TOPIC:

This question pertains to the mix levels on the LFE Mgt tab of the bass management plug-in for the 88A. In the Applications White Paper titled "Bass management with the DDRC-88BM", in Section 3, the instruction clearly states, "Check also that the mix levels are all set to 0 dB". No additional details are given to support this recommendation.

However, in a response to an 88A user who was experiencing the "static pops" issue, MiniDSP Technical Support stated "In the xml file, the input to LFE Mgt are all 0dB, when they sum up, the magnitude could be quite high (see LFE_Mgt.png), try reduce those magnitudes to see if that helps.

In discussions in the 88A thread on AVS, several knowledgeable contributors claim that the mix levels for the redirected bass should be attenuated to prevent over-driving the signal, potentially resulting in clipping in the LFE channel. A value of -10.2dB has been recommended as the appropriate level of attenuation.

I am looking for specific guidance with respect to the mix level setting. Since I cannot measure clipping without an oscilloscope, which I don't have, I am "flying blind" without additional guidance. I find it somewhat strange that there is not a better description of this setting in the User guide, and what effect the setting has if it is left at zero, or attenuated to -10.2dB.

AustinJerry wrote:

However, in a response to an 88A user who was experiencing the "static pops" issue, MiniDSP Technical Support stated "In the xml file, the input to LFE Mgt are all 0dB, when they sum up, the magnitude could be quite high (see LFE_Mgt.png), try reduce those magnitudes to see if that helps.

Did it?

I am the user Jerry mentioned. There was one other comment from miniDSP about my config may be relevant:

In the Dirac Live project, a custom target was setup for the subwoofer channel which cover the whole audio band, and the audio at 10k – 20k get boosted up somehow (see custom_target.png). If you click “Auto Target” button which is Dirac’s default, the curtain limited the optimization band to up to about 200~300Hz only, which avoided the high frequency band get boosted up (see auto_target.png), so, try to revise the target setting here to see if it helps.

I trimmed off 2dB from the inputs (guessed at what value to use), and fixed the subwoofer target (sub had accidentally been grouped with the mains when I set the target curve) and I re-loomed my wiring to reduce strain (using unbalanced rca in&rout). I did all three at the same time, then returned the input levels to 0dB. I haven't heard any further pops.

Thank you for your reply, but let's focus on my original question. There are already several threads on the popping noise, so no need to discuss that issue here.

The uncertainty is:

- Is there really a risk of over-driving the LFE channel if the mix settings remain at zero?
- How would one measure whether there is over-driving? One method I am aware of is to use an oscilloscope to observe whether there is clipping on the LFE output. But I don't have an oscilloscope, and don't want to incur the expense if there is an alternative way to measure.
- If the risk of over-driving is real, then what level of attenuation is recommended? Again, with a scope, the attenuation level could be dialed in, but I don't have one. A recommendation has been made that -10.2dB is the attenuation that will cover all cases, but this level of attenuation results in a 10dB loss of output levels for all speakers (since the mains must be attenuated as well in order to maintain the balance between mains and LFE).

john.reekie wrote:

Did it?

John, if you were to suggest a way to measure, I would be more than happy to follow your recommendation.

Since I don't have an oscilloscope to check for clipping, I thought that feeding the 88A a high-level signal and observing the output meters might be a good substitute to help answer my question. So, using REW's signal generator, I output a 60Hz sine wave at -3dBFS, with the AVR master volume set to 0 (reference on my system). I selected the center channel to measure re-directed bass, and then the LFE channel. In both cases, the channel 3 (summed bass) output level never exceeded -9dB, which is below the "red" area.

Is this a reasonable test to determine that there is probably no clipping going on? The first file is the center channel and the second is the LFE channel (10dB higher, as expected):

Sorry for the short response below but I am a bit pressed/stressed for time, will elaborate more later.

As long as the meters don't go above 0 dB, there is (should be) no clipping. The DSP is floating point and will have a "fair amount" of headroom internally. Unless something is seriously awry, my understanding is that the clipping, if there is any, will take place at the output where the floating point data is converted back to integer for the DACs. So increasing levels anywhere e.g. in the LFE mixer, then descreasing the output level accordingly, will result in no clipping.

You can use your computer and soundcard as an oscilloscope with some caveats. I use Signalscope on Mac (www.faberacoustical.com/apps/mac/signalscope/) but there will be similar for Windows. Perhaps this one is worth a try: www.zelscope.com

While you can monitor the waveform from the mic output, this will result in injurious signals to yoru speaker if you are looking at extreme conditions like clipping. An electrical test where the outputs of the DDRC are disconnected from your amps and connected directly to your soundcard input would be the way to do this kind of test.

The main caveat in this scenario is ensuring that your levels are such that you are not clipping the soundcard input A/D.

Appendix: the input level between your AVR and the DDRC is also a factor. Ideally for the test you would want the level to be just below the max input voltage specified (depending on which input and sensitivity setting).

You could also monitor for clipping in the REW RTA with say 1/48 bands - a sudden rise in harmonics will indicate clipping.

Thanks, John, you have given me several ideas to explore.

Hi Jerry

Some great comments already from John!
Our very specific answer to your question is "It depends"..

No seriously.. each install is unique, each comment we give (and we learn as we see many installs in front of us) will differ depending on the setup. With the DDRC-88A being analog in/out, we've seen our share of setups where people were indeed overdriving the input. Thanks to the RMS meter in the matrix we indeed can see that now.

If say you go unbalanced in, drive hard the 2Vrms in with a Sub out from an AVR (sometimes higher levels), then apply Dirac Live with 10dB boost in your target curve (seen many times), and apply bass management in a 5.1/7.1 Dirac live setup without attenuating as per our app note here ( www.minidsp.com/applications/home-theate...ement-with-ddrc-88bm )then even if you have floating point, you can get in trouble. So that was our suggestion.

With this said, not sure it's a generic comment. Just something to look into. The RMS meters is just the detail to look into here.

Keep us updated on your findings.

Devteam

Thank you for your reply. I have carefully read the "Bass management with the DDRC-88BM" guide and am confident that I have followed the guidelines.

I am using balanced inputs on the 88A, and have carefully measured the output voltages from my Marantz 8802A pre-pro, with results shown below. So, based on these conversations, and the readings on the RMS output meters, I am reasonably confident that I don't have clipping on the LFE Mgt channel. The voltage measurements were taken with a maximized signal, much higher than my typical listening levels.

To MiniDSP Support: Allow me to submit as a feature request a suggestion to provide peak readings on the output meters as well as RMS readings. Peak readings would be more useful to assess whether clipping is occurring.

Hi Jerry, I had always assumed they were a peak-detection type of meter, but if not, fair enough request.

I'm hoping I've cleared the decks for the next couple of days to hook up the DDRC to my soundcard and do some measurements. Will update....

john.reekie wrote:

Hi Jerry, I had always assumed they were a peak-detection type of meter, but if not, fair enough request.

I'm hoping I've cleared the decks for the next couple of days to hook up the DDRC to my soundcard and do some measurements. Will update....

I'll look forward to your results, John. I wasn't sure about the output meters until the earlier post from the Dev Team stating that they were RMS. It would be nice to be able to toggle the measurement type, similar to what can be done with the REW SPL meter.