The switch from audio peak-normalization to loudness normalization is one of the biggest revolution in professional audio. It is important for broadcasters to be aware of the loudness paradigm.
The EBU PLoud group created the popular EBU R 128 Loudness Recommendation, making sure there is a common, vendor-independent and relatively simple way to measure loudness. Ploud is part of the EBU's Strategic Programme on Production.
If you are interested in Loudness (PLOUD), join our group on this topic and participate in the discussions. Due to the massive interest in this topic, group participation in principle is limited to broadcasters and equipment manufacturers. Other organizations (e.g. post houses) are encouraged to contact the national broadcaster they are delivering to for loudness guidance in their country.
In August 2010, the EBU published the first version of its Loudness Recommendation EBU R128 (now in version 4.0). It tells how broadcasters can measure and normalise audio using Loudness meters instead of Peak Meters (PPMs) only, as has been common practice.
The switch from audio peak-normalization to loudness normalization is probably the biggest revolution in professional audio of the last decades. It is important for broadcasters to be aware of the loudness paradigm and how to adopt their systems and working practices accordingly.
Basically EBU R 128 recommends to normalize audio at -23 LUFS ±0.5 LU (±1 LU for live programmes), measured with a relative gate at -10 LU. The metering approach can be used with virtually all material. To make sure meters from different manufacturers provide the same reading, EBU Tech 3341 specifies the 'EBU Mode', which includes a Momentary (400 ms), Short term (3s) and Integrated (from start to stop) meter. Many vendors support 'EBU Mode' in their products.
EBU R 128 is the result of several years of intense work by the audio experts in the EBU PLOUD Group, led by Florian Camerer (ORF). The Recommendation is accompanied by specific guidance for short-form content (EBU R 128s1), loudness in streaming (EBU R 128s2), a Loudness Metering specification (EBU Tech 3341), a Loudness Range descriptor (EBU Tech 3342), Production Guidelines (EBU Tech 3343) and Distribution & Reproduction Guidelines (EBU Tech 3344). An EBU Technical Review Article describing the fundamental change in audio in broadcasting is available: On the way to Loudness Nirvana, as is an article on distribution aspects: Loudness in Distribution.
The EBU Loudness test set v5.0 offers various sequences to test loudness meters for (minimal) compliance. It also includes a pink noise signal to help set a correct reference listening level as per EBU Tech 3343 (paragraph 8.2). The reference listening level signal is available as a separate download as well: 500-2000 Hz monophonic pink noise @ -23 LUFS.
We are aware of the following parties supporting EBU R 128 in products/tools.*
* Disclaimer Above list of parties providing EBU R 128 products/tools is based on information provided by these organisations themselves, such as publicly available product specifications. Please note that the EBU does NOT verify conformance of the implementations. Inclusion in the list also does NOT mean the EBU recommends the party or its product(s)/tool(s). Users should verify the suitability of any product/tool themselves. Where tools have no clear company owner, the tool name is used.
What exactly does the abbreviation PLOUD stand for?
This abbreviation is based on the EBU TECHNICAL working structure at the time of creation of the PLOUD Group. The P originally referred to "Production", as the Group was managed by the Production Management Committee. The LOUD part should be obvious :) Back then the exact spelling was P/LOUD. After a reorganisation of the EBU's work, the slashes were dropped, but for PLOUD the P was kept - as the name was so well-known already. You could regard it to stand for "Project", or as some have proposed, for "Practical", because the Group has a strong focus on providing solutions for practical use.
You recommend a target level of -23 LUFS. How does it relate to -24 LKFS?
It is closer than you may think. The EBU recommends to normalise audio according to EBU R 128, which specifies a target level of -23 LUFS. In the USA the ATSC recommends -24 LKFS. Besides this 1 'dB' difference and the use of LKFS instead of LUFS (which is only a difference in naming), the idea is quite similar. But only comparing the target levels is not the full story, as the specs also differ in the tolerances and the maximum True Peak Levels which are allowed. One could argue the EBU spec allows for a more predictable approach, as it has a smaller tolerance window.
What is the difference between LUFS and LU?
LUFS is the unit that is used to express loudness levels on an absolute scale, while LU is the unit for differences between loudness levels, in other words, loudness levels on a relative scale. So, a programme that has a loudness level of -23 LUFS is 2 LU quieter than a programme that has a loudness level -21 LUFS. LU can also be used as the units for loudness levels relative to the target level. It was anticipated that common usage would be, "You are 2 LU low," rather than "You are at -25 LUFS."
If you are familiar with the decibel (dB) as a unit, you will know that a dB is an expression of the ratio of two levels - the level to be described, and a reference level. The postfix to the dB tells you the reference level, for example, dBm is referenced to 1 milliwatt, dBu to 0.775Vrms. LUFS is a measurement on a decibel scale and is relative to the loudness level of stereo (front left and front right) 1kHz tone peaking at 0dBFS.
I like compressed sounds. Can I still use dynamic compression?
Yes, as much as you like, but it won't make you sound much louder. The more you compress, the lower the fader will have to be to reach the same Loudness target level (-23 LUFS). So you can use dynamic compression, but not misuse it to gain loudness.
Do you encourage the use of a lot of dynamic compression?
No. We believe dynamic compression is used too much in current practice, reducing the quality of productions. The good news is that by using loudness normalisation, audio engineers who compress less are not 'punished' by losing loudness. So it encourages the artistic use of the available dynamic range, without pre-describing the amount of dynamic compression that can be used.
Are there tools to measure the loudness range of my signal?
Is the integrated loudness measurement known before the end of the measurement?
No. The final integrated loudness level can only be exactly known *after* the integrated measurement has ended, as it depends on all audio blocks between 'start' and 'stop' having been measured in an internal, ungated measurement first (to be able to determine which blocks are to be gated out by the relative gate).
Does this mean the relative gate cannot be used for live measurements?
You can still use an integrated meter to measure live programmes. The meter simply has to take into account the whole programme, from when you started measuring up to the present time. Basically the relative gate goes 'up and down' with the signal as the programme gets closer to the end, and audio blocks are excluded depending on its current value. Note however that no audio blocks are discarded until the final integrated loudness measurement has been done (as the final relative gate level will only be known at that point in time).
Should the relative gate threshold be -8 LU or -10 LU?
- 10 LU. Originally the EBU had specified -8 LU, but since August 2011, the gate is at -10 LU. This means the EBU and the ITU specification are using the same relative gating threshold.
Should the International Sound (IS) be at -23 LUFS too?
Yes. There are no specific rules provided for IS (without voice-over!), such as is for example provided with soccer matches (audience + pitch sounds only). Various audio mixers in the PLOUD Group have argued that the normal rules should indeed apply to IS too, as this makes for a simpler workflow and more predictable results.
I see people claim R 128 compliance and EBU mode compliance. What's the difference?
Compliance with EBU R 128 means working with a target loudness of -23 LUFS, a maximum true peak level less than -1 dBTP, and using loudness range as a guide. So, a programme can be compliant, or a loudness process, used to normalize programmes to meet the target level and treu peak requirements, could be compliant.
EBU mode specifies integration times and scales for a meter based on ITU-R BS.1770-4. A meter can claim to be EBU mode if it offers the 3 integration times of 400ms, 3s, and "start to stop" integrated, and uses one of the permitted absolute or relative +9 or +18 scales defined in EBU Tech 3341, and measures loudness range as defined in EBU Tech 3342.
Does EBU R 128 also apply to radio?
The EBU's loudness work is not limited to television alone, although it is fair to say that most focus has been on the implementation of EBU R 128 for television. You can find specific information on the use of EBU R 128 for radio in EBU Tech 3344.
Some EBU Members have already started to introduce loudness normalisation (using EBU R128) to radio production and/or broadcasting. In Norway for example the DAB/DAB+ network is loudness normalised since early February 2012:
- Brief explanation by Bjørn Aarseth (NRK)
- Audio example BEFORE EBU R 128
- Audio example AFTER EBU R 128
- DAB in Norway
If you are a radio broadcaster or equipment manufacturer, please join the PLOUD Group if you want to participate in this work sharing your experiences, ideas, concerns, etc.
Do you offer translations of the Loudness Publications?
EBU R 128 and related documents have been translated into several languages by EBU Members and other users. Simply go to the relevant publication in our Publications section and you will see the available translations at the bottom of the page.
If you have any further questions on EBU R 128 or related topics, which are not listed above, or if you intend to implement EBU R 128 Loudness normalisation in your distribution chain, then please send us an e-mail via: firstname.lastname@example.org.