In one of the first big lab sessions, we investigated what constituted broadcast quality signals as well as how to identify and measure them. This was part of our Broadcast Standards module, which was very technical at times yet was interesting all the same. Although this is the only piece I’m posting here, there was also an exam which was possibly the most difficult thing I’d ever done to date!
What follows is the document I submitted in its complete format, titled, “Identifying and Measuring Parameters of Broadcast Quality Signals”. As it’s got piles of useful information in the document, despite being a bit long and dry if you’re not into the technical details. This was my first time measuring broadcast quality signals, showing the impressive power of an oscilloscope. However, if you are curious about broadcast quality signals, then read on!
The objective of this report is to show understanding of the industry standards used in broadcasting audio and video signals by identifying and measuring parameters of broadcast quality signals.
This will be achieved by using standard lab equipment (detailed in the relevant sections below) made available at LJMU studios and labs and comparing to theoretical standard values which can be calculated as well as being made publicly available to allow for accurate calibration of equipment to a professional standard.
Finally, with both measured and theoretical values, there will be discussion on the quality and accuracy of the equipment available, as well as a general conclusion to show understanding on what has been undertaken in the measurement sessions.
Sine Wave Signal Generator
In this experiment, a signal generator was used to feed a 1KHz test tone into a pre-configured typical broadcast mixing desk. From the mixing desk, a representative resistive load of 600ohms was used for the test and calibration of the desk, while also having a RMS Voltmeter and Oscilloscope connected to the output to allow for the desk output to be measured and observed.
The following tables and charts display both the theoretical (th) and measured (act) values for both Power in mW and Voltage in V, as well as the error between these values.
To calculate the theoretical values for Power and Voltage, the standard log formulae are rearranged for each as follows:
20log10 (Vout / Vref)
Vout = Vref x 10(dB/20)
10log10 (Pout / Pref)
Pout = Pref x 10(dB/10)
|dB Value||P(act) – mW||P(th) – mW||Error|
|dB Value||V(act) – V||V(th) – V||Error|
Discussion of Results
Although there are significant variations in some of the actual measurements versus their corresponding theoretical values, that there are equivalent error values for Power and Voltage (remembering that the value for error for Voltage should be approximately half that of the Power value due to the different log formulae) shows at least that the equipment was performing consistently within the lab. For example, had the Voltage errors come out significantly different, then it would be possible that either the V(act) or P(act) values had been measured incorrectly and would need to be reviewed.
We can also see this when reviewing the measured values and remembering the general rules of thumb when it comes to Power and Voltage.
Firstly, with regards to Power, going from -3dB to 0dB does show an approximate doubling in the value, with the only outlier being the 10% larger value at 0dB over theoretical not allowing a more exact doubling in the value. This shows far clearer when comparing -10dB to 0dB for Power as this is a more exact 10th of the value.
For voltage, going from -6dB to 0dB also shows that the rule of thumb here applies also where the value from -6dB to 0dB can be doubled and that gives an approximate estimate of the value which was then measured. Again, going from -20dB to 0dB shows a 10th of the value.
However, even though the equipment was setup carefully by a technician, that there is a 10% error in the actual measurement for Power at 0dB over the theoretical does show how easy it can be for professional setups to come outside the range of compliance with EBU R128, as explained in EBU Tech Doc 3343 (when measured with devices as stated in EBU Tech Doc 3341). The smaller errors in theoretical versus actual measured values could also be down to a combination of the quality of the equipment and competency of the user.
Test Pattern Generator
In this experiment, a test pattern generator was used to display 100,0;100,0 bars, then multiple oscilloscopes were used to display the Y’ Pb Pr signals from the generator. These were then used to calculate and compare between theoretical and measured values.
The following tables and charts display the measured (act) values for Y’PbPr as well as the errors from theoretical values.
The table below is showing the theoretical values for luminance and colour signals in mV which were used as a reference against the measured values. These theoretical values are calculated according to section 2.5.1 of ITU-R BT.601-7 for the construction of luminance and colour-difference signals which is weighted as follows:
Y’ = 0.299R’ + 0.587G’ + 0.114B’
From this weighting, you can then calculate R’-Y’ and B’-Y’ as well as values for Pb and Pr (Pb= 0.564(B’-Y’), Pr= 0.713(R’-Y’))
Graph for sync pulse, burst and Y’ (act) signal
Y axis – mV ; X axis – Time µS (each square is 1 µS)
Graph for Pr (act) signal
Y axis – mV ; X axis – Time µS
Graph for Pb (act) signal
Y axis – mV ; X axis – Time µS
Discussion of Results
Considering the number of values being displayed in the video signal, the error amounts between the theoretical and actual measurements were overall small and a significant amount could be down to equipment and user error in reading data. This is shown as although the actual Y’Pb’Pr’ are generally below the theoretical, they are consistent and when shown on a graph still “look” right.
This is one of the advantages of the PAL system in general as although the signal itself can be weaker, so long as the broadcast has been calibrated successfully, the final viewed footage should still be acceptable when viewed as a whole image (not having flesh tones etc. being displayed incorrectly, for example).
Based off the percentage errors shown, the system is broadcasting within 5% which considering the equipment being used is acceptable. What is more, the values are generally being displayed as expected, showing that although some of the values are not identical to the theoretical, the system as a whole is performing acceptably.
In order to produce both audio and video signals accurately, it is clearly important to both calibrate equipment as well as monitor the equipment to ensure it is still performing to broadcast standards. The lab sessions showed how quickly this can be done with standard studio hardware, giving all students a far greater understanding on the theory behind the signals.
For audio, papers like Hanna and Easley (2009) show that despite correctly calibrating equipment at point of broadcast, “large loudness changes are often perceived during program-to-commercial transitions and while changing channels.” Showing that home user experience can still easily be compromised by external environmental factors, which is why the broadcast levels are so tightly standardised.
For video, although standard definition using BT.601 was used in the lab, the same high standards are applied today with high definition standards within BT.709, as shown in section 4 of EG 36:2000 , vol., no., pp.1-7 (2000). So having a high level of understanding over why these standards are used and how they are created is vital to a modern engineer in the field.
As broadcast of audio and video evolves with advances in digital streaming services, correct monitoring and calibration of the equipment used in the studio is becoming ever more crucial in providing the most accurate experience for the end viewer.
EBU (2016) R128 Tech Doc 3343 [online] URL: https://tech.ebu.ch/docs/tech/tech3343.pdf [Accessed: 9th November 2018]
EBU (2016) R128 supplementary Tech Doc 3341 [online] URL: https://tech.ebu.ch/docs/tech/tech3341.pdf [Accessed: 9th November 2018]
ITU-R (2011) BT.601-7 [online] URL: https://www.itu.int/dms_pubrec/itu-r/rec/bt/R-REC-BT.601-7-201103-I!!PDF-E.pdf [Accessed: 12th November 2018]
Hanna, Chris; Easley, Matthew (2009) A Survey of Broadcast Television Perceived Relative Audio Levels. Sound in Real Spaces [First Published online October 1, 2009]. URL: http://www.aes.org/e-lib/browse.cfm?elib=15091 [Accessed: 14th November 2018]
EG 36:2000 - SMPTE Engineering Guideline - Transformations Between Television Component Color Signals," in EG 36:2000 , vol., no., pp.1-7, 23 March 2000 doi: 10.5594/SMPTE.EG36.2000 URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7290633&isnumber=7290632 [Accessed: 14th November 2018]