Video Games and Loudness Standards: Interview with Sony’s Garry Taylor
The issues of loudness and dynamic range are common across all audio/visual media, but recently this conversation has been gaining traction within the game audio community. Following his presentation ‘Fighting the Loudness War’ at the Develop conference in early July, I contacted Sony Computer Entertainment Europe’s Audio Director Garry Taylor to discuss the subject further.
DS: First of all, can you tell us a little about how yourself, and your audio career in the games industry so far?
I left school at 16 and joined a band as a bassist. As well as constant gigging I also taught myself my way around a mixing desk and started engineering live bands at my local venue, The Square in Harlow, UK. After 10 years gigging and engineering bands, both live and in the studio, I bumped into a game developer friend who asked me if I wanted to do some music for a project he was working on. I enjoyed myself on that project and decided I wanted to work on games full time. I offered to work for him for free for a couple of months, and in those two months made myself invaluable, after which he offered me a full-time job. I stayed at Mythos Games (creators of X-Com) for 4 years before joining Sony Computer Entertainment Europe (SCEE) as a sound designer in 2001. After working at SCEE’s London Studio for 5 years, I moved to Cambridge to manage the audio team there. In the last couple of years I have taken responsibility for audio for SCEE’s London, Cambridge, Liverpool and Evolution Studios.
DS: Could you tell us about where your interest in loudness standards stems from? What issues are being caused by a lack of any kind of standard in games at the moment?
GT: The main problem for me was the complete lack of consistency between different titles on PlayStation, especially the audio that accompanies the icons on the Xross Media Bar (XMB). As a front end, I felt the XMB problem reflected very badly on PlayStation generally and needed to be fixed. But without any official guidance, there was very little that audio developers could do to counter the requests from producers to increase the volume on titles.
DS: Surely loudness standards are more applicable to a medium such as Television, and the infamous ‘brick wall’ limiting of advertisements?
GT: The level of adverts has been, and continues to be a big problem for consumers. But loudness is an issue for anywhere where many different 3 rd party groups are creating content for a single platform. Recommendation BS1770 is gaining ground in broadcast now, and any tools that help us in game audio should be embraced.
DS: You are a vocal advocate of loudness standards in game audio. Can you give our readers a brief overview of why this is an important topic, and what the end results would be if a standard was widely adopted?
GT: Ideally, consumers shouldn’t have to jump for the volume control when switching between different games or any other media playable on Sony platforms. Ideally all games should be a consistent average loudness regardless of the dynamic range of the content. In my view a lack of standards in games has led to bad audio engineering practices throughout the industry, and this is something I feel needs to be addressed.
DS: At Develop this year, you unveiled your work as part of the Sony Worldwide Studios Audio Standards Working Group (ASWG), can you tell us about the recommendations the group has put forward, and how you decided on those recommendations?
GT: Over the last year our group, consisting of senior audio engineers from the US, Europe and Japan, have measured over 120 games, films, TV shows and trailers. We spent time listening to a wide variety of different material at our studios in Cambridge, San Diego and Tokyo. We also talked to all of our internal teams (at 14 internal Sony studios), and got their reaction to loudness standards generally. We are also speaking to all the other groups within Sony that our recommendations will affect such as Sony Computer Entertainment Inc (SCEI), Tools and Tech and First Party Quality Assurance (FPQA). The recommendations from Sony Worldwide Studio’s Audio Standards Working Group is that all content should measure an average of -23 LUFS (Loudness Units, relative to digital Full Scale) for PS3 and -18 LUFS for PSVita, with a tolerance of +/-2 LU. These will bring us into line with broadcast specs for the US, Europe and Japan, which is important as consumers enjoy more than just games on our platforms.
DS: During your session at Develop, you discussed getting all Sony first party studios on board with the recommendations. How has the reception been from Sony’s internal teams? Have you encountered any resistance from those studios?
GT: The reaction from the audio teams within Sony has been universally positive. Although SCEI, producers and FPQA have raised very good questions about how the teams will conform to any recommendations, the reaction from those groups has also been very positive, to the point where we are now trialling loudness metering at FPQA, which is very encouraging.
DS: Will these recommendations become mandatory on Playstation platforms for third party developers? A benefit of this would be the potential for multi-platform titles to adopt these standards, raising awareness and improving the consumer experience in the wider games industry.
GT: SCEI is the division of Sony PlayStation which sets Technical Requirement Checklists (TRCs) for all 3 rd party development. Our standards group is part of Sony Worldwide Studios, so we are separate groups. At the moment there is no plan enforce the recommendations as TRCs, however, we are in constant communication with SCEI, and they are happy with the progress we’ve made so far. I am also on the committee of the Interactive Entertainment Sound Developers (IESD) group, so I am part of a conversation on loudness with the wider industry outside of Sony.
DS: How is the loudness of a piece of content measured? Are any specific tools or equipment required?
GT: We recommend avoiding AV Amps as part of the signal chain as amps tend not to give a 1:1 signal through their pre-amp outputs. At Sony we are using the Arvus HDMI to AES/EBU converters (www.arvusgroup.com) into a DAW or hardware meter, although at $1500 this may not be possible for developers on a budget. For a more cost effective solution I’d recommend the Gefen HDMI-2-HDMIAUD boxes at about $250. This will enable audio teams to get the output of their game into a DAW so loudness can be measured with any BS1770-2 compatible plugin. We recommend that teams measure their titles for a minimum of 30 minutes, with no maximum, and that the parts of any titles measured should be a representative cross-section of all different parts of the title, in terms of gameplay.
DS: During the Audio Standards Working Group’s study, were there any particularly bad examples of lack of consideration for loudness?
GT: There were a number of games that were extremely loud. I’m not going to name names. There were a number of developers that are already pretty much on the mark in terms of our recommendations. Media Molecule and Rockstar are good examples of developers that are already producing titles that conform to our specs, both in terms of average loudness and dynamic range.
DS: What are the main challenges that are currently being presented to implementing loudness standards? Is there a potential for a loudness war to emerge in games, as seen in commercial music production in the last 20 years?
GT: There are two main challenges. One is of education, and the second is reproducibility and consistency when measuring non-linear material. The situation we’re in at the moment isn’t a loudness war as such, but more like the Wild West, where most developers are just doing their own thing, leading to a massive variation in the loudness of content generally. As I mentioned, audio teams have been very receptive generally, but it’s producers that need to be educated in order to understand how reasonable levels and reasonable dynamic range can help their titles sound fantastic. Audio developers have been after guidelines on this for years to back up their own assertions that having content too loud, and as a consequence, having a limited dynamic range is not conducive to good engineering practices and good sounding games.
DS: Are there any recent releases which you feel are leading the way in this area? Some game audio professionals are thinking more about dynamic range in their titles, but is adhering to a standard of loudness an emerging trend?
T: Aside from the titles coming out of our studios which we have direct control of, Rockstar Games deserve a shout-out. Both GTA4 and Red Dead Redemption were very well mixed and levelled and adhere to our specs long before we started our study and made our recommendations.
DS: How can audio content creators incorporate awareness of the proposed standards into their work? Are there any analysis tools or reference points that can be used to guide this?
GT: I would encourage audio teams to speak to their producers and make them aware of the ASWG loudness recommendations. As far as analysis tools, I personally have yet to find anything close to the Flux Pure Analyzer application for measuring loudness, spectral analysis, true peak, dynamic range and other visualisation tools. As far as loudness metering generally, Dolby Media Meter 2, Nugen VizLM, Waves WLM, and Steinberg SLM-128 (free to Nuendo and Cubase users) are all very good.
DS: As Audio Director at SCEE, have you been able to introduce some of these concepts to the hardware teams, getting such standards implemented on current, or future hardware? Is there a possibility of perhaps including tools for monitoring in the SDK, or other methods of support?
GT: We, the Audio Standards Working Group, have been speaking to SCEI about our recommendations and about the needs of audio developers in the industry. If loudness standards are to be adopted by all developers, we need to make sure that they have the guidance and tools they need to make compliance as easy and as cheap as possible. We are currently in discussion with SCEI about how this is best achieved.
DS: Finally, where can people go for more information on this subject?
GT: The ITU paper BS1770-2 is available here: http://www.itu.int/dms_pubrec/itu-r/rec/bs/R-REC-BS.1770-2-201103-I!!PDF-E.pdf and the EBU documents on loudness and dynamic range are here: http://tech.ebu.ch/loudness/
If registered developers have questions, they can post to the DevNet audio forums here:
DS: Thank you very much for your time, Garry