Hi All.

This is sort of a summary/continuation of several questions I've asked in this room in the past.  Thanks very much to all the AVIXA folks who have commented on these subjects in the past and brought me far enough to post all this.  I hope it encourages healthy discussion. Hopefully I can find the clarity I've been looking for... and if my questions are apt, maybe the results will find their way into a future edition or addendum of the exam guide...

When specifying amplifiers I am constantly under pressure to reduce the wattage for cost reasons.  I have struggled to understand several things about the 'Audio Principles of Design' chapter in the Exam guide as they apply to specifying amplifiers with a minimum of over-engineering.

At the same time, I have been part of system configuration in the field and found that my calculated taps are (at times) only barely enough to provide the 68dB-72dB my calculations accounted for.  I would like to avoid under-engineering even more so than over-engineering.

There are four points of confusion for me:

1) Choosing appropriate headroom in the EPR calculation (p155)

2) Applying the RMS vs peak values yielded by the EPR calculation to amplifier selection (pp155-156)

3) Lack of clarity on the reasons we increase the summed tap wattages by 150% before selecting a 70V amplifier as shown in step #7 (p158, top).

4) Best practices for ensuring the specified system has an controlled amount of 'overengineering' in the tap values to allow for volume controls to sit at approximately the 'two-o'clock' position when delivering the intended SPL

Choosing appropriate headroom in the EPR calculation (p155)

The exam guide suggests 10dB of headroom for speech and 20dB for music.  This is simple enough in terms of speech, but 'music' encompasses everything from the highly pre-compressed output of an iPhone with the 'sound normalizer' on;  to live orchestral performances in amplified concert halls.  

I have shared my designs based on this calculation to several amplifier vendors/manufacturers and been told that 20dB is a frankly ridiculously large value for real-world background music.  Still, as I mentioned above, I've sometimes found it to be 'only just enough.'

Does anyone else have additional points of reference for headroom required?  Say for an mp3 player output playing rock and pop versus a restaurant lounge stage input where the board is set up for moderate overall compression on the mains output? And is 20dB sufficient for a concert hall, or does full-range sound reinforcement of an acoustic live performance with timpani and brass require 30dB?  40dB?

The only source of data I have been able to find on my own is via software created by participants of the online 'fight against loudness wars.'  I can use programs like DR14.meter to assess batches of the prepared mp3 files my employer uses as content and discern the differences between RMS and peak levels across our catalogue.  (I have not yet done this.)  From what I've read online, these values are typically between 8dB and 14dB for mastered recordings.  I would further guess that pop and rock recordings trend toward the lower end of that range.  

Will software like DR14.meter give me the information I am looking for?  (Have I correctly matched the software to the use-case and in compatible units?)

Am I correct in believing that the findings above suggest a headroom value of 15dB is appropriate for most mastered pop and rock music?

Am I also correct in believing that a compressed 2-mix off a restaurant stage would require something closer to that 20dB value, and that 'proper' sound reinforcement in the pro-audio world requires even greater headroom?

Any advice is appreciated.

Applying the RMS vs peak values yielded by the EPR calculation to 70V tap and amplifier selection (pp155-156)

As shown on the screen capture from Biamp's website in figure 6-12 on page 156, the EPR calculation can be performed twice... once with no headroom to determine RMS and once WITH headroom to determine peak.  I assume that tap values must then be calculated using the PEAK figure, as the peak is included in the default ERP calculation we are given, and no language in the chapter suggests omitting it.  (Nor does any language suggest how we might make use of the peak wattage value once it has been so calculated, if not for determining the speaker taps.)

However, this implies that after the tap wattages found by PEAK EPR calculations are summed, an amp with a greater-or-equal PEAK wattage should be specified.  Unfortunately, 70V amp manufacturers typically only publish RMS wattages on their spec sheets.  What's more, it is implied by the manufacturers and vendors that the published RMS rating SHOULD be used when specifying an amp based on summed taps.

I do understand that 70V amps aren't so much meant to provide the wattage they advertise, as they are meant to perform at a 70V peak output down to the IMPEDANCE suggested by the tap wattages in the line... (70.7^2 / wattage and all that...) but why would amp vendors confuse this issue even further by citing 'RMS' where it doesn't belong?  

Can anyone explain this discrepancy? What is the best practice here?

 Lack of clarity on the reasons we increase the summed tap wattages by 150% before selecting a 70V amplifier as shown in step #7 (p158, top).

On page 158, step #7 tells us to "increase the total tap setting power by a factor of 1.5."  No reason is given. 

The headroom of the audio signal has already been accounted for in the previous step, so the remaining factors I can imagine are:

1) Accounting for any additional impedance from 70V attenuators, cabling, and perhaps speaker transformer loss itself... and thus ensuring that the amp does not fail due to these 'invisible' impedances from 'outside the calculations.'

2) Allowing for some 'fudge factor' to ensure the volume knobs aren't at 100% when the specified SPL is reached.

Folks on this forum have suggested to me that #1 is at play here.  And this makes sense... if it were #2, it wouldn't be listed in the 'Specifying a Power Amplifier for Distributed Audio' section... it would apply to directly-connected systems as well.  

Can I get confirmation that this factor of 150% is meant to absorb impedance outside of the EPR calculation brought into play by 70V attenuators, line loss, and/or transformers?  

Further to this, I know impedance of a cable is typically very low, and I know that attenuators are lossy.  Does this mean I can forego or reduce the 150% calculation if my 70V speaker network has no attenuators?  (i.e. volume is controlled using a DSP channel or remote amp channel control?)  Or are the speaker transformers themselves enough of an added load to require the 150%?

Best practices for ensuring the specified system has the desired amount of 'overengineering' in the tap values to allow for volume controls to sit at approximately the 'two-o'clock' position when delivering the intended SPL

Finally, and as mentioned briefly in the previous section, we come to the last issue:  If I want the system to perform at a nominal level of 72dB-SPL-A at the listener, how can I adjust the calculations to place the 72dB playback level at approximately 2 o'clock on the volume controls when this level is reached?  (I mean, obviously I can overengineer the taps and back off the amp trim... or limit the DSP volume range... but neither of these save $$ on my clients' budgets.)

My guess would be that I should perform my EPR calculations with the MAXIMUM possible required SPL instead of the OPTIMAL/STANDARD SPL... but is there any guideline as to how many decibels should be added to put that dial around 2'oclock at a given dB-SPL?

Thanks to any and all who have read this far.  I hope to meet you in the comments.