TOPIC:

Hey!

I got my DDRC-22D a couple of days ago and I have finally gotten around to testing it out. I first noticed that without any filters active there is a channel imbalance, right channel is louder than the left so the stereo image is shifted towards the left. With filters on that is corrected though and stereo image is centered.

The biggest problem however is that it kills my bass.



There you can see the channel imbalance and also the predicted response. I chose the room feel eq here in this pic but I also tried the Harman curve.

To confirm what I was hearing I did a measurement with REW.



The purple line is response without filters, both channels at the same time. Green line is with the Harman curve and the yellow is with the "Room feel" curve. Below 50Hz I get a good improvement but between 60-100Hz it totally kills the bass and it doesn't do much else in the rest of the frequency band either. I think I could probably get better results with my little 2x4 unit and REW eq.

Why does my DDRC-22D behave like this? The speakers are DIY floorstanders with 2x Morel CAW638 and a CAT408 in MTM configuration in a vented box. DAC is a NAD D1050 and amp is a S.A.T Amplifix.

Measurements were done with the UMIK-1 and a 90-degree calibration file and mic was pointed up towards the ceiling. The sofa (and sweet spot) is right next to the back wall.

Hi Rullknufs

I assume your REW measurements are done in the sweet spot only, right?

Dirac corrects the response "sort of averaged over the measurement positions". Try measuring approximately the same positions as measured with Dirac and use the average function in REW.

Best regards

PK

Yes, the REW measurement was done only in sweetspot.

I had time today to do some new measurements. I began with trying out the "Chair" measurement preset in the Dirac software. Then I measured with/without Dirac in sweetspot and saw the same trend as before. Tried to do measurements in approximately the same positions in REW as in Dirac and did an average and ended up with this result. Green is with harman curve EQ in Dirac, averaged over 9 measurements at and around sweetspot. Red is sweetspot only but without Dirac.



All the measurements with Dirac had the same dip in the 60-100hz region. What is happening here that causes Dirac to go nuts? I mean, it seems to work perfectly fine below 50Hz and it seems to work for the mids as well.

I don't know if I shall bother experimenting any more with this now since I'll be moving to a house in 1.5-2 months time. Maybe it has something to do with my current room?

@Rullknufs

This is indeed very odd.. I wonder if it's something related to the measurement process. Few comments:
a) Please perform a measurement with the mic pointing at the speaker (not 90deg.)
b) You're making the 9 average measurements. Correct? How far away are you making the measurements?
c) I'm curious if it's worth to "not" boost the speaker outside their real performance. See how your natural response is diving at 30Hz on the blue measurement? That is what Dirac measured. You then applied a very heavy/tough target/boost to achieve with the target curve trying to get down to 20Hz. In room correction world, it's a big NO NO as you're basically trying to get your speakers to re-invent the laws of physics... Dirac can do some magic, but to some "extend". Per our experience, that 10dB boost is way too big.. Please move your target to follow the natural roll off of the speaker.
d) Please provide more information along with sharing your Dirac Live project. It would be telling to see what's behind the scenes here.

I finally had time today to do some more testing. Did new measurements with mic pointed at speakers. Also, got a shorter signal path now that I have bought a longer USB cable. Before it was Macbook Air -> Samsung smart tv via HDMI -> DDRC22. Now it's PC -> DDRC22 via optical cable.

The measurements for the sofa preset are taken across the whole sofa, which is approximately 2 meters wide. Height vaires maybe 50cm and depth maybe 50cm as well. For the chair preset (which I tried today) measurements are centered around the center of the sofa, with measurements taken around 30-40cm from sweetspot.

I tried today to move the window to just above 30Hz to avoid that boost. Didn't seem to help much.




Also, I don't know how to share my project since it says it is too large to be uploaded here. But maybe this google drive link works? drive.google.com/file/d/1K3f52YUWXQu8NJV...6jm/view?usp=sharing

@Rullknufs
Sorry for missing this thread but here are few comments on your project.
- In your current system, you have an obvious bump at 30~40Hz. That's where I think you're enjoying your bass..
- With the project received, we noticed that the target curve was almost flat and therefore you're loosing all that bass as the target curve is really not trying to follow the current response of the speaker.
Here is what it was.



Since you're enjoying bass, you'd want to keep that bump here in your system as it's a large one (more than 15dB boost). You must be having some sub in that system? Is there anyway to delay it to correctly timealign it to your system?

Here is the target curve you could try maybe for that purpose?



Do you understand the gist? see how we're trying to follow the current response of your system especially at the bass section?

If it's not doing the trick, you might have to look at correctly time aligning that sub.. (or moving it around). In a proper 2.1 setup, you'd want to a) timealign the sub to the main b) Correct the overall.

DevTeam

Sorry for such a late reply, been busy traveling and moving from the apartment to a house.

Did a quick initial setup a month ago in my new living room in the house but it was all done in a hurry so not perfect. Today I setup the speakers correctly in the room and aligned them symmetricaly with the sweet spot and made new measurements and some new target profiles.

It does indeed sound better with the DSP on compared to having it off, but the biggest change is in the mids and highs. In the bass there's suprisingly little improvement. I'm used to seeing a much bigger improvement in the bass with using the standard MiniDSP 2x4 before.

There's also a really annoying thing happening in the bass at louder volumes. It's like the speakers make a farting sound when certain bass frequencies are played a loud volumes. And no, it's not the drivers running out of excursion. It is most clearly audible when playing songs with clear basslines with not much else going on at the same time. Some examples are Tyga - Rack City orTrentemøller - Chameleon. It is not at the very lowest frequencies. I tried to find the problematic frequencies using the sine wave generator in REW but I couldn't get that farting sound to happen. If I bypass the DSP I can play loud again without that farting sound. Some hiphop/electronics songs play perfectly fine even at those loud volumes that give a farting noise in certain other songs.

I don't know if it's possible to see the filters in the Dirac program but I'm suspecting there's a narrow filter boosting some frequencies super-hard causing something to clip making that farting sound. Right now I'm using the default response curve and limiting the lower frequency limit to ~30Hz.

Also, there's no sub involved in my setup. Just two floorstanders with 2x Morel CAW638 6.5" woofers in a MTM-configuration in a vented cabinet. They are quite bass-capable for their size.

@ Rullknufs

We can't repeat it enough but the design of a target curve is KEY. It's an affair of taste to "match" your expectation (E.g your test #1 where you load a flat target curve = flat target cuve). i.e. you'll have to play with your target curve a bit more to get it to what you like to listen.

As for the dip in REW at 50Hz, I think that it's the natural response of your speakers/room somehow.. See how Dirac live measures that dip at 50~50Hz (blue lines) "before" it tries to do anything? Maybe worth investigating where that come from... Once you linearize your response at low frequency (i.e. less peak), you end up with less bass but dip is still here because it's something else..

DevTeam