need to be compressed in future, including new HD formats, which may be even bigger the current CD wav files.

    Data used throughout the report was the most up-to-date at the time. The bulk of the research was carried out as noted in the report’s Assumptions, remembering that it was based on a small part of extensive three-year future-looking project that began in 2009.

    You raise some interesting points, not least in your second response concerning alternative energy consumption figures based on Google research. Whilst we aren’t in a position to question these (nor would it be appropriate to), it highlights the lack of sufficient baseline data, for what is a very complex topic with multiple touchpoints. Such studies are entirely dependent on a large number of variables, some of which you allude to in your post, with other studies (e.g. Google) receiving any number of interpretations and challenges – from the FT to the paper industry (http://www.paperonline.org/uploads/MR_bad%20for%20climate.pdf).

    We’re pleased and grateful that you’ve raised important questions – our hope is that it will fuel further analysis and inspire debate in addressing these future challenges.

    Kindest regards on behalf of the MusicTank team.

    Jonathan Robinson
    Programme Director, MusicTank

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.