In this blog post I compare intermodulation distortion measurements being taken either indoors or outdoors to compare the effects of the room. I attempt to answer the following questions:
- Is the room having an effect on my measurements?
- Can the test setup and conditions be improved for indoor measurement?
Test Method
I conducted intermodulation distortion measurements both indoors and outdoors keeping the test setup otherwise identical in each case. I set my microphone at 30cm from my horn. I used the ES-290 Biradial Horn with the DCX464 compression driver. I used a low decade ten tone test signal from 150Hz to 1kHz.
To start I compared the natural noise of each environment with no test signal. Below is a spectral comparison of each environment's noise. The indoor noise are -103dB at 1kHz while outdoors are another 10dB quieter at -113dB.
I then conducted the IMD test at various SPL levels from 90,95,100, and 105dB SPL levels. I then compared the results side by side as shown below.
Comparing the 90dB SPL test revealed a large difference in the resulting noise artifacts. Indoors revealed noise at -47dB while outdoors was all the way down at -65dB.
Increasing the test tone SPL to 95dB showed a drop in the noise floor from -47dB to -51dB concluding that the signal to noise ratio is improving as we raise the signal. However outdoors we see no change at -65dB for both 90dB and 95dB SPL test signal levels.
Increasing the SPL level from 95dB to 100dB is shown below.
Using a 100dB test signal shows noise starting at -50dB down from the test signal. Outdoors we finally see the noise at the same -50dB level.
If we increase the test signal to 105dB we see noise match again at -45dB for both indoor and outdoor.
Conclusion
- The test signal needs to be of sufficient amplitude to rise above noise artifacts related to the condition of the room.
- It can be safely assumed that these artifacts are reflection sound within the room
- Room reflection noise is not limited to only frequencies generated by the test tone itself. In other words the reflected sound in the room somehow has the ability to change frequency and end up as side bands from the fundamental tone. I find this quite interesting! Apparently air has distortion.
- Very generally SPL levels should be at least 95dB SPL at 1 meter and the mic should be placed as close as reasonable possible to loudspeaker, at least enough to capture any sources of noise equally (ie. surround, dust cap)
Further Comment
In my previous blog I looked at the spectral content of music and how it relates to the noise levels observed in my multitone tests. I showed light grey textured areas between the natural harmonics of the bass flute instrument. The spectrogram of Paul's instrument is shown below. The light grey textured areas between the harmonics seem to follow the amplitude modulations of his instrument. From a speaker measurement context I categorize the artifacts generated between the side bands as noise from the room. However in a musical recording aspect this could very well be ambient information related to the space that makes the recording beautiful. At the very least, it's certainly not random noise unrelated to the music or recording. It is the recording.
