@Magnus Hi, tried to read all of the post and apply your suggested settings, but it seems original guide (at the beginning of the post) is based on older version of the software. Is there any update available? Tx a lot!!!
@Magnus , Iām also struggling to interpret your original post with reference to the latest version of REW. If you have any suggestions as to anything that needs to be changed for the latest version Iād be grateful.
Yea I know, I should update it with new screenshots (everything else is pretty much the same though).
Stupid question, but how do I limit the number of filters applied, letās say to 10 max? When I manually disable some filters after the correction, it messes up the whole thingā¦
I get this error message when trying to import the filters to wav:
āThe selected filter set has no filters that have an effect. Nothing to export.ā
Any tips?
I have also got that but I donāt remember the situation. Make sure you select the correct export though, its easy to mix them up (it should be File ā Export ā Export filter impulse response as wav), and that you have performed the āMatch response to targetā functionality in the EQ window.
Thanks, Iāve made sure that these are correct. Iāll try to take new measurements this evening. My friend doesnāt have this problem either so I have probably just mixed something.
I doubt you need to make a new measurement, but you have to press the āMatch response to targetā button in EQ window ā Filter tasks. This will generate filters to compensate for your tops and dips in the measurement.
Yes Iāve done that. I manually adjusted the filters also and only want to take care of the worst peaks in the bass region. So between 20-100hz or maybe 20-200hz max.
For now Iāve just used Roonās PEQ but I want to try the convolution so I can make separate corrections for both channels.
Hey now I think I got it. I had General EQ selected when I ran the target matching. I guess it needs to be rephase like in your screenshot
Iāll try again this evening when I get home.
That wonāt matter, and unless you want to use rePhase afterward (which is more complicated) you should use General EQ.
I did new measurements but still canāt export and Iām getting the same error message. I was earlier doing two measurements per ear (L and R separately) and then taking average of these. This was wrong. Now I did it with one single measurement moving the UMIK-1 around both ears with the same record interval. Obviously one measurement for L and one for R speaker.
After making the filter adjustments, shouldnāt I press the āGenerate measurements from predictedā to get the filtered curve to main window and use that for the exporting? I tried this and also āGenerate measurements from filtersā and neither one works for exporting. Thereās nothing about this part in your instructions. How to get to the part when I export the filters to wav.
Also the current version has different saving options in RTA window. I can save current, peak or both. Which is correct? Thanks for all the help.
EDIT:
Wait, is the EQ/filter information automatically saved to the MMM measurement data I have already in the main window? Iām not supposed to generate anything from the EQ window? So I export the original measurements after matching the response to target in EQ window?
EDIT2:
I tried ths now and I managed to export the wav files and import them to Roon. I think it works since it kills the peaks like I adjusted it to do. Also the difference to regular PEQ is clearly audible. Convolution sounds better.
EDIT3:
Damn, this really sounds excellent. Clearly better than PEQ which I used earlier for two small dip corrections. Its quick to AB test these just by enabling PEQ and disabling Convo from Roon DSP and vice versa. Much better PRaT and clarity with Convolution. I guess it has its advantages to have separate filters for both channels.
Have never messed with room correction so this thread has me very interested. I got the umik-1, the REW software on my mac, fired it up - and it looks different than the guide at the top of this thread. Is there an updated guide based on the current version of REW by any chance?
Thanks.
āPeter
Sorry no, I should update the guide with new screen shots. But the actual steps is still the same, itās just that some parts of REW looks a little different.
There are few settings which are missing from the guide but it still works. For example in signal generator window the pink periodic noise has possibility to choose āminimize crest factorā option. I have no idea what this does but I got similar results with it on or off.
A good idea is probably to use default value (I donāt know either what that is).
OK. A few questions / ideas then:
āPink PNā ā "Noise ā āPink randomā? full range
In the RTA I donāt see where to set āUpdate interval 1 and 0 for peak hold and peak decay.ā
In RTA do you do ācurrentā āpeakā or ābothā?
How long do you run RTA for? Your screenshot shows ā22 averagesā which I assume is not a setting but the RTA showing something. When I run it I donāt see anything about # of averages displayed?
In the EQ window I donāt see where to set āvar smoothingā?
Thanks!
āPeter
Pink PN means Pink Periodic Noise, and it should be full range. The peak hold/decay has moved to its own window but I am fairly sure the defaults are already the correct 1/0.
Peak or current probably donāt matter much (they typically seem to be very close but peak a few dB higher), but use current. The length is up to you, I typically run around 30 seconds on each side.
Var smoothing is set on the drop down from the cogwheel:
Thanks much Magnus.
I think Iāve run REW correctly. I seem to have very flat response in my room?
Probably doesnāt need correcting?
Or did I do something wrong / does something look wrong?
Left and right both look like this:
Much appreciated!
It looks like itās a very low volume, if you select the Limits button up to the right and set 20 - 20000 for frequency and 30 - 100 for dB (SPL), you will get a better look.
When you play the pink noise it should be around 75dB, which is a reasonable loud volume (loud enough so speaking to someone else will be hard).