0dBFS sweeps in REW may not represent the maximum SPL of the system

[ernestcarl]

I asked a rather off-topic question in the ff. post asking why my measured SPL peaks with music playback went above the measured 0dBFS full-range sweep traces generated in REW.

Ideas for more meaningful speaker measurements

Give the Highland Brewery a visit. One of 67, but I find their ales to be the best in the state. Thanks! … All that is technical, measurable, tunable. What is not is how your brain inherent, genetic, and training, processes what you hear to decide if it is "musical" or not. Well either...

www.audiosciencereview.com

It turns out I just completely forgot about the fact that music has a dynamic range (duh!) and peaks / crest which is also why one should always ensure adequate headroom is available to prevent any clipping/limiting.

I'm not using an AVR, but a home theater PC with JRiver media center software outputting directly to a multichannel DAC. Maximum reference level is set manually at a much lower level since I only have a single sealed 12" sub that isn't able to reach "THX" reference levels. I do not want to risk bottoming out the sub with LFE signals exceeding well above the subwoofer's compression point... so I've been using a global negative gain on top of JRiver's own built-in or internal digital headroom -- so basically an addtional layer of negative gain per output channel.

Anyway, I assumed that by disabling "volume leveling" and generating a 0 dBFS signal in REW, I would have reached my system's maximum volume limit -- above which the system would automatically "soft clip". However, playing any file or streamed content at max volume -- even with the volume leveling in my software media player turned off -- still gives out LZpeak SPL levels that exceed above the seen visible sweep magnitude trace lines taken at "0 dBFS" in REW.

For one thing, routing audio through the JRiver software automatically automatically applies a ~ 4.7 dB reduction in the digital volume gain (default extra headroom for DSP purposes) as opposed to directly sending the signal to the DAC. Whereas I got a reading of 87.5 dB with a -20 dBFS PN signal sent directly to the left channel of my DAC, I got a reading of 82.5 dB when routing the signal into JRiver's WDM driver (with volume leveling off). Now, with the volume leveling turned on, this will additionally apply a 10 dB negative gain yet on top of the initial -4.7 dB -- giving out a final reading of 72.5 dB at the MLP! So there is indeed already adequate headroom available to account for maximum volume playback.

manual routing with no PEQs running (leveling disabled)


manual routing with all PEQs enabled (leveling still disabled)


Hmmmn... thinking this through again, I would not recommend others, nor dare risk myself sending out a 0 dBFS signal directly to the DAC and speakers -- besides there's always the potential of clipping elsewhere in the chain...

In my case, 87.5 dB (-20 dBFS signal measured at 2.3m) + 20 dB more giving out an theoretical equivalent total of 112-113 dB FULL-RANGE SPL signal at 1 meter! My studio monitor is only spec'd at 102 dB continuous max SPL -- even though the maximum (transient) "peak" is written at 116 dB -- one must assume this is of a limited range (250 Hz - 3 kHz perhaps?)


-------------

BTW, I haven't checked other software media players. But I presume some of them might also apply a negative gain pre-emptively (?) -- the smarter ones with extra DSP features, at least.

 

[ernestcarl]

In my case, 87.5 dB (-20 dBFS signal measured at 2.3m) + 20 dB more giving out an theoretical equivalent total of 112-113 dB FULL-RANGE SPL signal at 1 meter! My studio monitor is only spec'd at 102 dB continuous max SPL -- even though the maximum (transient) "peak" is written at 116 dB -- one must assume this is of a limited range (250 Hz - 3 kHz perhaps?)

I performed burst test tone signals with the microphone positioned 1 meter away from the horn opening. The spec'd 116 dB peak is actually for two speakers or a stereo pair. The LZpeak for a single speaker (Sceptre S8) ranges anywhere between 110-113 dB -- measured in-room away from most boundaries except the sidewall. Below 80Hz in the bass, burst tones exceeding +100-2 dB (ports closed so already a ~3-4 dB loss incurred) would momentarily activate the internal limiter -- red flashing light of front LED. However, while the speaker will play at a higher SPL the more you push it, the acoustic impulse peaks will start to compress. If one wants to reach sub bass peaks well over 105 dB still with acceptable sub bass distortion, one would need to at least enable the built-in digital 60Hz high-cut or HPF of the monitors and augment them with a sub(s).

*In the past, I did insane sweeps exceeding 110 dB per speaker (pushing the input gain way up) but this also activated the protection limiters -- without the limiters, I would have blown something for sure.

 

[gnarly]

Hi,
If i understand correctly, REW's output is going to J-River digitally, which then passes on the digital signal to an external DAC.
Which makes the J-River / DAC combo stream essentially a sound card for REW's output.
Is that correct?

If so, i'd suggest calibrating the REW Generator's output, using the 'Calibrate level' button by measuring the voltage output of the external DAC.
Snips in first post show the Generator with REW's default FS sine Vrms of 1v.
That should solve most of the SPL disparity between REW's 0 dBFS, and J-River's file playback (without volume leveling and/or J-River fixed attenuation).

One thing I've found helpful testing vented subs at higher SPL, is to raise the Generators start frequency to a little below port tuning frequency, until I'm sure the subs high pass filter is handing potential unloading and over excursion.

 

[ernestcarl]

built-in digital 60Hz high-cut or HPF of the monitors

Oops... I meant low-cut filter -- always get the terms mixed up.

If i understand correctly, REW's output is going to J-River digitally, which then passes on the digital signal to an external DAC.
Which makes the J-River / DAC combo stream essentially a sound card for REW's output.
Is that correct?

You are right. I've set up REW's output directly to JRiver which controls the channel mapping, bass management, EQ, convolution, and other misc. DSP.

If so, i'd suggest calibrating the REW Generator's output, using the 'Calibrate level' button by measuring the voltage output of the external DAC.
Snips in first post show the Generator with REW's default FS sine Vrms of 1v.
That should solve most of the SPL disparity between REW's 0 dBFS, and J-River's file playback (without volume leveling and/or J-River fixed attenuation).

My mch DAC at full-scale output should be ~1.14v -- at least, according to Amir's review:

And I've already figured out what the remaining SPL difference was by changing REW's output setting directly to ASIO driver of my DAC instead, fully bypassing JRiver. Although, for quite a long time, I've always wondered why the SPL was lower whenever using JRiver even with volume leveling turned off...

One thing I've found helpful testing vented subs at higher SPL, is to raise the Generators start frequency to a little below port tuning frequency, until I'm sure the subs high pass filter is handing potential unloading and over excursion.

The S8 monitors I use currently are sealed, so no longer vented -- I've rechecked some older measurements, and it seems the SPL loss is probably closer to 5dB around +44Hz or so. Because of my use of a subwoofer, the loss of bass below 100Hz isn't so much of a big deal. At very loud listening volumes of specific bass heavy tracks, I thought I could hear the port "chuff" -- or maybe it's some combination of other box distortion escaping from the ports(?) Sealing these didn't really help as much as just using the internal low-cut filters of the monitors. In fact, despite cleaning up the response a little, sealing the monitors is unnecessary for the most part. I thought about re-opening them, but I do have a young kid nephew that likes to shove things inside holes; and, in fact, I've caught him a couple of times attempting to do so with these speakers... eh, I think I'll just keep them closed for now!