Trying to correct issues with hiss and DSP EQ

[HarvBow07]

Hello ASR, new member and first time posting. I recently found this site and am enjoying all the measurements. I have fallen prey to many of the ideologies and snake oil products that ASR tries to protect against, so thank you. I have a few issues with my current setup:

Sonos (digital) Oppo BDP105 (analog) Speaker antipode 2.0 (analog) minidsp2x4 crossover (analog) emerald physics EP100.2se x2 - emerald physics ep2.7. Speakers currently have a passive crossover for mids to highs.

When I bought the speakers and amps new, the dealer and engineer did a custom calibration and setup in my old house. They sounded wonderful except for some hiss that was and is audible from the listening position. Even though they are high efficiency speakers I would like to try to correct that. Furthermore, the work they did was saved in the minidsp 2x4. It sounds terrible in my current house and I have just lived with it for 4 years now. No matter what I need to change the eq settings. After reading a bunch on ASR I have come up with two options but they may not be the only options. I think one move forward could be removing the DSPeaker antimode and minidsp2x4 both of which seem to test poorly. At the same time I would like to remove the passive crossover and use an active one of some type. Two products that could fit the bill are the minDSP flex 8 and the Sublime Acoustic K231 analog active crossover. If I go with the flex I would get the Dirac license and try that out. The other option I see is to try to buy the miniDSP software that will let me mess around inside that unit. However because of the noise and the newer offerings I m leaning towards number one.

In a perfect world I would like it all connected with balanced cabling as I plan to buy one or more Buckeye amps. However I think in my case I could use adapter cables from the flex8 to the amps and be all right. Any problems or better solutions to consider? Also is it sound logic that replacing components with high noise and low sinad with better measuring components could help the hiss?

Sorry for length and thanks in advance,
Harvey

 

[HarvBow07]

I forgot to mention the hiss seems to be coming from he miniDSP2x4. When I remove it from he chain the noise goes away.
Harvey

 

[Sokel]

One good thread about it.
I don't know if the new firmware is out or if it applies to 2x4 thought.

DSP Measurements and Rising Noise Floor

In several threads related to DSP platforms I've seen requests for measurements with filters engaged. I thought it unlikely that implementing filters in a DSP would have measurable effect on noise / distortion other than the obvious change in the fundamental, however @Sokel recently mentioned...

www.audiosciencereview.com

 

[HarvBow07]

Thank you. There is a lot of data there which I don't quite understand yet. I will read it a few more times. I think the takeaway is the filtering can generate noise but is still inaudible? Towards the end the comments describe my setup very closely. I have very high efficiency speakers, with high gain amps, and low sinad control components. However I do not have a 2x4hd, it’s the non HD version without sharc so I do not think the firmware would help me.
Harvey

 

[DavidMcRoy]

SIDEBAR: I have no experience with any of the miniDSP products. I use the Dayton Audio DSP-408 exclusively. (It's 4 in, 8 out.) I have never encountered any noise issues with them. I should buy a brand new one and send it to Amir for testing. The specs say "Chipset: Analog Devices ADAU1701 SigmaDSP 28-/56 Bit Audio Processor." I believe they operate at 24/96?

 

[mdsimon2]

Your setup is not well optimized from a noise perspective. You have three digital to analog conversions and two analog to digital conversions. I don't know about the Oppo, Antinode or Emerald Physics but the miniDSP 2x4 is certainly high noise and will compromise the entire system. I would not attempt to switch from a passive crossover to active crossover if you have hiss issues as it will likely make them worse as the passive crossover likely has padding on the tweeter.

To achieve low noise you want to only have one low noise digital to analog conversion followed by a low noise amp with appropriate gain. The Sonos in to a higher performance miniDSP product like the Flex or SHD followed by a Hypex amp would be a good choice from a noise perspective.

Michael

 

[HarvBow07]

Thank you for the replies. That's an interesting unit David but I think I want a little more functionality than that one offers. Michael, I can certainly start off by leaving the passive mid-tweeter crossover in place. I do like the idea of all active crossovers though. I could try it and see if I bought the flex 8. If I hooked the oppo to the miniDSP over spdif I could go down to no AD conversions and only one DA conversion. So far that is the way I am leaning. It appears my options for a commercial offering are limited if I want a minimum of 2 in and 6 out for a unit that does crossover plus DSP room correction.

Harvey

 

[DavidMcRoy]

Thank you for the replies. That's an interesting unit David but I think I want a little more functionality than that one offers. Michael, I can certainly start off by leaving the passive mid-tweeter crossover in place. I do like the idea of all active crossovers though. I could try it and see if I bought the flex 8. If I hooked the oppo to the miniDSP over spdif I could go down to no AD conversions and only one DA conversion. So far that is the way I am leaning. It appears my options for a commercial offering are limited if I want a minimum of 2 in and 6 out for a unit that does crossover plus DSP room correction.

Harvey

I can see that the miniDSP can do some things the DSP-408 cannot, although they aren't anything I need, FIR filters, etc. I had assumed that the widely reported noise issues with miniDSP units in general were to do with user-end gain structure...maybe not hitting the units' inputs hot enough to overcome internal noise, but again I have no hands-on experience with them so I might be wrong.

 

[HarvBow07]

I ended up getting the miniDSP flex8 and have installed it. I must say I really like this unit and its remote. I hope Amir tests it at some point. I ended up gutting the cabinet of many components and cables. The noise has been greatly reduced to the point where I can not hear it from the listening position at all. I have to get within about two feet of the tweeter to hear anything even with 30 db gain amps. Sound has improved quite a bit with the old EQ no longer in the chain. I haven't ran Dirac yet so I am excited for that. I am running the speakers all active and they are doing very well. Efficiency is up quite a bit. Thank you to all for your input.
Harvey

 

[HarvBow07]

I have been trying to optomize my system for a month now. I am pulling my hair out. I hope to get some guidance on dealing with a problem room. My speakers are open baffle and each has 2 15" woofers a coaxial 12" midrange with 1" polyester compression tweeter inside. The crossovers recommended by the mfgr are 250 and 1200. I have played with different crossovers but the original settings seem to be the best sounding. I refuse to believe (fingers crossed) I can't get decent bass from them. My room is exceptionally large as it is not enclosed at all. Everywhere around my listening position is a bass black hole. I don't mean it's thin, I mean none. Think 4" mini monitor. I have moved them extensively to see if I can get the resonnances to change and I cannot. I have placed a sub woofer in the listening position while mapping the room and cannot find any place to put the speaker that has any reinforcement. I have tried equalizing the drivers using REW and then exporting the PEQ to the miniDSP flex 8 DL. I have ran Dirac at every position possible. Dirac does amazing things but it is unable to fix this. I ended up placing a subwoofer 6 feet from my listening position as there was reinforcement there when I had the sub in the LP. It helps greatly, even for a crappy sub at low volume. However I'm losing almost all energy from 120hz and down and without Dirac I lose energy at the 1200hz crossover. So I still have a huge hole in the midbass. These speakers have been subless their whole life because they truly will go to 20 hz with DSP. I have no problem buying subs but I'd like to make sure I have done all I can to optimize the speakers first. PS the midrange a treble are to die for and the soundstaging and imaging leave people in disbelief.

The "room" is 21x16x9-16 but it is open to the whole house by huge openings. I will attach some REW files in case it helps. Please let me know what I should be measuring if I haven't. I have tried aligning the spectrogram plot which was decent to begin with, measuring time delays with the step function, flattening responses manually, changing polarity of different drivers, etc. Even when Dirac makes the speakers follow a Harman curve by brute force (attached "No sub with Dirac"), the measurements show bass but ears will not. I'd like to get them down to at least 60hz as I am able to locate a sub when crossed at 80 or higher. I am using UMK-2, I don't know how to do FIR filters and the rest of this is pretty new to me. Although I am getting more familiar with REW over the last month. The speakers currently rest 58" from front wall, 36" from side wall (center of tweeter) and I listen 50" from the rear wall. Speakers are 11' apart and I listen 12' from them. There are two 5.5" thick GIK panels between the speakers on the front wall at the front wall's first reflection. In case it matters, I don't care about any listening position other than my own.

Thanks in advanced for any help, please let me know if there is more information needed from me.
Harvey

 

[ozzy9832001]

Yeah I would post your .mdat file from rew, saved as a .zip file.

Measure the left and right speakers separately and then both together (just incase there is some weird summation issue).

Let's see if we can't figure this out.

 

[DVDdoug]

I can't get decent bass from them...

...These speakers have been subless their whole life because they truly will go to 20 hz with DSP.

That seems contradictory... Have you ever heard this 20Hz bass or is it "theoretical"?

If you're into open baffle speakers you probably know this, but the soundwaves from the back of the driver are out-of-phase with the front and without a cabinet they "leak around" the baffle, mix, and mostly cancel-out. It's worse at low frequencies (long wavelengths). Maybe with lots of DSP bass boost and enough power you can overcome the cancelation. But in general, you can get more out of a woofer/sub in a properly "tuned" sealed or ported cabinet.

 

[HarvBow07]

Sorry I thought the files already posted. Trying again it says the file is too large. Compressed it is 8 mb. I'll need to try another way.

Have you ever heard this 20Hz bass or is it "theoretical"?

Yes at my previous house where they were intitially setup and used for years.

 

[HarvBow07]

I can add a drop box link. I can't seem to get the file small enough with winzip or windows' built in compressor.

No sub with dirac 081323.mdat

Shared with Dropbox

www.dropbox.com

 

[AnalogSteph]

Hmm. No matter how much I window the responses, the mids and highs remain a bit of a rollercoaster, with fairly little difference between both speakers. That suggests to me that it is in fact the anechoic response of the speakers themselves that'll need some work in this area (thank goodness for DSP). How far away was the microphone, and what kind is it? I would suggest taking some additional measurements closer to nearfield (maybe 4-5' off?) just to be sure.

Regarding the mid-high crossover, you may need to measure both midrange and tweeter separately in order to get to the root of the problem. 1.2 kHz probably is a bit of a stretch for both drivers for different reasons, and phase at crossover seems to be way out of whack. The first thing I'd try seeing that big dip is inverting tweeter polarity, maybe it really is that simple. You're always going to need some FR / phase / delay massaging for things to line up prior to applying crossover in the stricter sense. For one thing, I cannot imagine that a tweeter stuck in the throat of a 12" cone has the same apparent sound origin, being located behind all of the cone... some delay for the midrange (and in extension woofer) is likely to be necessary to line them up in the time domain.

With fully active speakers, you get to have all the fun of a speaker designer.

The speakers currently rest 58" from front wall, 36" from side wall (center of tweeter) and I listen 50" from the rear wall. Speakers are 11' apart and I listen 12' from them.

OK, let me translate this imperial insanity so us metric folks can wrap their heads around it:
"The speakers currently rest 1.47 m from front wall, 0.91 m from side wall (center of tweeter) and I listen 1.27 m from the rear wall. Speakers are 3.35 m apart and I listen 3.66 m from them."

From my limited experience with dipoles, my gut feeling says that your back wall distance may be too small, and you may have issues relating to SBIR. In a dipole, this should strike whenever the path difference is a multiple of a full wavelength (as opposed to a CB speaker, where it would be 0.5 wavelengths, 1.5 and so on). The smaller dip at 115ish Hz would fit this pattern. Not sure about the big one at 66ish Hz though. I would try out REW's Room Sim feature and model the room for further insight.

For the time being, I would try getting the speakers 3-4' closer to the listening position, which should be no problem now that the hiss is lower.

 

[HarvBow07]

Thnaks so much for the input. I do realize one mistake I made. When I tried to PEQ each individual drive I had the intended crossovers set in REW. I realize now I should PEQ them to above and below quite a bit, then apply crossovers in miniDSP. I will try that again. I do have mdat on each driver from the listening position with no crossover present and widebanded. I will post the link for that:

Individual drivers no xo preset 3.mdat

Shared with Dropbox

www.dropbox.com

Can you gain any insight regarding better crossover frequencies from that? The tweeter is a BMS 4545ND. The specs are conflicting as it says under features that it is crossed at 1.2khz. But under specifications it states recommended cross at 1.9khz. I have not been able to ID the mid or woofer. The rumor is they are made by Eminence.

I don't think the wife will let me pull the speakers another 4 feet into the room. She is very tolerant but 9 feet is a lot to ask. I guess step one would be to measure drivers near field (3 feet?) and PEQ them in miniDSP. Then do the crossovers in miniDSP. Then remeasure each driver. Then calculate the delay for the tweeter mid and woofer. For delay do I just use the delay stated in the text box after a measurement? I was hoping to do that but when I reference the L tweeter, then measure the L tweeter I get a negative delay. I would think it would be zero or close to it.

Does that sound correct?
Harvey

 

[Philbo King]

You might try the opposite approach - move your speakers as close to the front wall as possible. It should bring the low bass up significantly.

 

[HarvBow07]

You might try the opposite approach - move your speakers as close to the front wall as possible. It should bring the low bass up significantly.

I tried as close as 2 feet and heard some change but not very much. It also closed down the soundstage quite a bit. I am playing with this as we speak. PEQ applied to each driver as stated then measured again and low and behold no change. Something is going on. miniDSP says the sample rate for biquads has to be 48 khz. The miniDSP2x4HD preset in REW defaults to 96 according to the title. Somewhere I read to use that preset for the flex8. I think I will have to redo everything for generic 48khz in REW. And of course REW randomly closed and I lost measurements... I'll tell you what.

 

[HarvBow07]

Ok using 48khz generic worked! All data I have for near field I will post a link for. I PEQed each driver and have before and after. I ran full range sweeps of each speaker before and after. I found out that I did indeed need to invert the tweeters. I also changed the crossovers based on the plots. I chose 300 and 1.9k. I'm interested if any of you would have chosen otherwise. And lastly I ran both speakers full frequency from the main listneing position and it looks bad. But the near fields look much better! I haven't listened to it yet so I will hold off celebrating until then. Any other advice based on the new data would be welcome.

New each driver no xo PEQ try 081923.mdat

Shared with Dropbox

www.dropbox.com

Harvey

 

[dualazmak]

Although quite belated, I just noticed this thread.

I assume my DSP-based multichannel multi-SP-driver multi-amplifier fully active stereo audio project wound be somewhat of your interest and reference.

You can find find here (on project thread) and here (remote independent thread post) the Hyperlink Index of this thread and some of my related posts in remote threads.

Jus for your quick reference,,,
- The latest system setup of my DSP-based multichannel multi-SP-driver multi-amplifier fully active audio rig as of August 3, 2023: #774