Decibel and Meter – IV

Fig, 1 loudnessnorm.jpg

Before proceeding to analyze the LUFS meter that have defined legal limits and recommendations of use in the production of mix and master audio after years of battles to make music and sounds with the highest values of loudness for the song or audio track that sounds “louder,” it’s good to take a few steps back and see how the concept of loudness has evolved over time.

 

Background

The management of the loudness is a phenomenon born with the first recording systems and mechanical audio playback, such as phonograph and gramophone, in which have a high average recording signal allowed to maintain a SNR (Signal to Noises Ratio) high and therefore a noise of relatively low bottom and allow for a greater clarity of the audio signal.
Subsequently, with the spread of music and acoustic recordings (initially mostly environmental shooting through funnel horns of different sizes as can be seen by looking at the phonographs, then subsequently through the use of electrical systems (around 1925) as environment microphones and close miking and amplification systems), loudness is worked on the same principle of removing the noise from the audio signal even through the environmental structure, then controlled acoustical environments, with defined reverberation times (0.5 – 0.6 ms, T60 ) and in the management of the microphone locations such as to have minor falls or falls possible environmental controlled.

In help also it came the technology with the emergence of the first electromagnetic recording systems for the transfer and etching audio signal before on vinyl records (from 1948) and then magnetic media, which allowed a much higher signal to noise ratio than the phonograph and/or gramophone.
With the emergence and spread of digital technology and the increase in hardware and software performance for both analog and digital processors are allowed considerable potential increased loudness is being recorded that mastering and playback (thanks to the advent of optical media (1982) and digitals), as to retire any obsolete recording equipment first of all phonograph and gramophone (already retired to vinyl exit times), then switch to vinyl and cassette.

The fundamental problem is that evolution even allowing to exploit recording and playback media that are hardware or software with a dynamic potential that can now be used in some cases even beyond the levels of perception and acoustic sensitivity of the ears, from the point of view of Use has been exploited to achieve an ever-higher loudness and impact, neglecting the possible exploitable dynamics (then quality and naturalness of listening).

It follows that today the dynamic level left to the music tracks especially those from pop to rock to metal and disc is comparable to the maximum dynamic level obtainable with the phonograph (this what said of the record as a matter of “market war” in which as It said it is believed that a track with greater loudness sells more than one with less (although this actually has been experienced not be true), in addition to this as a matter of competitiveness (also dubious here), because if the average produces with a certain level of loudness then you need to record and burn an album or single with the same or higher values to be out of business. It ‘also true that if you listen to two songs one with a higher level of loudness than tending to another the focus moves to the song with the highest loudness.

Through analysis of multiple audio files recorded from the birth of the first CDs to date (as discussed below), it was concluded that in more than 30 years the average level ad of music such as pop has had an increase of loudness +20 dB, this has enabled older recordings to be remastered according to the average loudness levels year after year on new albums and collections with the aim to make buying the new album being in comparison with the original much more loud, we are now in loudness trigger levels for digital technology on the market and the resolution of our ears, and so in addition to regulatory issues you think the war to achieve the highest possible loudness is over.
Another problem to increase more and more the level of loudness is the considerable increase of the distortions (especially if you use dynamic processors) and listening fatigue (the first because the excessive compression tends to introduce artifacts in the audio signal, the second because listening to a piece of music without presenting crescendo and diminuendo causes psycho-acoustic stress). Fortunately, as we shall see they are beginning to come out recommendations and raw regulations that seek to regulate the loudness market, considering both the qualitative aspect of the music that the purely commercial.

 

LUFS Meter

In 2006, with the revisions to date, ITU and EBU to better adjust the recommendations and as we see the regulations created to limit the dynamic war that was going to grow in the broadcast and audio productions, have created a standard meter that analyzes the most of the audio signal in compliance with regulations.

ITU with the legislation ITU BS.1770 which regulates these new digital meter, has also introduced a new term the LUFS (Loudness Unit Full Scale), a term currently used exclusively for this type of meter, constructed and regulated with tolerances and calculation methods as legislation to go to adjust the maximum volume of loudness of various music programs as shown in figure 1. So that you are in the Broadcast (television programs, advertising, radio, cinema) or Studio (recording and mastering) is important to have just seen a digital meter updated reference to the standards (generally the digital meter of the present info on the technical characteristics and the various standards for the calculation).

The meter that follow these regulations are also recognizable as called LUFS Meter.

As you can notice all these regulations are mostly for broadcast transmissions are analogue (requires the use of digital meter for loudness relief) or digital and recorded music (studio), there are still at today legal references for the loudness in the live music (only limited by law the maximum and average values ​​of human exposure to noise). In the digital environment a professional technician (for example, through the aid of a digital mixer with a meter that meets these standards or through a meter on an external analyzer that is software or hardware) which tries to follow the rules defined for the broadcast studio and so that also repurpose live environment the same dynamics and loudness of the recording to which people are used to listening before going to a live event. In the case of using analog components (for example, analog mixer) it is well equip with digital meter (that just follow the standards visas) to be interfaced after the master output of the audio matrix (audio mixer). If you use a hardware and/or A / D converter sample to detect the audio signal via the software in the computer is necessary that the quality of both converters as standard.

In figure 2 and figure 3 are some examples of LUFS Meter.

Fig. 2 lufsmeter_1_4_0_main.png

Fig. 3 meters-large.jpg

 

Generally LUFS meter have fixed settings, with a red, yellow or different rms value when the audio level exceeds – 23 dB (Program Loudness or Integrated Loudness), a color for the peak level and one for the level of range rms and lower bars, but in many it is still possible to set colors according to the level reached both rms and peak that other values, as well as time of analysis windows (fig. 4), in such a way as to be able to also use the meter for several references, such as those in figure 1.

Fig. 4 2016-09-10_18-19-41.jpg

Many have meter and loudness meter separated peak.

Links to the detailed explanation of the regulations:

ITU-R REC-BS.1770ITU-R REC-BS.1770

EBU r128-2014

OTHER

The digital meter that comply with the ITU and EBU standards possess an algorithm of different calculation from that seen until now, for the RMS value in consideration of the frequency response in the ear (noise contours) (fig. 5) until a few years ago was used an empirical weighting curve called Leq RLB (fig. 6) that compensates for the measurement whereas the frequencies above 2 kHz with a roll-off starts from 100 Hz, such as to maximize the listening of low frequencies even at low volumes listening as the human ear at low volume has a difference of perception of low frequencies than at high much higher (the bass are perceived with a much lower volume than listening to higher levels).

To date, the weighting curve has been refined and is the Leq R2LB, formally known as Leq K (equivalent level K) (fig. 7).

n.b. Compensation allows during dynamic processing to increase the efficiency of low frequencies before reaching the maximum allowed limit.

Fig. 5 FHfitxa15_ilus_1

(See in detail the equal loudness curves in topic psychoacoustics)

Fig. 6 mm

Fig. 7 slide_6

 

The peak values ​​instead are marked dBTP (True Peak) and the meter scale is extended up to some value above 0 LUFS or dBFS in the event that a meter is applied in the dBFS wording. Compared to traditional meter provide a better indication of digital headroom, detecting via algorithms, those inter-sample peaks with a value higher than 0 dBFS which are the cause of distortion introduced by the codec and the D/A converters.

The traditional professional meter as well LUFS meter that reflect the EBU R128 standards sometimes have the ability to determine the window of time that analyzes the audio signal and the average RMS value (a time window is a must as they analyze each sample the digital audio signal would not reflect the real human ear perception), generally a time average window is 400 ms (momentary loudness with the presence of gate) (generally used) or 3 seconds (short term loudness without gate) (used for comparison or as a second window that is mediated to the short for enhanced security in the return of standard parameters as the short-term is very varies over time). Other meter that do not follow EBU directives may have different time windows.

The meter regulated by these standards for the calculation of the rms value then involve the use of Gating, That is a processor that works, and then operates the calculation and compensation processes only above a certain threshold and not considering the audio under the same threshold, so as to limit and less stress on the calculation processor making it work at 100% only in the useful zone and trying not to consider the pauses and the silences that could alter the final average value (as it happens in the normal meter). Some manufacturers, however, propose alternative control algorithm (some are improvements, other worsening) of the RMS signal and Peak keeping intact the constructional characteristics of the digital meter for regulatory compliance, such as a Short Term instead of the Gating for the control consumption of the processor and the elimination of breaks from the temporal average.
n.b. In traditional digital meter when an audio signal pauses gradually decreases the value of the signal level (according to their performance) until to arrives at – ∞ dBFS, The more the decrease will be quick and respects the immediacy of the pause, and the less it will go to mediate if the sound resumes without the meter returning to 0 to pause.

The Gating used in LUFS meter follows the directions given in the table in Figure 8.

Fig. 8 2016-09-05_18-20-56.jpg

The gating proposed by scholars and various associations are G20 which excludes from the calculation of the rms value the signal that decays of 20 db, the G10 which excludes the signal that decays by 10 dB and the G8 which excludes the signal that decays of 8 dB, the gate followed by ITU BS.1770 standard is the G10, in some LUFS meter however also possible to change this value to have different references. In the table in Figure 8 is a comparison of all these gates also with a traditional meter without gate. The traditional meter as seen in the media include the silence and pauses unlike Gate. A gate may be narrower as the G8 has a focus on the RMS value more targeted also excluding any background noise.

The ITU has opted for the G10 as it was considered more like the kind of ear response to perceive sound and music.

In figure 9 an example of how the RMS value changes depending on the use of Gate or less.

Fig. 9 2016-09-05_19-37-03.jpg

It is noted in figure 9 as the use of a gate (G8) determines an RMS value different from that seen with a standard meter, generally with higher value as the standard meter as seen also average the pauses and the silences.

Another factor generally find in LUFS is the LRA (Loudness Range Amplitude), represents the average variation between the peak value and the RMS and is always measured in LUFS or LU. The loudness range that must have an audio program is always a recommendation and not a law as instead the maximum rms level (- 23 LUFS ).

Some meter LUFS also have another parameter, the LUFK which considers the average RMS value of the duration of the audio program taken into consideration during the measurement, which is also generally the Max Loudness which identifies the maximum average level reached LUFS (the point at which the average value reaches the maximum level).

A LUFS meter has 2 of analysis windows, the Momentary Loudness (average audio samples every 400 ms) and the Short Term (average audio samples every 3 seconds), these values were estimated for the Momentary Loudness as the average of the system perception time human auditory and for the Short Term as a preventive analysis to stabilize and retain the fast variations of LUFS Momentary Loudness values of which should be difficult to make a stable identification of LUFS values averaged over time, works as a sort of prevention in order to limit the Maximum values of loudness that could easily exceed regulatory limits. Average LUFS of LUFS meter is generally an average of these two weighting windows and often called (Program Loudness or Integrated LUFS), but in general also possible to see the values of both windows (fig. 10).

Fig. 10 screen480x480.jpeg

N.B. As per regulation to EBU R128 standards, the Max Loudness for an audio music program can get to – 18 LUFS when measured in Short Term Loudness and – 15 LUFS if measured in Momentary Loudness.

Short Term and Momentary are immediate and display the signal value averaged for the period of the time window. The Integrated Loudness is an average that considers the entire analysis time (Elapsed Time).

The fundamental problem of these meter (although technologically improved over time) is that having a dynamic compensation of the ear’s response standard cannot be properly applied (even if used) to the LFE channel in a surround sound program, which works best with standard meter or meter LUFS processed for the analysis of the LFE channel.

Some may think what they are for all these regulations, may not agree on the maximum RMS values ​​defining them not quality for a correct perception of the audio dynamics for that kind of music program (especially in those many dynamic or high levels of compression that if unrealized through quality instrumentation could lead to generate high levels of harmonic distortion values), or that severely limit the technical and creative skills and quality of a sound engineer. All this is comparable to the reasons why there are laws in the life of every day and to which all do not agree on all, without legislation even in audio environment there would be the great chaos, for many a lot more dirt and poor quality cannot offer never reference points to who hears this sound.

LUFS is therefore an excellent meter that allows you to directly analyze the RMS average of the entire piece without considering the breaks, analysis instead should be done by hand (by selecting only the effective parts where no sound, then unusable for analysis in real-time ) whereas the other types of meter as previously seen, and it is a meter that better reflects the human ear’s perception.

n.b. Considering a meter LUFS for conversion of interfacing with analog devices the 0 dB analog is – 24 LUFS.

In figure 11 other recommendations – standard for normalized loudness of audio programs according to the type of transmission platform.

Fig. 11 2016-09-06_18-54-512016-09-06_19-03-04

It is noted in Figure 11 as the high definition broadcasts for HD televisions but also UHD enable and are technologically prepared for a higher peak and loudness range, having listening a greater dynamic and loudness normalized and less compressed than in listening from mobile devices and non-HD TV (for which more qualitative). Quality due to a higher transmission bitrate (if digital) and in general to a lower background noise with respect to the mobile digital platforms and standard televisions (analog or digital).
The recommendations to obtain a quality audio signal in broadcast as well regulated standards are:

HDTV: True Peak Level Max – 1 LU, Program Loudness – 23 LUFS, Loudness Range <- 20 LU, Max Loudness – 20 LUFS (short program).

No HD TV: True Peak Level Max – 1 LU, Program Loudness – 23 LUFS, Loudness Range <- 12 LU, Max Loudness – 20 LUFS (short program).

Mobile platforms (smartphones, tablets, MP3 players, which includes podcasts and streaming): True Peak Level Max – 1 LU, Program Loudness – 23 LUFS, Loudness Range <- 8 LU, Max Loudness – 20 LUFS (short program).

N.B. For a music program as mentioned above there are some tolerances:

Max Loudness: – 18 LUFS when measured in Short Term Loudness and – 15 LUFS if measured Momentary Loudness.

Others use their standard of ITU BS.1770 always drawing as:

iTunes and iTunes Radio: Program Loudness – 16.5 LUFS (via an internal algorithm standardization of loudness Ext Sound Check always present in the iTunes Radio selectable choice in the iTunes).

Games: The audio for the games instead follows the – 23 LUFS Program Loudness.

Compressed files with AC3 Dialnorm: – 31 LUFS.

Figure 11 shows how useful it is when it comes to gaining low dynamic and peak audio programs, such as non-hd and mobile TV programs, up to a maximum of -20 LUFS allowed for Max Loudness.

N.B. As you will see when you talk about mixing and mastering, creating mix and master reflecting the dynamic as regulations and / or recommendations proposed using LUFS depending on where the audio track’s going to be played, it helps a lot the final overall quality techniques the song because that this hardware processor to play that works and monitors the audio file going to amplify or attenuate the signal level based on how it is read to bring it close to regulatory levels (always according to standard for transmissions broadcast which covers TV, radio and mobile, with limit as mentioned in – 23 LUFS), it is less stressed and works less on their already adjusted files in a dynamic phase of the mix and mastering, so as to pose fewer possible alterations and leaving unchanged the timbral quality of the mix and / or mastering done, especially if the control processor and hardware correction used for playback is not quality.

 

Difference between standardization processes

EBU R128 (Input signal always normalized, saved, and then diffused, no metadata is used).

iTunes and iTunes Radio (The loudness metadata (information about the instantaneous level of each audio sample) come inserted within the AAC file, this metadata are interpreted by the software and using the Sound Check algorithm creates a Gain Shift to reach the normalized level in the process of reproduction, the gain is adjusted automatically during playback), (the fundamental problem is that in iTunes playback can also be performed on uncompressed files as AIFF and WAV files that do not have the ability to carry metadata, in this case creates a metadata package separately added to the stream for encoding by Sound Check).

ATSC (This standard uses the Dialnorm from the Dolby AC3 bitstream, a process similar to the EBU no metadata is used).

Among the three that considered the lowest quality is the audio standard from iTunes and iTunes Radio.

 

Management of the normalization from iTunes and iTunes Radio

Unlike the EBU and ATSC standards, iTunes and iTunes radio as said first read the metadata and then normalize, another problem of this process occurs when you try to read at iTunes such as an Audio CD, the CD does not contain metadata about information loudness and the algorithm is unable to read and perceive the loudness information in real time, so the algorithm needs to cope with a database (called Gracenote database) to detect metadata and apply its standards ( no precise process), if the audio file is not found in the database an attenuation of – 6 dB is applied by default and then back to 0 dB once the track and start over with a new analysis to the next track. In some cases, it carries out error, namely that detects a track within the database, but that is not the one being played, this leads to an incorrect normalization.

For these reasons it is not recommend to play audio CD from iTunes.

Another factor which may prefer other playback systems especially compared to iTunes Radio (then streaming radio) is the type of normalization performed, That is the Single or Track Normalization which gave an average value defined, the process tends to “crush” then compress or expand the levels of the tracks to the set value (in iTunes Radio is automatic and cannot be edited), this makes you lose the tracks the real dynamic difference and mixing environment (for example, a song may be intentionally mixed and mastered with a wider dynamic compared to another, or an average value deliberately lower, all these dynamic variations between a song and another will be lost in Single or Track Normalization, but it is important to know that each song retains its dynamics), these systems generally provide also a limiter or compressor to limit the possible increase of the peak values ​​that could lead to distortion levels.

Broadcast that instead follow the directives EBU or ATSC are using Album Normalization, which allows (when playing an album or a series of audio tracks) to bring (through analysis of an algorithm) the values of the loud audio tracks at the same average level and the values of soft audio tracks at the same average level (the loud with loud and soft with soft), (even this system can submit limiter or compressor in order to prevent phenomena of clips).

The Album Normalization allows the end user more freedom to decide through its desired level of the volume controller for listening without losing dynamics and depth (limited only by the ambient noise and noise of equipment used bottom).

iTunes for playback instead of albums and collections allows the choice between Single or Track Normalization and Album Normalization.

In figure 12 an example of the difference between Single or Track Normalization and Album Normalization:

Fig. 12 2016-10-29_15-20-35.jpg

It can be seen during the Single or Track Normalization or the average value of each track is brought to the same level, while during the Album Normalization are maintained relationships between tracks with more volume (1 – 3) and those with less volume (2 – 4) even if normalized.

 

Other levels of weighting

Some LUFS meter also have the ability to weigh in a different way the average level of the audio signal analyzed as weighing Leq (A – B – C), such as to be able to perform and define different types of loudness according to necessity, such as the average level of an audio signal within an environment in which they must comply with regulations on maximum sound pressure (because the maximum sound pressure in entertainment environment is measured in Leq dBA), (all this achievable calibrating suitably the average level of LUFS meter example with LeqA weighting which reaches the limit level when the signal detected by a measuring sound level meter indicates, for example, 95 dBA, which is the average sound pressure level limit in sight environment), (fig. 13).

Fig. 13 (LUFS Meter with the possibility to weigh the level of the audio signal on Leq, the choice of the weighing type is indicated by “weighting” parameter).

Fig. 13 2016-11-02_11-34-20.jpg

 

Meter for the control of the level of the signal output from the codec

With the advent of numerous formats codecs mp3, aac, mp4, codec for streaming like youtube, spotify, itunes, pandora, etc … which compresses audio with different technologies producing artifacts, distortions and changes of the dBFS level of the original audio signal, most often by increasing their level risking to bring the decoding distortion levels, are born meter that give the ability to predict both the level of dBFS audio signal output from the codec to perform a pre-listening of how this sound will be heard, which visualize the frequency response (fig. 14 – 15 – 16).

Fig. 14 2016-11-14_16-01-20.jpg

Fig. 15 2016-11-14_16-07-06.jpg

Fig. 16 2016-11-14_16-03-06.jpg

 

Loudness War

The history of the recent past to the near future teaches us that the more years pass and prompted greater dynamic compression of audio programs and an increase of RMS value, why?

According to the record companies a piece of music with higher volume (thus increasing the RMS value) sells more than a piece of music at lower volume. To avoid phenomena of war between various music companies and movie studios that could have led to poor quality audio levels (because the more you try to compress and the more will be introduced and will result harmonic distortion factors is the same piece of music that from tool used to compress and so it is necessary to use equipment of the highest quality) (for this reason was born on Mastering, to optimize and work best on the dynamics of the audio file), ITU and EBU standards have begun to define the levels regulatory RMS values ​​used. Through studies it was then realized that the final listener in most cases feel a musical program with a higher RMS value does not bring any benefit in terms of investment. The end user always tends to buy their favorite songs independent of its loudness level than the other or on the same if he would lower RMS value.

Then there are then regulations designed to control the compliance with the same standards of maximum loudness limit, such as CALM (Commercial Advertisement Loudness Mitigation) which obliges the US not to transmit broadcast advertising to a higher RMS level of the transmitted program.

On the other hand we have the fact that music programs like classical music that needs all the possible dynamics for tonal representation of each instrument suffer a deterioration of the final quality having to undergo a hyper compression up to the defined standard values, thus bringing high values ​​of distortion. This in broadcast as the adjustment of the dynamic levels is generally fixed (normalized to – 23 LUFS or in any case cannot exceed this value) for all programs and is revived by dynamic processors before the spread of the transmission, compression placed on audio that already it was working (and therefore compressed) in the recording studio and / or mastering. For live events, teleshopping and promotion are granted fluctuations of +/- 1 dB.

n.b. Legislation that regulates the maximum average volume to – 23 LUFS (exclusive to broadcasting) is not a mere recommendation, but it is a law that, if not respected, can come to light in legal complaints and disputes.

Fortunately, as previously mentioned in live context, there are still no precise rules for which it is still possible to maintain the required quality standards. In the studio, however, it is still possible to adjust the correct dynamic levels by trying to follow the rules defined in the musical program as in Figure 1.

Even the same weighting technique used for RMS value (R2LB) may be a reason for discussion, especially as it does not really respect human perception of sound, but only compensates for the perceived sound tone response in order to change the value RMS of a music program to match the tone values of listening (the technique as seen works well if you consider low listening volumes), thereby altering the perceptive natural response.

At the professional level to be produced and marketed is highly recommended propose projects (even more if you work for broadcast, as may be the radio and television advertising broadcasts) respect the regulatory limits ITU and EBU or standard depending on the country in which you are.

n.b. Some programs, especially those mastering and audio editing are among the options the possibility to normalize the RMS values ​​following the reference standard (often EBU R-128, with reference to – 23 LUFS standard for broadcast, although in reality the algorithm calculates the rms level of the music program analyzed and increases or reduces the rms value bringing it closer to – 23 LUFS put as a limit of harmonic distortion and alteration of the frequency response that when detected limit the phenomenon of gain or attenuation, so that it is easily detectable for high or low rms values, the non-effective normalization to – 23 LUFS. Generally, without causing distortions and alterations of the frequency response, these algorithms are able to work well in a range of +/- 5 dB, so avoid using the normalization factor for rms values of the music program far outside this range, if very high cater to mitigate undergoing mixing or mastering procedures the final RMS level, if too low to provide the same manner to raise the final RMS value, doing so will surely get a higher quality).

Below are some graphic representations to see how loudness has changed over time (analysis and statistical charts were performed by Ortner in 2012, on a sample of + 10,000 songs from 1951 to 2011 focusing mainly on CD recordings and mastering) (Figure 17).

Fig. 17 2016-10-26_15-50-28.jpg

As seen in figure 17 from ’79 until ’03 the average loudness level has gone to increase, demonstrating the battle to reach ever higher loudness before defined. From ’03 to ’11 the level of loudness went on to equalize, this somewhat for the loss of quality obtainable by raising further the level and a bit to the recording medium (in this case the CD) that is not capable of recording and play higher levels of loudness correctly.

From 1997 to ’05 it is also noted that the peak level is higher than a few dB at 0 dBFS, from + 0.5 dBTP to + 5.4 dBTP causes serious distortion problems, especially in digital but at the same time to make it a More loud sound was overlooked.

Fig. 18 2016-10-26_16-02-28

In Figure 18 we see the Loudness Range section always in consideration of the analyzed audio samples. It is seen as the difference year after year is not so high if not in the periods from ’85 -93 and ’03 -05 to signify that have been most commonly used compression processors and / or a greater spread of electronic sounds (then compressed). For this you can also understand how the battle of loudness is not from a musical point of view dynamic extension, always well respected.

In Figure 19 an expanded representation of the Loudness Range and the mean level based on the analysis period.

Fig. 19 2016-10-26_18-56-52.jpg

Where, however, much work has been done to reach remarkable levels, it is from the point of view of the Loudness Program and the headroom in peaks limitation and elevation of the overall level, leading to dynamic crushing and perceiving always a flatter sound (Figure 20).

Fig. 20 2016-10-26_16-14-43

In figure 21 an expanded representation on the maximum peak level (Maximum True Peak) based on the period of analysis, it is clearly noted from ’91 onwards there is a tendency to exceed the 0 dBFS.

Fig. 21 2016-10-26_18-59-56.jpg

To informative purpose we also see the course of the average value in the time for Momentary Loudness (fig. 22) and Short Term Loudness (fig. 23).

Fig. 22 momentary.jpg

Fig. 23 short.jpg

Below is a comparison table including loudness of each factor shown for each period of analysis, including Integrated Level (Fig. 24).

Fig. 24 2016-10-26_19-05-59.jpg

Below instead a table representing the use of processors to achieve the desired loudness level (Figure 25):

Fig. 25 2016-10-26_17-28-25.jpg

It is known as initially until the early ’90s, compressors and limiters were mostly used to balance the level of loudness of the master and limit peaks in recording and mixing stages of the analog components (in addition to the use of VU Meter as reference scales which alone restrict the opportunities for loudness management). Then with the advent of digital audio and plugin it began the era of the abuse of software, digital hardware and loudness war (through increasingly aggressive compression and use of limiter to avoid distortion phenomena beyond the 0 dBFS also though in some cases as shown has arrived even further, also passing by an equalization abuse especially in high frequencies to optimize the listening even in areas in which the ear fatigue in sensitivity “high”), until today in we arrived at the limits beyond which you cannot compress if getting poor results.

In figure 26 we have instead a graphical representation targeted to highlight the difference in time between the soft passages and the program loudness of the various music tracks over time.

Fig. 26 CAM00707.jpg

Note how in the analog era a bit limited in equipment and low dynamics processors and speakers used the difference between the value of program loudness and soft passages is narrower than the relationship between them after the advent of the digital age (especially from ’94 onwards). This is because during the era of analogue background noise of all the equipment used was higher than today, and then you tried to close as possible to the soft, light passes to the average perceived value (and yet were obtained more dynamic than listening today) during the digital age instead the phenomenon of compression and high dynamic usable allowed to work more on the difference between the soft and loud passages making it more natural sounds (pity though that tends to crush all dynamic towards highest possible peak value).

Some think that the loudness war has begun with the advent and spread of audio in automobiles and portable devices, starting from the premise to realize loud songs to compensate for background noise caused by cars that both internal and external environmental noise and products for use by consumers and not the Hi-Fi.

Below I propose charts of the performance of a song of the ’80s (Michael Jackson’s Beat It) and its evolution over time, to represent how discography points to remastering old songs and albums according to the average values of the period they make Reference to sell again new collections of the same artist (figures 27-28-29):

Fig. 27 (trends LUFS parameters)

2016-10-27_14-03-17

In figure 27 can be seen clearly as any parameter has risen to a level in the course of time to indicate the increasing compression in order to adapt the level of loudness to the market. The Crest factor indicates the headroom between the maximum loudness value and the Max TP. HLSD (High Level Sample Density) indicates the level of sample present in the passage between – 1 dBFS and 0dBFS, it is clear how the evolution has led to the use of increasingly limiters and normalizers to raise the level of the signal dynamics.

Fig. 28 (evolution of gain in frequency response using as the equalization reference in 1982) 2016-10-27_14-03-43.jpg

It can be seen especially in the low frequency has been given a considerable increase in dB (over 10 dB of gain). Even in high frequency, a significant increase in dB (over 7 dB) has been shown to indicate the best response frequency response (hence a wider frequency band usable) of the recording and playback devices used since 1990 with the ‘ Advent of the CD.

Fig. 29 (evolution of the frequency response)

2016-10-27_14-29-49.jpg

In Figure 29 we have the frequency response, even in this case it is clear that there is a considerable increase in the average value (pink and red) and a peak value normalization (probably through standards bodies and / or limiters) (blue).

A final figure to fully understand the development of the loudness in history is in figure 30 in which are represented all the peak levels normalized to 1029 Hz over time:

Fig. 30 2016-10-27_16-43-45.jpg

In figure 30 it is clear that during the first 50 years of the tools used do not permit the extension in dynamic and frequency response at high frequencies above 5000 hz (considering a – 10 dB tolerance) and low frequency below 80 Hz (considering a – 10 dB tolerance). It is also seen as especially in recent years there is a tendency to exceed the 0 dBFS accepting a little distortion but ensuring greater loudness.

 

Considerations on standardization

As you might guess from all analyzed in this article, each media and playback device has basic automatically or manually some type of normalization to prevent and limit this Loudness War, especially in the broadcast. To date the only devices capable of accommodating more freedom in loudness management are CDs, DVDs and Blu-ray player or some software such as JRiver, which do not have if not for manual selection processes of normalization.

For this reason many manufacturers choose to process and mix with different loudness audio files for these media.

For this reason it is believed that optical supports are hard to die!

Even live music including sequences (if not normalized) guarantee full freedom in managing dynamic, limited only by the legal regulations on the maximum noise exposure.

Today the standardization can be a problem from the point of view of the creation of a mix and mastering because as you can easily guess from everything explained it tends to change the natural character of the original mix and mastering. So as not to lose hours of work and commitment it helps to know the type of destination that will have the single or the album produced in order to create a master following the effects of processing and normalization of the playback system, which is broadcast, portable player or stream in order to understand the effects and create mix-master mediated or specific for that type of target (this can also be seen as a considerable increase took place mostly work as a mastering operation), so as to avoid peaks and distortions, dynamic crushing, reducing the image stereo or surround, depletion effects and frequency response.

 

Conclusions

In conclusion, a question can be asked:

It’s better to listen to a loud signal but distorted or a more dynamic signal?

The correct answer would be a more dynamic signal and increase or decrease the gain via the amplification potentiometer of your playback device according to when “loud” wants to be perceived.

LUFS Meter is therefore to now an indispensable tool for any kind of application from mixing to mastering, that both studio or live, thanks to it is possible to achieve dynamic levels, depth, spatiality of a mix cannot be reached with the normal analog or digital meter . So, I therefore advise those who have not already done so to replace their meter with a LUFS Meter (we will see in other arguments how to use LUFS Meter to mix and do Mastering).

For more information on the evolution of loudness where there are a lot of charts here proposed in a more detailed explanation:

Loudness Evolution

There are other ITU and EBU reference standards as well as those just seen, mostly linked to television broadcasts:

EBU.3342

EBU.3343

ITU R BS 1770 “Algorithms to measure audio program loudness and true peak audio level” (09/2007)

ITU R BS 1771 “Requirements for loudness and true peak indicating meters” (07/2006)

ITU R BS 1726 “Signal level of digital audio Accompanying television in international
program exchange “(04/2005)

ITU R BS 1864 “Operational practices for loudness in the international exchange of digital television programs” (03/2010)

ITU R BR 1352 “File format for the exchange of audio program materials with metadata on information technology media” (12/2007)

 

More on Decibel and Meter:

Decibel and Meter – I ( Decibel and Standard Types )

Decibel and Meter – II ( Analog Meters )

Decibel and Meter – III ( Digital Meters and Software )

Decibel and Meter – V ( Meters in Audio Equipment )

Decibel and Meter – VI ( Loudness Manager, Loudness Engineer )

Advertisements

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s