Testing car audio amplifiers

[RAW-CAt]

Hello, lovely people. I have a small YouTube channel that focuses on car audio measurements and testing. Since I joined SQ (sound quality - opposite of SPL) community, I realised that a lot of car audio is based on myths and fairy tales from "back in the day". My goal is to start fresh and test and measure as much as possible and give the community actual data rather than opinions.

In order to do that I would like to measure full range 4ch car amplifiers, DSPs and DSP/amps. I would like to do that as cheap as possible with the equipment that I already have. I would like to follow Amir however, I do realise that the reliability or my measurements won't match audio precision...

A little bit about my gear. As a power source I will use a LiFePo4 100Ah bank that is capable providing ~100A of continuous current. On top of that I have a switching power supply that can output ~35A. I have a DIY resistor load bank consisting of 4x4ohm and 4x2ohm loads up to 400W per channel. My measuring device is Focusrite Scarlet 2i2 (2nd gen). I am planing to upgrade it to Motu M4 as according to Amir's testing it outperforms 2i2 by quite a lot. I will be measuring with REW, I recently discovered all the stepped sine measurements the software can perform. Singal will be fed by the interface for now (up to 2V unbalanced), however I am planing to use a quality head unit capable of outputting up to 4V, in order to keep the gains on the amps at minimum.

Now going to the point, I need help deciding what actual measurements to make that would be applicable to car audio and would be realistic and meaningful. For example I think that measurements at 1W or 1V are not realistic, as car audio speakers in general are not sensitive and require more power for "daily listening" while driving. I think 10-15W would be a more realistic value. The audibility thresholds have to be different as well. I'm thinking of presenting the results in percentages as much as possible, as they are a bit easier to understand than dB.

Any advice would be appreciated. I have a few videos made about this venture:

 

[alex-z]

In my experience, car audio speakers have similar sensitivity as typical bookshelf speakers, 80-85dB @ 2.83V @ 1 metre. The 5 watts Into 4 Ohms (4.472 volts) that Amir tests with is relevant for daily use.

As for dB vs percentage, I would argue dB simply because it is a more intuitive scale for visually determining audibility thresholds. In a perfect world the software would have an option to apply perceptual masking for you, but that hasn't gone mainstream in the hi-fi world. Telephone networks use standards like ITU-T P.863, but that is modelled specifically for speech, not music.

 

[fpitas]

I wish you nothing but luck. But, I have to think a lot of people would rather not know the truth. It's that kind of market.

 

[Online_olympian]

Awesome. I have hundreds of amps, decks, crossovers, dsps, components, subs. If you want access to my Google drive to look through it. Let me know. I am willing to loan some stuff for testing. I have been collecting old school gear for years but I do have quite the assortment of more modern gear as well.

 

[DVDdoug]

Now going to the point, I need help deciding what actual measurements to make that would be applicable to car audio and would be realistic and meaningful. For example I think that measurements at 1W or 1V are not realistic, as car audio speakers in general are not sensitive and require more power for "daily listening" while driving.

Amir measures SINAD at 5W. At low power the limitation is usually noise, and I believe the idea is to use a standard (lowish) level because if you measure relative to the amp's maximum output, a higher power amplifier will give you a better signal-to-noise ratio even if the background noise is the same and equally audible.

Then he plots distortion vs. power.

The real noise usually comes from the car's 12V power supply or from ground loops so the bench/lab measurements may not be that useful.

Power measurements could be VERY helpful. I'm pretty sure most manufacturers fudge, and some by a LOT!