SPL measurement gone horribly wrong?

Mandachuva · 22nd October 2009

Hello!

Just purchased a SPL meter to calibrate the level of my monitors, namely this one:
Digital Sound 8928 Level Meter

OK, I'm a cheapo, as it's not licensed, but it's supposed to give somewhat accurate readings, right? Soundcard (Echo Audiofire) at +4 level, balanced connections, M-Patch 2 passive volume control, Adam A7 monitors.

Well, I've done the Blue Sky sattelite calibration steps (http://www.abluesky.com/user/data/bs50081b.pdf). The 1kHz sine outputs 20dB FS, the 500-2500Hz filtered pink noise outputs around -16dB FS (on my soundcard). I've set up the attenuation level on my M-Patch 2 for the individual channels to read 85dB on the SPL meter (isolation headphones on - I still try to forget the headache I got when doing some acoustical test with pink noise for a few hours).

All is well, right?

Wrong! Almost every song completely blows my ears out at that level, even jazzy dynamic ones. Even by lowering the output level of my soundcard by -18dB FS, it's still a tiny bit louder than I've ever heard my Adams, be it for listening to music or mixing down songs. (and I think this is after ear-drum compression kicked in after testing it at the funkin' "calibrated" level)

What the hell could be going on? Even if I'm not used to listen/mix loudly, as I have somewhat sensitive ears (don't ever practise without ear protection on), and the loudness wars are raging, and the bloody SPL meter is off by 5-6dB, it's still way to loud...

Help?...

What the hell am I doing wrong?

Mihai Toma

Pascal Sijen · 22nd October 2009

The calibration level you are using is derived from a film standard and when compared to movie soundtracks, music tends to be much more compressed, with reduced dynamics and greater overall level. Because there is no recognized standard for monitoring level in the music industry, music mixes may seem too loud when played back on a system calibrated for an 85dBc reference level. Typically for music applications, we recommend calibrating for a lower reference level, such as 78dBc.

With regard to the 500-2500Hz filtered pink noise measuring at around -16dB FS: The bandwidth limited pink noise will read higher than -20dBFS, because pink noise has a high crest factor (12dB), which will throw off the RMS / peak reading on the meters in your DAW. In order to get an accurate reading, you would need to use an averaging technique, where multiple time samples are captured and averaged over a ten second interval to calculate the RMS level. This is why we included the 1kHz test file, which is used for setting the electrical / digital levels. Once you set the faders / output of your DAW to -20dBFS / +4dBu / odBVU using the 1kHz tone, then you should leave them in this position, when doing the acoustic calibrating, using the bandwidth limited test files.

I hope this helps to explain what you are observing.

Cheers!

Mandachuva · 23rd October 2009

Thanks for the quick reply, Ms. Sijen.

I understood the calibration technique using the sine why - dunno really why I mentioned the peak meter readings for the filtered pink noise.

I'll give it a shot tomorrow to calibrate it at a lower level, hopefully with better results... In truth, loud mixes were showing as 90-95+ dB SPL on the meter, using the previous calibration technique.

All the best,
Mihai Toma

22nd October 2009	#1
Mandachuva Gear Head Join Date: Feb 2008 Location: Timisoara, Romania Posts: 44 Thread Starter	SPL measurement gone horribly wrong? Hello! Just purchased a SPL meter to calibrate the level of my monitors, namely this one: Digital Sound 8928 Level Meter OK, I'm a cheapo, as it's not licensed, but it's supposed to give somewhat accurate readings, right? Soundcard (Echo Audiofire) at +4 level, balanced connections, M-Patch 2 passive volume control, Adam A7 monitors. Well, I've done the Blue Sky sattelite calibration steps (http://www.abluesky.com/user/data/bs50081b.pdf). The 1kHz sine outputs 20dB FS, the 500-2500Hz filtered pink noise outputs around -16dB FS (on my soundcard). I've set up the attenuation level on my M-Patch 2 for the individual channels to read 85dB on the SPL meter (isolation headphones on - I still try to forget the headache I got when doing some acoustical test with pink noise for a few hours). All is well, right? Wrong! Almost every song completely blows my ears out at that level, even jazzy dynamic ones. Even by lowering the output level of my soundcard by -18dB FS, it's still a tiny bit louder than I've ever heard my Adams, be it for listening to music or mixing down songs. (and I think this is after ear-drum compression kicked in after testing it at the funkin' "calibrated" level) What the hell could be going on? Even if I'm not used to listen/mix loudly, as I have somewhat sensitive ears (don't ever practise without ear protection on), and the loudness wars are raging, and the bloody SPL meter is off by 5-6dB, it's still way to loud... Help?... What the hell am I doing wrong? Mihai Toma

22nd October 2009	#2
Pascal Sijen Gear maniac Join Date: May 2003 Location: La Quinta, CA Posts: 220	The calibration level you are using is derived from a film standard and when compared to movie soundtracks, music tends to be much more compressed, with reduced dynamics and greater overall level. Because there is no recognized standard for monitoring level in the music industry, music mixes may seem too loud when played back on a system calibrated for an 85dBc reference level. Typically for music applications, we recommend calibrating for a lower reference level, such as 78dBc. With regard to the 500-2500Hz filtered pink noise measuring at around -16dB FS: The bandwidth limited pink noise will read higher than -20dBFS, because pink noise has a high crest factor (12dB), which will throw off the RMS / peak reading on the meters in your DAW. In order to get an accurate reading, you would need to use an averaging technique, where multiple time samples are captured and averaged over a ten second interval to calculate the RMS level. This is why we included the 1kHz test file, which is used for setting the electrical / digital levels. Once you set the faders / output of your DAW to -20dBFS / +4dBu / odBVU using the 1kHz tone, then you should leave them in this position, when doing the acoustic calibrating, using the bandwidth limited test files. I hope this helps to explain what you are observing. Cheers! __________________ *Pascal Sijen* Director, Product Management, Cinema Solutions @ Dolby Labs \| Former Co-Founder Blue Sky Int. & Audio Design Labs Inc. \| I'm also a geek in my spare time.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread starter	Forum	Replies	Last Post
Measurement Mics	aussie_techie	So much gear, so little time!	1	2nd April 2008 07:37 PM
Recommendations for Measurement PC	thadman	Music computers	4	20th February 2008 05:59 PM
measurement microphone ?'s	Pistolpete	Low End Theory	11	2nd January 2007 08:52 AM
When High End Goes Horribly Awry!...	meteor	Remote Possibilities in Acoustic Music & Location Recording	14	16th September 2006 04:14 AM
when protools goes horribly wrong	chrisgraff	So much gear, so little time!	29	2nd November 2004 08:52 AM

23rd October 2009	#3
Mandachuva Gear Head Join Date: Feb 2008 Location: Timisoara, Romania Posts: 44 Thread Starter	Thanks for the quick reply, Ms. Sijen. I understood the calibration technique using the sine why - dunno really why I mentioned the peak meter readings for the filtered pink noise. I'll give it a shot tomorrow to calibrate it at a lower level, hopefully with better results... In truth, loud mixes were showing as 90-95+ dB SPL on the meter, using the previous calibration technique. All the best, Mihai Toma