MiniDSP - Page 50

Since my denon 4311 has subeqht should I use two separate rcas for my left and right subs? Right now I have a y splitter running to my separate amps. Two amps for the subs.

Are you using a minidsp? Do you have measurement gear? Do your subs match?

It's kinda up to you. Do you want to set distance/delay on the AVR or minidsp. How do you want to fine tune the trims?

I would
-Gain match my subs if they match
-Determine best location for subs based off a raw(no audyssey no minidsp) FR plots trial and error from available locations
-hook in the minidsp as a passthough doing nothing, continue to use the 2 subs outs from the SubEQHT AVR
- Run audyssey to get the distance and level really close, look at the FR graphs
- See if you can improve the transition area by tweaking delay on the minidsp
- Save best graph so far
- Turn audyssey off, Make FR plot as good as possible with minidsp. Try their auto correction, try a bunch of stuff, hopefully educated trial and error and a bunch of research on good vs bad EQ techniques. Too many filters bad, less is more, boosting bad for head room is a myth, cutting has the same impact. Don't make big boosts. You can't boost your way out of a true null
-Save best plot like that
-rerun audyssey
-measure, is this the best yet?
-enjoy, think about what can be better based on what you learned, tweak when you have time if you enjoy that sort of thing.


I have seen folks not use the SubEQHT once they have measurement gear and a minidsp like gadget. If you use one sub out you get the distance on the AVR right for the furthest sub and add delay via minidsp to the closer sub to optimize the FR of the transition region.

Yeah I have the 10x10. The subs can't be moved but are in line array so distance is the same. I have measurement gear and am having issues that's why asked about the sub eqht. So running each output from avr separate is optimal or not? Right now just using one input/output of dsp for subs. Will I get higher voltage running two inputs/outputs?

And, when in the xover what is difference in using sub option and peak option?
Edited by audiovideoholic - 2/28/13 at 6:12pm

It's just a matter of where you make the settings then. Does SubEQHT come up with equal distances and levels? Since you can manually address the two things SubEQHT sets (distance and trim) It's truly just a matter of where you set it. I'd be tempted to measure once with SubEQHT doing it, then 0 out those settings and do it with the minidsp equal settings and measure. In theory it should come out identical though.



I don't know. Minidsp experts help

marantz av8801 it does have subeqht, thats why i need atleast 2 discrete inputs for the minidsp so i can smooth things down some more if necessary after audyssey.

Ok, need major help. Have searched net with no luck.

My subs and room by itself have similar FR. I have a spike from around 35hz that is smooth to about 75hz that is aprox 10-15db hot. This is a big hump that I have been trying to deal with for about a week now with no real luck. I've tried cutting, boosting, raising q to bring up the lower FR to match it, all with no luck. I'm running the 10x10 with 8 subs in a line array across the bottom of my screen that are powered by two FP10ks out of 2 outputs from the mini.

What would be the best settings to tackle this? Using the peak eq, sub eq etc...? The sub eq limits fr control to 50hz and my problems are above that. All my attempts at smoothing it out have failed (10 plus hours of measuring). I'm beginning to think the device just won't do it or maybe I'm in a much larger situation that just don't know. Would using all 8 outputs help since using two split into 4 for each amp isn't helping. I'll also add that this hump is present in all locations that I place the mic.

Should the input levels be zerodbs? Should the input settings be flat? I'm also wondering which tones are more accurate to measure with as the graphs improve great on omnimic with the distortion track (long sweep from 5-20000hz) or short fast sweeps? If the distortion tests that run rewlike sweeps are the best options then maybe the mini is doing a little better than it appears when running the fast sweeps.

I'm at a loss and starting to regret the purchase of the mini since I don't see changes no matter what I do.

Note, I've tried delay and phase but no change in readings.

Sorry for rambling but after 10 hours of messing with this thing I'm getting fed up. It worked great for my LCRS but not changing the LFR hardly at all with my attempts. Maybe I'm doing something wrong with the peak, sub, etc settings.

"My subs and room by itself have similar FR."

what does this mean? it doesn't seem to make any sense because subs will have a response that rolls off while rooms will have a response that provides gain the lower you go.

Lol yeah I know! I took snapshots while was out in theater but haven't loaded to photobucket yet. Will show!

I'll try to explain. When I have the measurement gear all plugged in and ready to play a sweep the mic is recording the room response with only CPU running and that's it. Volume is turned up on equipment but nothing is playing and yet my graph shows a very similar response/pattern/slope as when the measurements are actually taken.

For instance, the measured sweep shows gains in 35-75 hz range (a hump that is smooth) and before I press play the room floor has the same hump but a tad more narrow. Like 50-65hz spike while nothing is playing. I have asked in another thread about a room having its on unique fr measurement and someone said yes that is possible. But what I'm trying to figure out is if my drivers actually have this funky curve or if the room is somehow dictating it making it almost impossible to eq it out.

I'll post the graphs tomorrow. Maybe I explained it well enough to understand what they will show.

"Volume is turned up on equipment but nothing is playing and yet my graph shows a very similar response/pattern/slope as when the measurements are actually taken."

one thing that you may wish to do if you already haven't is a close mic response (mic right in front of dust cap) of one of your subs just so we know what is coming out of the subs. i'm sure you will get it sorted out.

Yeah. That's my next objective after zeroing out the dsp settings. We shall see but just weird since have same hump in both rows of seating. I have a feeling it's electrical 60hz hum and is getting amplified along with volume increase.

That's easy enough to test; just look at the spectrum via REW/RMAA with no signal and you should have a nice defined 60Hz spike.

Anyone know how the plugin is transferred to the buyer when you sell a MiniDSP, or do they have to buy it again?

Put your copy on a flash drive or email it the buyer.

I use the same plug-in for both my Minis.

great, thanks, Oliver

Ethically, that depends on whether or not you still use the plugin.

If you don't still use it, then send it by any physical or electronic means (you can always download the original dmg file from miniDSP if you've purchased it, I believe.)

If you do still use it, then it depends on what representations you've made to the buyer, but one of you should ethically buy a new copy of it.

I'm be selling the MiniDSP so wouldn't be using the plugin.

plug in downloads, updates etc are tied into the account you made on minidsp's website. but as others have stated, you can pass it out via usb or email etc.

I couldn't find a plugin software license on their website which is where I'd suspect to find the transfer of ownership lingo. I'm of the opinion you own that software and you can sell it. Can you give him your user account on the website? Ideally he would be able to get the future upgrades too.

I'm going to be jumping on the miniDSP bandwagon here soon. I'm thinking I need to keep it simple and just go with the 2x4 balanced with the 3/4 way advanced plug-in.

On the other hand, should I future proof myself and pick up the 10 x 10HD just in case? What, if any, is the advantage of having a higher sampling rate? 96kHz for the 4 x 10HD vs 48Khz for the 10 x 10HD and almost everything else?

This question has probably been asked schfifty-five times. Don't hate me...

i believe the hardware is the same, +/- a few connectors

its the plugins different i am pretty sure the 4x10 plugin works with the 10x10, its just that with a 10x10 plugin, the minidsp does not have the processing power for that higher sampling rate.

Ok, but...a lower sampling rate = what?

Loss of sound quality?
Unable to process information as fast?

Is there a notable benefit to having a higher sampling rate?

Hi Popalock, There are a number of trade-offs between sample-rates, but most of them are inconsequential in an application like the miniDSP. The only material difference that I see (without thinking hard) is the quality of the anti-alias filter.

If you sample at 48kHz, then your incoming analog signal has to be completely free of energy above the Nyquist frequency of 24kHz. So your anti-alias filter has to roll to zero between 20kHz and 24kHz. That's pretty steep.

If you sample at 96kHz, then your incoming analog signal has to be completely free of energy above 48kHz. So your anti-alias filter has over one whole octave to roll to zero. That's much easier.

The higher sample-rates can translate into lower cost or lower noise, and sometimes both.

Ok, thanks for the input Mark. All factors considered, sounds like the 4x10HD is the best bet for me.

What sound quality implications would that sharp roll off have if any? Meaningful material above 20K is nonexistent isn't it? I mean I cant hear above 18K effectively and most speaker systems don't extend above 20K anyways, so whats the effect?

Is it safe/correct to say if one is not concerned with >20KHz audio content then 48kHz sample rate is sufficient?

Hi Nick, The sound quality implications are pretty bad, because anything above the Nyquist frequency will not stay there - they will "aliased" into a lower frequency and become audible noise.

Lets say you are sampling at 48kHz, and have a less-than-perfect anti-alias filter (they all are). It happens to have an -80 dB lobe in its stop band at 35kHz, and therefore lets energy into the ADC at that frequency. That 35kHz noise will be digitized as -80dB 13kHz noise (symmetrical around the Nyquist frequency), and now cannot be filtered without affecting the original content.

Any frequency above the Nyquist frequency will be "mirrored" to a frequency below the Nyquist frequency by the sampling process, as in the 35kHz becoming 13kHz when Nyquist=24kHz. So it is very important to remove all energy above Nyquist prior to digitizing, because it can't be removed after digitizing, because it is now mixed in with your audible content.

With a sampling-rate of 96kHz, any noise would need to be above 48kHz in order to be aliased down. And even then, it would need to be above 76kHz in order to be aliased below 20kHz.

That's interesting Mark. I hadn't thought about A/D converters much.

Looking at the ADAU1701 (miniDSP chip set) datasheet, its anti-aliasing filter is of the "half-band" type, so it's only 6 dB down at 24 kHz for a 48 kHz sample rate. I don't think I like that very much.

There's no anti-alias filter on the miniDSP board? Is there an image filter at the DAC output?

Curious - Don

Hi Curious - Don, I was wondering the same thing.

Normally, the initial analog anti-alias filter is external to the ADC. But the ADAU1701 shows a filter-plot (Figure 9. ADC Stop-Band Filter Response, on page 14) which seems to indicate that a filter is included internally. ^

That doesn't mean that the miniDSP people might not have added an analog filter. On the otherhand, they also might have assumed thet there would be no content above 20 kHz

Hi Andy, Yes, that is an old trick that you can play when you're only designing for audio. The filter is down to -83 dB by the time you get to 27kHz, and everything between 24kHz and 27kHz will be aliased to between 24kHz and 21kHZ. So you are moving the frequency of the inaudible noise to other inaudible frequencies.

Where did you see the reference to the half-band filter? I could not find it. I have always thought of a half-band filter as digital.


^ Notice that the filter plot shows a lot of stop-band ripple, up to -83 dB. Coincidently, the THD+N spec is -83 dB. So the performance of the chip appears to be limited by this filter.

I just assumed it was. AFAIK, almost all audio A/D converters use digital anti-aliasing filters with upsampling nowadays, maybe with a little help from a simple analog filter. Using analog filters for the whole job would be very difficult to implement on-chip.

Hi Andy, Ah yes. I understand now. My familiarity with half-band filters is also with regards to up-sampling, but that is when you do anti-aliasing in the digital domain. With regards to digitizing audio, anti-aliasing needs to be done in the analog domain prior to the ADC, because trying to digitize frequencies above Nyquist will give erroneous results. With the ADC's that I've used, it takes the form of an RC filter at the input pin.

I guess that there must be some form of low-performance analog filter inside the ADAU1701, as there is that frequency-response plot (fig. 9), and there also are no instances of any analog filters in the reference designs that they show. Every Sigma-Delta ADC that I have used shows a recommended analog anti-alias filter in the datasheet. However, having never worked at the low-end, I have yet to use a chip that was as "all-in-one" as the ADAU1701.