Measuring Equipment for 7.1 ULF potential...

[maxmercy]

Just to make sure the sinewave doesn't get flattened or mildly clipped by the BD player.  In theory it shouldn't, but in theory, this thread shouldn't have to exist either.

 

The reason clipped waveforms summed is so bad is this:

 

"Gibb's phenomenon. Occurs when limiting the bandwidth of a wide-band signal (or truncating an impulse response). This is particularly important when the signal is clipped in the digital domain, but it3 applies generally. What happens is that a square wave (or hard clipped signal) can be viewed upon as a sum of individual sine waves of frequencies 1, 3, 5,... times the fundamental frequency. The flat top of the square wave depends on the presence of all harmonics at the right levels and phases. If some of the harmonics are removed by lowpass filtering, the peak value of the signal rises. When converting from digital to analog a low pass filter is always applied, so the analog level may be higher than expected."

 

Take 8 clipped waveforms, sum them, apply a 20kHz lowpass, then a ~120Hz lowpass, and you get 0dBFS+.  There is a reason Immortals sounds louder thanit measures.  The LFE signals are square waves in many of the scenes. With a lowpass applied, they get louder.  I'll demonstrate in another post.

 

http://www.audioholics.com/audio-technologies/issues-with-0dbfs-levels-on-digital-audio-playback-systems

 

JSS

[maxmercy]

20Hz Sinewave hard limited at -2dB from the sinewave's peak:

 

 

Then add an 80Hz lowpass:

 

 

Then multiply the problem by 8.  With TF4, Godzilla and others already at 0dBFS in nearly every channel, you can see where this would cause problems.  I only measure digitally now.  When I used to measure from SW out, TDKR showed me several instances of 'OVER' on the meter calibrated for a max of 128dB output level, and that means 0dBFS+.  It is happening, and with no "I ran it at +10dB on the sub" silliness, just all trims set at 0, no house curve.  And I do not believe it was clipping, because I would run films back at -10dBMV on my AVR when measuring the old way to ensure some headroom, but mainly because it was the level I would listen at anyway...had I known of these problems beforehand, I probably would have measured the old way at -20dBMV.....

 

JSS

 

 

[maxmercy]

My final settings to avoid problems since I listen at -7dBRef maximum:

 

1. Turn down all digital outputs via nanoAVR by 0.5dB.

 

2.  I no longer use nanoAVR for bass management, as it introduces too high a noise floor for my tastes (summed bass becomes 0dB to the digital LFE input, so 10dB of headroom is essentially lost, and having to turn up the amps that much more results in amplification of the noise floor).

 

3. Since the Denon can sum WCS cleanly at -11dB sub trim, I have it set at -7dB, giving me 3dB of headroom.  To avoid too much amplification of noise floor on the LCRS, I turned up the trims for them to +4dB, leaving 3dB of headroom before clipping at my -7dBRef maximum. 

 

I now have a quiet system that does not clip.

 

I do not have the displacement to run at Reference without significant distortion <20Hz, and Reference is quite loud in my space.

 

 

JSS

[SME]

Hmm.  Interesting points about the effects of low-passed square waves.  In your post, you say the limiter is at -2 dB, but I see -3 dB.  I assume that's a typo.

 

The peak level in the low-passed version is definitely high, by about 1/2 dB or so.  However, even if that occurs on all channels at once and they are all summed together, the result is still only an increase of 1/2 dB or so.  Remember that decibels are basically a multiplicative measure.  When you increase by 6 dB (approximately), you double the level.  Mathematically, multiplication factors out of sums, so the overall change is still only +1/2 dB or so.

[maxmercy]

The bigger the flat top, the larger the boost. Check out Immortal's LFE channel and apply a DTS lowpass and be amazed.

 

JSS

[SME]

What do you mean by "the bigger the flat top"?  Do you mean,  "the more square it is (or sharper the corners are)"?  Or do you mean "the higher in level it is"?  If you mean higher in level, then the LPF will still increase output by the same multiplicative factor and the same number of dB.  In terms of absolute level change, the increase will be greater of course, but we already take that into account when we say things like "+3 dB increase requires double power".

[maxmercy]

The greater the amount of clipped off portion of the waveform, to the point that they look like square waves. The slope of the lowpass has a significant effect as well, hence the mention of the DTS lowpass, which is a high order filter.

 

JSS

[SME]

Sorry, but what exactly does DTS stand for?

[Shredhead]

Dolby Truly Sucks.

[maxmercy]

DTS = Digital Theater Systems.  Here are the lowpass details:

 

"In the DTS-HD codec only [DTS-HD Master Audio™ (lossless) and DTS-HD High Resolution Audio™ (lossy)] the DTS-HD decoder applies a low-pass filter at 100 Hz (-3 dB) with a 60 dB/octave roll-off onto the LFE channel."

 

 

JSS

[SME]

Thanks!  I learned something new and interesting.  That is a very steep roll-off.  I wonder how it's implemented?  It would be interesting to simulate its effect on a square wave to see just how bad it can get.

[maxmercy]

I would guess it is implemented in the licensed decoders.

 

Simulation (cascaded 48dB/Octave + 12dB/Octave Butterworth) on a -6dBFS 20Hz square wave:

 

Square Wave:

 

Square Wave w/ DTS LowPass

 

~35-40% increase in digital signal amplitude.

 

JSS