Super! Freshen it up a bit, and the next generation of REW Roon users will be on the way. =)
For my part, REW made quite a difference to the sound in the very lived in livingroom, and Roon made quite the difference in how we listen to the music, an most excellent combination.
Maybe I should rewrite it to use sweeps and average them instead, that would make it easier to use homeaudiofidelity.com in combination with the guide. For example, the first 5 steps could be same for both REW and HAF, and then step 6 and futher would be if you want REW to do the actual corrections.
I have also discovered some new stuff, like how you should stay away from big Q values, especially at higher frequencies, since those corrections don’t look the same in different sample rates. Generally speaking, I have got best result with manual corrections using as few filters as possible, even if they look less flat in REW. Manually correcting against a “Psychoacoustic” filter is not a bad way to go, but more complex to perform.
Here is my current correction, using 10 filters and no Q higher than 4, showing with psychoacoustic filter:
With var smoothing, its not at flat as if you let REW do the auto-correction, but sound better:
And don’t laugh now, my room is horrible (3.5x3.5 meter and concrete walls, floor and celing). Even with lots of acoustic treatment (2 cubic meter stonewool) its horrible, but here is the uncorrected measurement:
Seems like you got REW to do exactly as it should, and make your room better.
I really like the moving mic RTA pink noise measurements. Why not add sweeps, sure, but keep the option of RTA, because from what I’ve read since using your guide, it seems that it’s a pretty big pain to use sweeps when you want to correct a large area - the entire center of your room, for example.
Not having to use a microphone stand, trying to measure the absolute exact same area for left and right channel, is a huge deal. Using moving mice for 40 or 60 seconds in a slow large spiral from sitting ear-level to standing ear-level is great.
Hi guys,
Not sure if I should post here first, as before trying to make a filter I am just trying to use at best my acoustic panels.
I have two AStri diffusion panels on front wall (actually a large window) in between the speakers. Two AStri bass trap on top of Ikea Kallax shelves both left and right on the front wall. Two GIK bass trap with diffusion panel on left and right walls between the Ikea Kallax and the speakers.
At first reflection point I have two absorption panels on the left from AStri Audio. There is a large window on the left wall, this is where I have put the panels. At first reflection point I have one absorption panel on the right from AStri Audio.
Behind the listening point I have a difusion panel from AStri Audio. Left corner on the back I have a GIK bass trap. Right corner on the back I have a AStri bass trap.
Room is 380x370cm (12.5 ft * 12.1 ft).
Speakers are 137 cm/4.5ft from back wall (not accounting the Kallax with comic books or vinyls) and 72cm/2.4ft from side walls/window. Speakers are 242cm/8ft apart. Measures are taken from center of the speakers front panels.
Distance from listening position is 220cm/7.2ft.
Listening position is 80cm/2.6ft from back wall but actually 60cm/2ft from back diffusion panel because there are also CD shelves on the back wall.
Now to the point. I have measured the responses left and right of my speakers with all the panels in place and without any of them.
This is what I am getting.
I am seeing some differences on the left readings probably due to the interaction with the window…but on the right channel really I am not seeing any differences! At least as new to this, this is what I am reading!
Could you help me out here on what should be the next step for me please?
Should I zoom in the graphs or there is no differences anyway?
I had set up the volume at 75db on the SPL meter but strangely the graphs show a lower average volume around 50dB, so I may have done something wrong!!! Having had a second look at RTA preferences, I did not had “Adjust RTA levels” ticked. Should I re-do the measurements?
RTA analysis is not the right tool to assess the actual performance of acoustic treatment. RTA is good for the overall assessment of the tonal balance of the speaker in the room, as well as identifying the main resonances / anti resonances.
You need to run sweep measurements and evaluate:
There’s not much to be seen on the SPL. Even good bass traps rarely reduce the amplitude of room peaks by more than 2-3dB : this is OK, what’s more important is their ability to dramatically reduce the decay at the problem frequencies.
A very basic DSP question re Anti-Mode 2.0 Dual Core:
I just installed and ran Anti-Mode 2.0 sweeps on my 2.2 system. Running digital (Fidelizer Nimtra Signature Roon Core > AM 2.0 (USB); AM 2.0 > Modwright/Oppo Sonica DAC (TosLink).
Improvement to bass and mids not subtle at all.
Question: It seems the highs are a bit hotter with the new settings. (Speakers are Harbeth m.30s/subs HUS ULS-15 MK2s). Since AM 2.0 only goes to 500Hz, it seems odd that upper frequencies should sound different.
That said, how would I bring down the upper freq’s in Roon?
Roon’s Parametric EQ Roon Band presets are
100 Hz Low Shelf
200 Hz Peak/Dip
2000 Hz Peak/Dip
6000 Hz High Shelf
In-active Band presets:
20 Hz High Pass
20 KHz Low Passc
Would 6000 Hz Band be the target?
Would adding an additional Band, say 8KHz or higher be advised?
What parameters for adjustment would you suggest? IOW, understanding that my ear will be the final judge (not interested at this point in REW, etc. as they seem far beyond my technical abilites) by what dB increments should Gain attenuation be adjusted?
What does Q setting mean?
Thank you for indulging a newbie here.
Try experimenting some, for example try a high shelf at 10khz with gain -3 and Q 0.3
As @alec_eiffel said, acoustic treatment is not very visible in frequency response, but it would be more visible in for example a waterfall graph (but to get that you need sweeps and not RTA). Acoustic treatment also makes it easier to get EQing to give a good result, so try following the guide and see it it improves the sound.
They are at 0dB so they don’t do anything, so might as well delete them.
Hi @alec_eiffel, @Magnus,
I have followed your recommendations and also read through the “pdao” threads on Forum-Hifi for which @alec_eiffel had put a link in that topic before.
Most importantly I have been able to convince my wife to open the large window doors between the Home Theater room and the Hi-Fi room. The opening is 44% of the surface. It is dead center. The volume of the room has been multiplied by 2.4! It is now 92m3: L925cm, l375cm and H265.
I have made following measurements. Please let me know your comments. Especially if I have done something wrong in the measrurements or using the templates from “pdao”. The sweep I have used are the one provided by HAF for left and right channels, as I am intending to use their services later on.
All the acoustic panels I have mentionned previously are installed. There is no acoustic panels in the Home Theater part of the room (yet!
)
Wow! Just stumbled across this thread as I was about to fire up REW and do this for myself, thanks so much to everyone who has contributed so far!
I’m moving to a more minimalist setup with my NUC Roon Core -> RPI/DigiOne endpoint running into a Devialet Expert 200 and want to get some of the correction benefits I see in my current 2.1 setup which has a Marantz AVR running Audessy in the mix.
Step 4 - should generated be set to left, right or both?
Step 7 - save current values, peak values or both current and peak values?
Step 4: it don’t matter much, but use both. This is just a guidline to get a decent volume, 70-80dB seems to work good (one channel will be 3dB lower than both)
Step 7: save current (when the guide was written there was only one “Save”)
Thanks Magnus. Might be worth editing the instructions at the top of the thread to avoid confusion (or maybe it’s just me) 
The guide could use with an update, some new features in REW and changed GUI does confuse a little bit.
Hello and thanks to you Magnus for this guide (but also to all the other contributors of this thread for their valuable inputs). I have had the miniDSP microphone purchased as well as REW installed a couple of years back but only recently went for a Roon subscription so it was good time to use them
During the past week I have tried to follow the instructions. Some first impressions:
Overall it seems to me that this is on the one hand a very promising but (at least with my current results) not really acceptable solution.
Whilst I get a much more controlled behaviour in some horribly sounding bass heavy tracks I seem to be gettting “truncated” (for a lack of better word) bass notes on some others which sounds very unnatural.
The other thing that was unexpected for me is that voices (especially female) seem to get much “drier”, like the difference between a “church” sound and an anechoic chamber sound. This somehow persisted even when I limited the correction to the 20-400 Hz range. I am puzzled as to whether this is a correction (i.e. closer to the “right” sound) or a coloration.
I will need to repeat the measurements using the “Dirac box” concept.
A couple of questions:
in order to avoid meddling around with cables I downloaded the PinkNoise from REW into a file and using Audacity created Left and Right channel only copies which I use for my measurements. Do you think there is any problem with this approach?
whilst switching between Convolution “on” and “off” in Roon I noticed that when I play an MQA file witn the convolution DSP “on” my DAC recognizes is as a plain PCM file. I guess this is one drawback of the convolution filter, right?
thanks in advance,
Greg
If done correctly, digital room correction will result in a more neutral sound with very little coloration. However, its not always easy to get it correct. Also, coloration tends to sound “neutral” after a while, and when removed you miss it.
Make sure you have “Headroom Management” on in DSP with “Show clipping indicator” enabled. If you get clipping, you need to lower volume (I use automatic volume levelling but a fixed decrease also works).
You can DSP -> Procedural EQ -> Mute Channels to get left/right only sound from Roon.
I don’t have a MQA DAC myself, but I think you can make Roon “put back” the MQA after DSP, but the second part of MQA is only an up-sample so it won’t make a big difference. Make sure you let Roon handle the first MQA unfold though, since that needs to be done before any DSP (use “Renderer only” in device setup for MQA Capabilities).
Are you sure about that as MQA call it rendering and suggest more is going on?
