Due to the necessity to learn and understand how mixing and mastering works in the real world it’s time to discuss loudness standards and for me to figure out which ones are best employed where so fit the specifications. I order to do this I’ve had a long read on different specifications and will try to summarize them below. First off however, I’m going to quickly explain LKFS, LUFS, and LU.
LKFS, LUFS, and LU.
LKFS is an abbreviation for Loudness, K-weighed, relative to Full Scale which is the standard measurement to normalize audio for broadcast within ITU BS.1770 and ATSC A/85. LUFS stands for Loudness Units Full Scale and is employed by EBU among others. Both are absolute measurer meanwhile LU, Loudness Units, have been implemented to get a relative measure. One unit equals one dB and this is true for all three methods.
LKFS is measured by calculating the average amplitude over the duration of the video although, EBU has presented the ‘EBU Mode’ which allows for Momentary at 400ms, Short term at 3s, and Integrated which measures the full duration. You can read more about this in EBU Tech 3341.
When measuring the audio level, it is a standard to use K-weighting which is a filter curve that aims to bridge subjective impression and objective measurement. What’s being measured here is known as the perceived loudness and thanks to CRC and McGill University among others we now have an adequate curve to do so with.
Loudness Range, Program Loudness, and Descriptors
Loudness Range, LRA, measures the lowest to the highest parts of media. By leaving out the lowest 10% and the highest 5% the measurement ensures that high peaking sounds, such as explosions or high-pitched screams, and extended quiet parts are not misrepresented while calculating the average level of the media. This is calculated in LU.
Program Loudness calculates the average amplitude of media in LKFS/LUFS.
Descriptors are the above values shown by loudness meters that employs the EBU Mode which I mentioned earlier.
Gates directly affect the Program Loudness by pausing when the amplitude of a media drops below -10 LU which is very interesting due to its allowing of loudness aligning cross-genre material and normalizing it to the same level.
As per the title this is the target levels where the different loudness standards want the media to sit at. EBUs target levels is at -23 LUFS and they make us of a gate to allow for loudness aligning meanwhile ATSC A/85 aims for -24 LKFS and does currently not use a gate. If EBU removed their gate their target level would per default become -24 LUFS.
As the old way of metering with sample-peak meters while using the newer loudness standards can result in clipping the use of true-peak meters are today a necessity to maintain quality audio. These are commonly implemented into loudness meters and differs in an interesting way by measuring inter sample-peaks rather than only sample-peaks. These can find peaks that appear in-between samples which the sample-peak meter simply can’t see. This helps the normalizing process by ensuring that there are no hidden peaks that can cause distortion on broadcasted media.
Onwards to loudness standards.
International Telecommunication Union is sets globally recognized broadcast standards and most other specifications are based on their algorithms. Their specifications are however, being processed for update due to EBUs addition of new tools.
The European Broadcasting Union are basing the R128 on BS.1770 as well as setting new standards by creating new tools to improve the normalization process. As mentioned above the specifications is -23 LUFS with a gate.
The USA uses the ATSC A/85 and is as well based on BS.1770. It uses the full-length measurement for commercials and the speech based method for regular programmes. Its target level sits at -24 LKFS for both methods.
The Japanese version based of BS.1770 that is using a gate at -24 LUFS.
This is the current Australian loudness standard which as well is based upon the BS.1770 specification and suggests that both and full mix and speech based method to normalizing audio however, shorter programs ought to be measured in the universal method of full mix measuring.
My conclusion of this is that it is more viable for me to use the OP-59 specification now as I’m currently located in Australia. I believe that I will have to adapt to the specifications depending on where my target audience is located. The ITU BS.1770-4 specification might be the most appropriate if I am to start looking to online distribution of my work which will make me look in to chosen platforms own target levels e.g. YouTube has -14 LUFS.
EBU. (n/d). Everything about 23 LUFS and EBU R128 broadcast audio requirements and loudness control. Retrieved from http://www.r128audio.com/
EBU. (n/d). LOUDNESS (PLOUD) Helps Members to understand loudness metering and levelling. Retrieved from https://tech.ebu.ch/loudness
ATSC. (2013). ATSC Recommended Practice: Techniques for Establishing and Maintaining
Audio Loudness for Digital Television (A/85:2013). Retrieved from https://www.atsc.org/wp-content/uploads/2015/03/Techniques-for-establishing-and-maintaining-audio-loudness.pdf
Free TV Australia. (2013). FREE TV AUSTRALIA OPERATIONAL PRACTICE OP- 59. Retrieved from http://www.freetv.com.au/media/Engineering/Free_TV_OP59_Measurement_and_Managemnt_of_Loudness_for_TV_Broadcasting_Issue_2_December%202012.pdf
Jordan. L. (2016, February 7). Premiere Pro CC: Loudness Radar and Average Audio Levels [u]. Retrieved from https://larryjordan.com/articles/premiere-pro-cc-loudness-radar-and-average-audio-levels/
TCElectronics. (n/d). What Is Loudness And Why Is It Important?. Retrieved from http://www.tcelectronic.com/loudness/loudness-explained/
TCElectronics. (n/d). Broadcast Standards. Retrieved from http://www.tcelectronic.com/loudness/broadcast-standards/
Masteringthemix. (2016, May 8). Mastering audio for Soundcloud, iTunes, Spotify and Youtube. [Web Log Post]. Retrieved from https://www.masteringthemix.com/blogs/learn/76296773-mastering-audio-for-soundcloud-itunes-spotify-and-youtube