With convolution you should be able to go as far down as you want at the low end.
Providing the processing engine is working to that rangeā¦
What I meant is, Roon should not be imposing an artificial limit. Convolution is a well defined mathematical operation. But yes, the rest of the hardware that come after (DAC, Amp etc) could be a limitation. For what it is worth, I EQ headphones down to 10 - 15Hz.
@KMan is right. Convolution is a mathematical constructāthere isnāt a way for us to make it not work in a certain range without changing it into something that we couldnāt call āconvolutionā anymore.
With the Parametric Equalizer, itās differentā¦it could work down there, but we donāt currently allow for parameters to be set that low.
With the logarithmic scale on the X axis, that 10-20 range takes up a decent chunk of space. I think weād need to hear more feedback from more people before spending the real-estate on the screen that way. In the mean time, weāll make it so you can type in frequency numbers down to 10Hzāthat way the adjustments are still possible for the people who want to play around down there.
Perhaps a zoom and slider control like the one used in DMG? let it default to 20Hz, but can be zoomed out to 10Hz?
Yes, in the mean time letting us enter values down to 10Hz would be great.
1 Like
Pan and zoom would be great anyway.
So would using more screen space for graph and less space for the EQ list.
Havenāt read up on REW yet. Two questions:
- Some systems, like Dirac, talk about working in the time domain, not just frequency. Dies REW?
- I understand REW can export PEQ filters. Does it also export convolution specs?
I donāt think REW does time-domain stuff but I could be wrong.
Yes. They export as a single wav fileāthatās all you need for room correction. Much faster than trying to transfer over PEQ stuff, especially if you do any iterating/tweaking on the generated filters.
You can also export one per sample rate and stuff them all into a zip file together to avoid having the filters resampled in Roon. This is a common practiceā¦just avoids one less thing mucking with the signal.
I think @joel did some manual transfers of measurements out of his Meridian MRC room correction config into our PEQ and found that he got really nice results from it. Not sure exactly what he did, but I know you have some Meridian gearā¦so it might be another avenue to consider if it makes sense.
This already would be wonderful. Thanks a lot!
Greetings,
Sander
Thanks. Iāll look into REW.
I never did RC in Meridian. Kinda clumsy in the G-series and the 818 doesnāt do it. There were rumors that they would add it, integrated with MQA, but so far nothing.
@brian - think you are correct about REW not doing time-domain stuff. I havenāt done it yet (need to install Virtual Box to enable me to run Windows apps on Macbook) but I think that output from REW can be imported into rePhase and used to generate FIR filters. Another possibility is FIR designer (which is available for OSX but is not freeware).
Exporting WAV file and using in Roon DSP convolution is, as @brian says, much easier than typing coefficients into parametric equaliser.
REW has mac builds.
Thatās what Iām using - it is rePhase that doesnāt have OSX build.
REW on its own os worthwhile. Images below show before and after using REW to adjust from 200Hz to 10KHz. Solves a major issue in the 1k-2k region and a noticeable improvement in listening experience when this is applied in DSP convolution.
Area below 200Hz might be better solved by treating room rather than with DSP (correction shown uses up all 20 of the filters available in in REW) - although Roon could be used to apply a couple of parametric filters before or after applying Convolution using the WAV file exported from REW.
@brian - Please pass on thanks to all at Roon who got DSP included in 1.3 - yes, there can be further improvements but it is a great step forward.
1 Like
I have now installed REW and a UMIQ-1, and read the manual, and Iām experimenting.
My measurements are occasionally disturbed because the parrot likes the 20 - 20,000 sweeps and joins in.
16 Likes
After 2 days of dabbling with REW and Roon this is my result. Yellowish curve is both the speakers without convolution, red are both the speaker with applied convolution. It seems Roon is doing the right thing. so it is up to me to come up with a decent target curveā¦
1 Like
What we need here is real time squawk detection coupled with beamed active cancellation anti-squawk dps to cancel squawks at the listeners ear. We may need a wi-fi mike in the cage to provide sufficient lead time. Will your parrot promise not to eat it ?
2 Likes
Thanks for sharing. It looks like you have managed to sort out most of the major anomalies.
How does it sound now?
What should an āidealā plot look like? (If there is a rough unerversally accepted benchmark for this)
Iāve had REW for a long time, but before Roon 1.3 I was not motivated enough to dive into it. Iāll be ordering my Mic now if I canāt find the calibration file for my old trusty SPL meter (Check Mate CM140).
Still dreaming of Dirac-like capabilities, but Iām happy with exploring the benefits one step at a time.
I love the idea that I can measure and apply room correction filters for each zone!
I am going through a selection of tracks playing at spot the difference. I would say that the sound seems a tad bit more in focus, more defined. Doing A/B comparison with āGoing Homeā of Leonoard Cohen, the voice seems to become clearer, probably due to some ringing happening before at the lower frequencies. Not a day to night difference but if you listen carefully it is there. My wife said she couldnāt hear difference but my younger daughter was able to spot it straight away Her comment when I asked if she was hearing a difference was āyep, the voice is clearerā.
For being the first time I use REW and I dive in the convolution world I would say I am pretty satisfied but after 2 days of study I feel I have just scratched the surface.
Please, can someone add a how to REW-convolution-Roon?
1 Like