Deqx pre8

and I should quantify I'm an IT engineer, deeply understand computing, enterprise networks and electronics. So if it was tough for me, I couldn't imagine others (watching your response above about modifying the host file / vpn's, resolving .local DNS entries and the likes made me wince! - not because you were wrong per say but imagine the audience even understanding what that is lol).

A fascinating discussion - found it by googling the new Pre-8.

Years ago I played (suffered) with the HDP series. Larry made things effortless and was a phenom but at the end of the day I found that this devices is more for people that like to play rather than listen to music. It was simply too complicated to set up, but even after setting it up it suffered from bugs that would remove the romantism from listening to music. Often times it would start to pop, or bug out / crash and so forth. This was an HDP-5.

I, too played with ribbons. Apogee Scintilla and then other open baffle speaker designs (of my own, or 3rd party). I finally gave up once I heard the units were being discontinued and thought maybe I'd give it another go with the newer "beta" units. But from what I'm reading here the issues are there but in a different way. And yeah, the strange random lack in comms from the HQ company didn't help. They would pick and choose when to respond and you really didn't feel supported at times. Somehow whenever I get the bug to go "active", after staring at more amps, more wires and dealing with bugs I get the "let's go back to basics", plug in the stock crossover, less amps and wow the sound is just better. Or maybe it's just more reliable. Forrestc I hope you keep it up and convince me to try it again. 

FYI: I have been experimenting with a couple of 384kHz/32-bit FLAC music files and I found that the Pre-8 plays these files with no problem when rendered directly from my NAS through Volumio. There should not be any down conversion going on since the ESS Saber D/A chip in the Pre-8 should be able to easily render these files.

On the other hand, when the same files are served through LMS (Lyrion Music Server) through the LMS DNLA/UPnP bridge plug-in, the file appears to be playing when viewing its progress on the Volumio player screen, Though the Volumio screen states that the file is in-fact 384kHz/32-bit and progresses right along, I do not hear any sound. I’m still trying to figure this one out.

I don’t think the software is there as of this latest release. You may want to submit this as a "Suggest a New Feature" request.

So far I’ve only submitted two request but I’m sure that I'll be adding more  soon.

The following are the two request’s that I have submitted:

1. For a provision to display the room response showing any automatic PEQ correction (if any) applied after all calibrations have completed. Also with the ability to auto-correct AND manually PEQ correct the room response.

2. For a provision in Volumio on the Pre-8 to allow installation plug-ins for the Volumio plug-in store.

As for the first request, if you have ever used the legacy DEQX-Cal, it has this feature. What was nice about it, is that it did have the Auto-Correct feature which by default would apply PEQ correction very conservatively. It gave you a great starting point and from there you tweak those setting or change it up completely. It applied the updates in real time where you could immediately hear any change. Right now as it stands, you cannot visualize room response at all. I assume that room response measurements are taken and adjusted automatically in step 3 of the calibration sequence.

The second request is asking for a feature at all other implementations of Volumio already have. Hopefully the Volumio folks can code this one for DEQX so that can keep working on enhancing the core product’s software.

Has anybody found a way to EQ individual drivers?  As opposed to EQing the summed frequency response. Thanks

Has anyone been able to connect to deqxy3.local through a VPN? I haven’t had any luck. So far I had to kill my VPN service in order for the connection to be made and use the Pre-8. I thinking that AWS (the cloud provider) must not allow a VPN connection but I don’t know for sure. Any thought on the matter?

The problem that I run into during the integration of a pair of stereo subwoofers using the Pre-8 is that all drivers’ relative sound levels, for both the main speakers and subs, are measured during the first step. Essentially, you place the measurement mic very close to each driver for left channels then repeat the process for the right. After all the driver measurements are collected for the left channels, graphic results are displayed of the response curve for each driver relative to each other. This in my case includes the left subwoofer sitting about five-feet behind the left main speaker and the same is on the right side as well. This seems to work well for drivers on, or nearly on the same plane, but for the subwoofer that I just took a near-field measurement, five feet (1.5m) behind my main speaker, is not going to have the same relative sound-level when measured at the plane of the main speaker.

In DEQX-Cal, used for legacy DEQX processors, you could match the level of your subwoofer to your main speakers by matching the subwoofer curve amplitude to the main speaker’s curve as measured from the sweet spot, but this was a manual process. You could also match impulse response of your subs to each other then both to the impulse response to your main speakers in order to set the correct delay. It appears now that subwoofer time alignment as well as time alignment of all drivers is performed automatically in step two with a provision to override (in distance).

But how could step two possibly set the subwoofer’s relative loudness when relative levels are set in step one and the subwoofer is substantially on different plane than the main speakers?

For me when I perform driver measurements in step one and close-mic my subs, the final outcome is low bass "lite".  As a kluge, which actually returned fairly good results, I made step-one measurements of the main speakers, in my case at 16" (40cm) then moved the mic along the same plane as I measured my main speakers to where the mic was directly in front of my sub and still on the same plane as my main speaker. This placed the measurement mic about 66" away from the front of the sub and I then took the measurement. I also limited the lower speaker correction (but not room correction) to 200Hz crossing over at 60Hz12dB/oct since the Apogee woofer panels go very low. The bass is pretty damn good now, but I know this couldn’t be best way.

Always open to suggestions.

Actually that sounds like it may be a great idea for measuring the subs and I'm going to be thinking about doing just that, but unfortunately that's not going to work, at least in my case, for a pair of 81-inch high 300lb (each) speakers.  Even a couple of inches to the right/left or forward/back becomes a quite a chore.  Thanks, every idea helps!

Welp, I'm going to throw out what I know may be a terrible idea.  Sorry.

What about taking a cue from ground-plane measurements?  Often used for subwoofers.  The idea is to put the subwoofer up against the floor so there is no reflection point.  The microphone is also nearly at floor level. 

What I am thinking you might want to do instead is to put your planars up against a wall and attempt the same thing? 

Sorry I'm not familiar enough with DEQX to help more.

Here a tip to help minimize aggravation when measuring speakers.  

If it's going to take a bit of time to reposition blankets, wall coverings, sofa cushions, furniture, etc, between measuring your left and right speakers in step 1, consider LOGGING OUT while moving these items to set up for measuring second speaker then log back in just before your start to take measurements again.   What I have found is that after taking so much time to move and reposition these items, the Security Token would expire right in the middle taking driver measurements requiring you to start all over again on that speaker.  This is especially aggravating when you have completed your main speaker and the token expires after you move to its associated remote subwoofer.

After some offline discussions with mijostyn about the difficulties in taking accurate measurements of dipolar speakers such a the Soundlabs that he has and the Apogees that I have, as well as any other dipolar speaker for that matter, I found that his experience and advice in this area proved to be correct and very helpful.

>> YOU MUST BLOCK THE REAR WAVE OF A DIPOLE SPEAKER FROM REFLECTING FROM THE WALL BEHIND IT IN BOTH STEP 1 AND STEP 2. <<. Otherwise, you will get comb filtering which produces all sorts of undesirable side effects. Even though your graphic results may not show what appears to a traditional repeating comb "tooth" pattern over a wide frequency band, you will probably immediately notice a problem in step 2 when the driver distances appear. Your first clue is when you see unrealistic relative distances between the drivers in your main speakers. I have found that these driver distances should be within the hundredths-of-a-meter without manual adjustment. I’ve also found that manual adjustment won’t fix this problem either. If you take these unrealistic defaults you will most likely have very noticeable phase issues. If you manually correct the distances, you may correct the noticeable phase issues, but it will still not sound right. This problem may be because there are driver phase/group-delay correction issues that cannot be corrected by manual relative driver-distance manipulations alone. Remote subwoofers are a different story and may need manual relative distance intervention. I’m still looking into the best way to deal with remote subs. Any suggestions in this area are certainly welcome.

I my case I was able to effectively block the rear sound wave from my Apogees with a heavy sound-absorbent blanket similar to what you would find in under-hood sound insulation in a automobile but of higher density. Please don;t ask exactly what it is or where I got it, it just appeared in garage one day, but any heavy-weight sound absorbent material such as a heavy quilt should work just as well. Just be sure to position the material in such a way in order to cover ALL of the rear of your dipole speaker. As I understand from our conversations, due to the exceptional acoustic transparency of the Solundlabs and probably most other ESL speakers as well, mijostyn informs me that simply blocking the rear of those speakers with a quilt was not enough and for this reason, he is building a Sallie for his unique situation.

Finally, be sure to remove the rear covering before starting step 3 since in this step acoustic properties of your entire room are quantified from your listening position.

@caglioti 

You should be able to tailor your curves with the EQ function. I roll the top end off from 1000 Hz amount depending on the volume. I put a corner at 20 kHz. and sweep the Q all the way over as far as it will. Then all I have to do is move the corner (or very low Q filter) up or down depending on the roll off suitable for the situation. I also EQ the bass up 10 dB at 20 Hz. What are you using for speakers? 

@forrestc 

I was not expecting plug and play at all. However, there are a lot of bugs and idiosyncrasies. The basic functionality is there and you learn your way around the idiosyncrasies. What worries me most is the lack of communication lately. Something happened. Hopefully nobody is ill or hurt. Maybe they are just tired of listening to me b-tch. 

ESLs have an extremely light diaphragm. It is transparent to reflections. You can hold a perfectly normal conversation with someone standing behind the speaker. The microphone picks up everything coming off the back wall and assumes it is coming directly from the loudspeaker. The curves are a beautiful example of comb filtering from 500 Hz up it looks like waves with +3-4 dB peaks.  Anyway, I'm building the solution. Not much above 200 Hz is going to get by this device. 

This is my first DEQX. I used a TacT 2.2X for 20 years. About a year ago it died. The 8 was not available yet. I got a MiniDSP SHD to hold me over. Dirac Live works surprisingly well and for $1500 the SHD is an amazing value, but it's DACs are not up to the quality a top system demands. Benchmark Media systems uses a SHD Studio with their own DACs and are very happy with the results. The SHD is now in a friend's system. The DACs in the 8 are top notch and even in bypass mode it is sonically superior to the SHD. 

Don’t give up too quickly. I know that it can be done. Yes, dipoles do radiate to the rear. Some less than others, but like you state, it is a major factor.

The Pre-8 is my fourth DEQX since 2012 and I have great success not only with with this pair of Apogees but earlier with a pair of Duetta Sigs as well. I’ve always used "conventional" indoor DEQX recommended methods to measure the ribbons. I will admit that there has been a lot of trial error with mic placement, crossover point and slope selection as well as speaker placement and toe-in (or in my case no toe-in). I’ve also had great success with precise subwoofer integration with both sets of Apogees as well.

This Pre-8 was purchased to replace my third DEQX which is a Larry Owens heavily modded HDP-Express II. To be honest, there was nothing at all wrong with the sound. In fact, the sound was drop-dead great. Fantastic imaging, great slam, deep smooth bass, very realistic sounding piano and voices - sometimes I had to wonder why am I messing with something that ain’t broken!

I’m thinking that the fourth generation DEQX will actually go to that next level. It may not initially as the software is still a beta release. I fully believe that the hardware is a SOTA as humanly possible at this date. No other manufacture of ANY audio product of ANY type is at the technological level of DEQX. Yes, you’ll see plenty of new, latest and greatest stuff for sale out there saying that they have the latest technology, promising to take you to the next level and built with unobtanium suspended in pure ether, and of course at a price that you would need a second mortgage in order to purchase. And in the end, it’s all the same stuff but with new lipstick.

I know that this doesn’t apply to anyone on this thread, but most audiophiles I talk to don’t mind spending the money but they want it simple. They just want to plug a few components together, read a review or two about a power cord or such, buy it, two minutes to install it, a week to break it in, and life is good.

If anyone who has electrostatic, planer magnetic, ribbon, MBL, or other non-conventional speakers AND DEQX, it’s never going to be plug-n-play. DEQX wants to make it easier for their customer and they gone a long way to make that happen. I’m sure that many potential DEQX customers have been scared off in the past by its level of hands-on technological prowess required. On the other hand, they can’t abandon their nuts-and-bolts type customers like us as well. I doubt many any beta tester customers expected plug and play.

The knowledge may be proprietary, but at this point, I for one would like to know what, at least in general, goes on in each step of the cloud processing. A block diagram would be great. That way we look at it and say, "ah-ha" that where _____ happens. Right now it’s a black box. Data in, data out. If it doesn’t come out right, change a parameter and try again.

OK, I’m done.

ForrestC in Tallahassee

@forrestc 

The problem comes with the first step, defining the drivers. In this mode the DEQX assumes everything is coming from the driver. It uses this measurement to determine what is room effect and what is the driver. Getting a near anechoic measurement on a dipole is next to impossible unless you drag them outside. Remember, dipoles rediate in a figure 8 fashion and sharply beam vertically. You only have to worry about the rear wall and reflections coming from it.  I tried blocking the rear wave with a quilt draped over a tall tripod. It worked up to about 1000 Hz then the comb filtering resumed. I am going to build a 24" wide "Sallie" 7 feet tall which should do the trick. If I can't get a decent measurement then I'm afraid that will end my relationship with DEQX. 

mijostyn, Im glad that you pointed that out to me. You got me thinking. Even though I used lots of very sound absorbent padding on the floor, walls, etc, and I’m sure that the cloud processing uses software algorithms to minimize first reflection interactions, I’m still about six feet from the side walls and just under four feet from the floor with the mic. I just now took another very close look at the impulse response in Lin X mode and i noticed some rising "fuzz" on the decaying curve starting at about 5.5mS. That’s probably more realistic. Later tonight, I’ll go back and run it through to the end and take a listen. I appreciate the heads up!

Oh no, I didn’t want to hear that about the Volumio installation. I was hoping to use the Softsqueeze plugin in order to use LMS (formally known as Logitech Media Server) directly to Volumio. Right now, I’m using LMS with the UPnP/DLNA Bridge plug in (on the LMS side) developed by an LMS community member.. It works but making a connection isn’t always smooth when connecting to Volumio. Sometimes you need to mess around with it to get it working.

@forrestc I have a problem with volumio. I reached out to Volumio and they referred me back to DEQX who have not been helpful either. I use Volumio to plug in Qobuz which works everywhere in my house except on the Pre 8. I went to up date Volumio as I was behind in versions. The unit crashed during the update. I reset it to factory settings and now other than the volume control the front panel does not work. I operate the system from my computer. There is a set up procedure for Volumio and I assume until this is done the Volumio attached functions will not work.

You are right staying back from the midrange and tweeter ribbons. Tri amping that speaker is difficult because the impedance of the individual ribbons is close to zero. My Divas had a huge back of resistors to compensate. 

10 ms is 10 feet. You are 10 feet from the rear and side wall? 

I see the same problem at the crossover points. I always check the results with an independent USB measurement system and there is a dip in the response at the high frequency crossover points. This only occurs at the midrange/tweeter crossover. The subwoofer cross seems normal although the delays on the subwoofer are OTL (Out to Lunch). 

I also have a phase problem. Usually, one of the high frequency channels will be 180 degrees out of phase, but it can happen with a lower frequency channel. I set the microphone positions with a tape measure. I have no idea why this is happening. 

I have not used any previous DEQX models. I used a Tact 2.2X for 20 years and it's correction system was faultless. It also offered much more flexibility, but the unit finally expired and the company is out of business. I looked at the Trinnov Amethyst, but I did not think it's bass management was flexible enough. 

Hi mijostyn, Alan replied back to an mail I sent may 30 on June 3. I figured that he must be swamped between further development and answering user questions. I questioned him because was perplexed at the final response curve that I was seeing. I saw significant dips in the final response curve at the crossover points and moreover, it just didn’t sound right. I also mentioned few items relating to the legacy DEQX-Cal v3 that they may want to consider to eventually integrating into the software. Alan assured me that the response curve we are now seeing with dips at the crossover point(s) is not accurate and in reality those dips do not exist in the actual response. He went on to say the they are presently working to correct that issue.

FYI. I get my best measurements at about 16" (40cm) from the ribbons. For me this yields right at 10mS to the first reflection on each speaker and I set the Trim in the Chirp Config there. I really cannot go closer with the mic since the mid-range to tweeter ribbons are passively crossed over and I’m happy at 10mS. I also set the lower limit of correction to 200Hz and not 20Hz as is the factory default. It probably wouldn’t hurt to go as high as 400 Hz I’m thinking, and I’m sure that I’ll try it eventually.

In the meantime, I did work out the issue of bad sound. The issue turned out to be a slight to sever phase difference between the left and right speaker at different frequencies. I ended up playing with the driver distance relationships under the "Create a Speaker" tab, starting by equalizing the distance left to right of the non-zero drivers. It’s probably not perfect, but it now sounds really good now. I have not integrated the subs as of yet. I’m trying to work out as many of the issues with just the main speaker first before adding a whole new mix of issues surrounding sub integration since it is done way differently than in DEQX-Cal.

Of yes, the Apogees I’m using are the original "The Apogee" which was their very first model with all steel frames and weigh about 300 lbs each.

http://www.reality-audio.com/full_range.html

Oh, one more thing. Does anyone know how to do a Volumio Plug-in on the Pre-8?

I opened a ticket with Volumio  about two weeks ago but it's still in the "pending" state.

Hi Fellows,

I was in the first wave. There was a plan to set up a forum for users. Have either one of you tried to contact DEQX lately. They have not answered any of my inquiries for three weeks. I'm a bit worried that COVID took to much of a financial toll on them. 

I have a difficult system for the Pre 8. I use dipole ESLs and the comb filtering they produce confuses it. I'm working on a solution to absorb most of the back wave so the 8 can get a measurement on it.

@caglioti you are right. we need more flexibility, manual control of delays and separate EQ for each channel. Right now I have set the crossovers and run it in manual mode, EQing manually. The imaging is not yet what it should be. Rome was not built in a day.
 

@forrestc Which Apogees? I use to have Divas. You have the same basic problem I have, Dipoles, perhaps not as bad as mine as the ribbon is not as transparent to the rear wall reflection. Place the microphone as close as you can to the ribbons, within and inch!  Look at the response curve. If you see it bouncing rhythmically along the x axis you are picking up the comb filtering.  If you go to the Sound Labs web site check out the "Sallie" rear wave absorption panel. I am building a 24" wide version with 8" wedges instead of the stock 6" wedges. If you are not comb filtering then it is a mute issue and you should be able to get a decent measurement.  

Thanks for responding. I suggested to DEQX that they start a forum for users to communicate experiences and suggestions, but we’ll see.

i made some profiles and got it up and functioning. It’s very transparent. I plugged REW into it and got some strange measurements though, vey confusing. 
I’ve been using a manual DSP, Xilica, so I’m used to setting driver timing, crossover slopes, and PEQ’s, so I’m missing the ability to control more parameter.  
mic placement seems important.  
I know it’s still in beta, but I’m hoping for increased ability for user tweaking.  
I’m running a usb cable from my laptop into usb input to run REW through DEQX and I’m getting the REW sweep sounds, so I think my measurements are accurate, but like I say, the results are not what I’m used to.

Typically I like a 8db or so slope from 100-15khz, not sure how to achieve that yet. 
the overall sound is very good, but I’m trying tailor the sound more to my liking, and the unit seems so focused on automation I’m having trouble. I’m optimistic about it though, wondering what the final software release will change.

Hi Ted,

Like me, you must be on the second wave.  Mine showed up a little less than a month ago and it been a tedious process.  I have been told by someone on the first beta wave that we're fortunate to be on the second wave since some of significant the first wave issues have now been worked out.  As you have probably figured out by now the beta UI is stark and not all that friendly, but as I've been assured the final consumer version will be much improved over what we are seeing now.

Have you made your first correction profile yet?  The videos made by Matt at DEQX show the essential basics. 

I've made a few profiles so far but I'm not there yet.

I hope to see more flexibility in the web-based cloud software in the future.  I'm so use to what you can do in the old Windows based DEQX-Cal V3, but Alan told me  "to rethink how the software now works and it has not relationship to the original DEQX-Cal software".  That was good advice.  Yes, it is fundamentally different than DEQX-Cal.

DEQX Pre-8 to using 6 channels for The Apogee original Full Range fully rebuilt and upgraded, powered by 2x H2O S250 SE with two subs.