Decibel and Meter – III

Digital Meter

The digital meter (fig. 1) are the meter that identifies the output values and the signal input to digital scale used in digital audio equipment (after the A/D converter and before the D/A converter when it comes to analog inputs and outputs and software.

Fig. 1 2016-01-06_19-00-15

At the hardware level they are such and such to the true peak meter that is LED display (fig. 2) only managed by integrated circuits whose voltage domain is that digital (square) and the level is that digital standard as the audio signal once it converted from analog to digital is raised to the voltage levels for the digital standard.

Fig. 2 2016-01-06_00-52-41

As defined in digital the maximum value before the distortion is 0 dBFS or clips.

Generally they are led or colored graphics green bars up to the reference level for the analog standard (-24 dBFS for SMPTE and – 18 dBFS standard for the EBU standard), so as to allow the correct interfacing between analog and digital devices, and orange up to -1 dBFS, while is red the 0 dBFS to indicate distortion values.

At the software level, especially in DAW (Digital Audio Workstation) and especially for the Master controls as for Figure 1, the digital meter are often enriched with additional graphics and control capabilities, such as the Peak Hold for both the peak for the RMS value of the RMS signal value and peak on separate bars to ensure proper visualization of Headroom (range that goes from the rms value of the peak) in order to manage the loudness of the audio program, complete numerical data accurately to the point, pan, the signal phase and more. In addition to sometimes find different colors depending on the signal level, and not just green, yellow, red, but shades to cooler colors for low value signals to hottest color upper to red at 0 dBFS.

The Digital Meter is in all and for all the faster and more accurate meter in providing data on the level of circulating signal. As only flaw was the fact that they depend on a CPU processor and also GPU whether in software, that depending on its quality and its performance can make the system unstable and create errors in the processing and then visualizations, the problem cannot be found instead in the True Peak Meter.

Typologies of Digital Meter

There are 4 variants of the digital meter:

Cheap Digital Meter

The poorer are those cheaper to offer a low-cost product offering with strong meter inaccuracies because only consider parts of the circulating digital signal, so as to be able to use processors and simple and inexpensive circuits precisely. Very often seeing the inaccurate of measure present value scales with 5 dB steps or more.

SPPM (Sample Peak Program Meter) is the first digital meter as well as the most commonly used in consumer level, it is a medium-sized digital meter as, set a sensitivity (pre-set by the manufacturer) only shows the peaks falling within the samples range established (an approximate range better than cheaper but less precise than the professional ones). This is because it would be not very relevant for the purposes of an analysis, from the consumer show the peak level signal of each single sample, for this it is average by setting a tolerance range and give a generic visual of reference (more than to visualize and give information on limit value beyond to obtain distortion value that should cause damage to the equipment).

n.b. We’ll see instead at the professional level the sample analysis as its fundamental to manage quality audio signal.

Example: to realize a digital signal as we shall see in future arguments, the analog sinusoid (fig. 3 red line) is taken from the A/D converter and brought to digital level according to rules for sampling and quantizing (fig. 3 black lines which follow the red sinusoid), sampling is shown on the horizontal line and each line is a sample, so the higher the sampling and many more samples will be taken the analog sine wave input so as to be the digital domain closer to that analog. The quantization is on the vertical plane and represents the amplitude of the wave (its voltage value), the higher will be the quantization the more samples taken will be given a width, so as to compose the shape of a sinusoid increasingly similar to analog input.

Fig. 3 800px-pcm.svg.png

In SPPM, for example, by setting a tolerance range of 10 (referring to the example in Figure 3), the meter will only display the voltage levels of the samples present up to 10, which will then be mediated to not send in Resonance and distortion of any CPU and/or graphic processor that has the task of processing this data.

As we understand it is not a precise meter as actually considering taking each single sample the overall signal level could easily be higher, generally from 3 dB to 4 dB.

In the professional field, instead, they use the Over Sampling Peak Program Meter (OSPPM).

The OSPPM have lower tolerances and more precision since before the analysis the signal is over-sampled, generally x4 or x8 times to obtain more accurate values having many more samples to be analyzed. From this one can easily guess how OSPPM require processors with higher performance compared to OSPPM and therefore cost even more.

N.B. Oversampling are today exclusively performed in a digital domain, as at a convertor level an analog signal can be sampled up to a maximum of 384 kHz.

Both SPPM that OSPPM generally have a scale with step of 1 dB in the proximity of 0 dBFS.

Some Professional Digital Meters distinguish the 0 dBFS from the Oversdrive (distortion values above 0 dBFS), while others already indicate the value of distortion since the meter indicates 0 dBFS.

The difference lies in the fact that a digital signal at 0 dBFS is not said to be distorted but it is only at its maximum value if the system detects at least 3 samples always at 0 dBFS then thinks this is a distortion value and then signals with The red led of the Overdrive

An alternative to OSPPM is the Intersample Meter (ISM), similar as functioning on the oversampling principle but with the difference of not creating mediation between samples analyzed through a time window. These meters are the most accurate, they analyze and provide sample information per sample, with the ability to visualize any distortion and use processors such as limiters, compressors, and sample-resolution gateways.


Plugin Meter

Finally, some Plugin emulators of analog processors have meter constructed with algorithms that tend to emulate the responses and graphics of the analog meter (fig. 4).

Fig. 4 2016-01-25_18-51-42.jpg

Every meter in addition to these basic parameters for compliance with regulations, can be equipped with other components and various employees applications by the manufacturer, so it is good once you know to use a professional meter that complies with the regulations in force, to study all its functionality to order to use it to its fullest potential.

At the software level (either Plugin or simple standalone software) it is possible to identify different scale values in the structure of a digital meter, which is not yet widely distributed in hardware environments on mixers and audio equipment, except in hardware meters. For hardware equipment, the main function is to identify the signal level such as to send the amount of audio signal needed from one device to the other by avoiding distortion phenomena, while in the software environment the meter as we will see is used a lot to identify also the values RMS media and manage its Loudness (audio level perception, volume you perceive) in order to create mixes and qualitative mastering.
In figure 5 all the possible levels of signal that can be identified in a digital meter on software.

Fig. 5 2016-09-07_14-19-54.jpg

In this case, the meter extends its scale of values from – ∞ to 0 dBFS on the vertical plane, but it can also be found on the horizontal plane. The maximum point reached by the audio signal is called Peak Level, you can then identify how we will see the RMS value divided by Max RMS for its maximum point and Min RMS for its minimum point, sometimes called Program Loudness or RMS. The range goes from Max RMS to Peak Level is called Headroom (not perceptible to the ear but to be careful as it is leading to distortion and breakage of the hardware components, not to hearing impairment). The range goes from Min RMS to Max RMS is called the RMS Range or Loudness Range (this is what makes the volume level and dynamic sound to our ears feel).

Under the RMS value we have the Foreground and Minimal Audience area (little audible, especially at high RMS Max values and compressed audio (so with a small Headroom and RMS Range), creates a kind of rug, helps to spatiality of the image and is the area where you generally work audio effects.
Below we still have the Background and Noise in which there is background noise and lower-value audio parts including effects.


RMS or loudness in digital

As we know, the digital signal is in the form of square wave so it does not have average values and therefore RMS, but then how can digital meters display the RMS value of the circulating signal?

In the analog domain we have said that the RMS value of the sinusoid is calculated as the 0.770 of the peak value (without considering the use of dynamic processors) and is shown as the sum of RMS values mediated at a certain time depending on the resolution and the capacity of the meter used, however, in digital domain, it is still necessary to find a way to represent the dynamics and loudness of the audio signal otherwise it would be a potential analyzer limited to peaks reading for the interfacing of digital devices and useless to use tools such as processors of dynamics and equalizers.

To do this the RMS value in the digital domain is nothing more than the calculation of 0,770 on the peak of the sinusoid recreated after the process of sampling and quantization (always without considering the use of dynamic processors).

In digital meter at software level, the RMS value follows the numbering and coloring used for the peak meters, which are only reported differently.
As can be seen in Figure 1, the red of the central stereo meter that represents the RMS value begins well before 0 dB, these coloring parameters are generally adjustable and variable by the user (for the correct calibration it takes knowledge and experience), they are indicate the maximum value that can be reached beyond which an overcompressed signal is obtained and no longer qualitatively suited to the standard generally used for the type of program you are considering.

The associations ITU (International Telecommunications Union) which defines and sets standards for telecommunications and radio waves, (BS.1770 rev. 1, rev 2, rev 3), EBU (European Broadcasting Union) (EBU R128), ATSC (Advanced Television Systems Committee for the development of standards for digital television), (A/85), the latter for American broadcasting and also OP-59 (Australia) and TR-B32 (Japan), have defined over time the standard headroom levels (recommendations) that should have the musical signals according to the type of program you are considering to be perceived by our ear in a qualitative and clean manner (considering that these programs are played in environments acoustically fit), as well as the method of calculating and the structure of the digital meters for their calculation. The ITU and EBU standards are very similar to each other and are considered internationally. Some countries such as the United States, Japan, Australia and others still adopt their own standard of ownership, however, following the ITU guidelines.

In Figure 6, some examples on the recommendations, which also include the average value of background noise and music background, the average mean value of the foreground and minimum audience, and the average value of the headroom (here is presented together with the range rms), Values that should have the genre and playback music program to be considered a quality sound and delineating standard reference levels.

fig. 6 loudnessnorm.jpg

n.b. It speaks of recommendations and not of real normative standard.

Looking at Figure 6 in the Peak Level Normalization side you can see which Headroom must have the music program based on its application, for example the Min RMS or Program Loudness for the cinema should not be less than 24 dB, for Jazz music should not be lower To about 15 dB, for Rock and Pop 6 dB.

On the Loudness Normalization side there is the normalized representation of the standard ITU BS.1770 and EBU R128 which instead represent a true law for broadcasting, the average value of the Loudness Program – 24 dB, to 2016 – 23 dB reference but variable depending on the type of transmission platform of some dB as we will see in part IV of this set of articles.

For a professional sound and recommendation of digital meter on software (generally can be done can only by professional ones) are generally pre-set by the manufacturer but are then calibrated according to the audio program to be analyzed in their RMS value by setting as red dimming the dBs that exceed the threshold we set up according to the music program we’re working with.

For example, if we are mixing and processing or even mastering a pop music to be downloaded as a master, for example, on an audio CD or digital download file, we will follow the recommendation in figure 6 on the left side where we have a headroom where it is also present is the range rms with a program loudness value of – 6 dBFS, to facilitate the analysis it may be useful to set the red-color of the RMS meter (such as the central one in figure 7) to – 6 dBFS if the RMS signal exceeds that threshold, the meter will indicate it by coloring the graphic bar that exceeds that red threshold (figure 7 central bars) so you must immediately know that the limit level as defined in the recommendations is reached.

Fig. 7 2016-01-07_21-59-22

n.b. It’s always said of average value over time, a calculation example for a music track is to calculate the average RMS value of the whole track. Some mastering programs let you do this so you can know if the song complies with the recommendation (also because it would be impossible to keep the same level of headroom for the duration of the song, only the big distorting phenomena would be created).

This would already be enough but if one wants to be even more precise in creating mix and mastering quality can continue defining color ranges for the Foreground and Background area, always following the recommendations.

Generally in the digital meters it’s possible to also set colorations for the peak values, headroom (yellow in the outer bars in figure 7), and rms range values (ocher color that is seen in the inner bar in figure 7 between the – 12 db and – 6 dB) and lower levels.

For the bars representing the signal peaks (the external ones in the meter in figure 7) it is useful to put the red – 3 dBFS red color for the previously seen reasons of preventing during the conversion and audio coding the rise of the peak level Distortion carrier, and as yellow range ranges from program point to loudness at peak peak.

n.b. The point of loudness range limit is an average value, it is allowed to have a + 1 / + 2 dBFS more for the Max RMS.

As we will see when we talk about LUFS meter, if you work for broadcast audio programs (tv, radio, podcasts) you will need to follow the table in figure 6 right side, as the standard – 23 dBFS (by 2016) is a norm law.

For a proper calibration of the audio program it is important to analyze only the section of the audio track, without considering the input and output pauses, since the meter measures every single input sample, so it would give values lower than the actual ones.

In Figure 8 an example of selection for preparation of an audio track analysis.

Fig. 8 2016-09-01_17-42-14.jpg

In Figure 9 an example of calculating the average value.

Fig. 9 2016-09-01_17-45-05

n.b. In this case, the software also provides the maximum RMS value and the minimum reached, as well as the average value of the cursor, all for each channel of the audio track.

The song being considered is Pop music, following table 4 we have the Hyper Pop that falls within the 6 dB of recommended headroom, since Pop is lighter and more dynamic than the Hyper Pop are allowed 2 to 3 dB of more headroom .

There are, as mentioned above, and we will see later (LUFS Meter), which automatically analyzes all of these things, and in addition they do not consider the breaks inside the track, while the meters just seen are for this reason less accurate.

The method of construction and operation of the algorithm of these meter calculation is defined in AES17-2015 regulations.

AES (Audio Engineer Society) is a company designed to publish and revise standards for the Audio and Media industry.

There are also variants on the algorithms proposed by the ITU and EBU associations to visualize the RMS signal in digital domain so as to always have the reference of the maximum reachable RMS value, the most used ones (such as variants) are those proposed and realized by the Master Engineer Bob Katz who are :

K-System 12, System K-14, K-20 System (fig. 10)

Fig. 10 2016-01-07_21-02-35.png

Figure 10 shows the 3 variants in version Peak Level Normalization and defines the Headroom that must have its music program.

In some cases these meters can also be useful not only to control the signal status on the basis of the recommendations given on which a technician may point to those proposed by the ITU and EBU standards, but also because they have parameters already pre-set by the manufacturer, So just select one of the 3 meters available, and as you see for ITU and EBU standards, it is necessary to set the meter parameters as needed.

For comparison the K-System have a more generous algorithm leaving more room to the dynamics especially in contexts of music concert.

This makes us understand how these sophisticated meters at software level and unfortunately still poorly-used at hardware level in mixers and digital processors, allow you to mix and process audio signals up to loudness limits that are of the highest quality for the type of audio program you are considering.
M-S Meter

In a digital environment, limited to some software and plugins that can perform MS (Mid-Side) processing operations, (referring to future articles and other online discoverable concepts for the processing and processing of an MS signal), rare cases in a hardware environment where processing is always displayed using a peak meter or digital meter like any other audio signal, there is a chance to find MS meter as in figure 11 (software), fig. 12 (hardware).

Fig. 11 2016-10-31_15-48-43.jpg Fig. 12 bmc-2-front.png

Mid-Side meters, which usually always accompany the peak or digital meter (Left and Right), in analyzing an audio signal are better suited to adjust and constantly monitor the independent signal level of Mid processing and Side, so you can increase or decrease more precisely one of the two components and understand immediately if one of them goes into the clip, which is not possible either by analyzing a Mid-Side signal with a peak or digital meter L-R as it can you highlight the peak but it does not tell you if the Mid or Side is distorted, but at times you reach the level of distortion but the peak or digital meter does not signal any clips (this can happen for example if the level of the Mid or Side is distorted but the other Mid or Side signal is very low so as to have a sum that does not distort the signal level L – R).


Considerations on mixing and mastering through meter

As can be seen from figure 6, mixing and mastering via a digital meter allows you to obtain very different dynamic ranges, both between musical genres and between increasing and decreasing the same song, because the voltage and dynamic range allowed by digital equipment is much higher than Analog, as well as more accurately defining spatiality and depth of image thanks to other plugin meters that we will see later. For this reason, today to have a competitive and/or creative and/or natural mix is good to use digital meter.

For comparison, once they used Peak Meter and even earlier VU meters which had much smaller scales and voltages than digital ones, they could mix and master with less loudness and dynamic range. For example, at the time of the vinyl records, the average of any song did not exceed the 6 dB of dynamic range of difference (in this case also limited by the method and structure of the incision processes that we will see in other arguments).


More on Decibel and Meter:

Decibel and Meter – I ( Decibel and Standard Types )

Decibel and Meter – II ( Analog Meters )

Decibel and Meter – IV ( Normalization and LUFS Meter )

Decibel and Meter – V ( Meters in Audio Equipment )

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


Leave a Reply

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

You are commenting using your 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