This is a thread for you to show me (and others I presume) how to measure my room and input the values into Roon.
I need baby steps please,
With no -TRP17 phase equalised anti- magnetic toroidal microphones preferably with negative deconstruction algorithms.
Or - phrases like , running from command line.
I would hope there is a simple enough way of getting these measurements for input into Roon, so how are Roonies doing this?
It would be good if we could get maybe 2 or 3 generally accepted ways of doing this?
First I’d suggest getting a umik-1 USB mic, (about £80) and a long USB extension cable (I got mine from Amazon).
This will enable you to use free and paid-for measurement tools and filter creation.
If someone can write a simple guide, I’d be glad to read it! Sadly, apart from Dirac - who have made the task about as simple and effective as it can be for a new user - the tools out there and processes required to do room correction properly with convolution are far from simple.
I’m no expert but in simple terms you’ve got two obvious options after measuring:
Use parametric EQ to even out the peaks and troughs and get a more balanced sound.
Generate full convolution filter which will do the above, but also take care of phase alignments, and reverberations, giving you an incredibly clear sound (hopefully) without adding any unwanted artefacts.
There’s a lot to read and learn. Hopefully someone more experienced can break it down into simple steps. I fear I would fail this task, but am happy to offer any advice based on what I’ve tried.
REW is free and (in my limited knowledge culled from web searching) quite well regarded.
I tried Dirac trial - was very impressed but decided that, with DSP coming in Roon, that I couldn’t justify the cost of Dirac.
What we need in Roon (@support) is the ability to take an input from REW and send it to an endpoint - this would make it a massively easier task to implement room correction.
Have you compared Acourate (340 euros) with REW (free)? It would be a significant cost for me - would have to add the cost of something like parallels for OSX (Acourate is windows only) - so would value any thoughts on the comparison.
Actually I didn’t compare both. All I can say is that room correction made the biggest improvement to the sound I had ever before, this was more significant than any dac or for example a linear power supply. Acourate comes at a price, but it is well worth the invest compared how much you can pay for other things that have a minor effect to thesound!
Acourate is the the gold standard in room correction and highly regarded, so I went this way.
But at the end all that is important is that you get a good filter for your room!
You can pay much money for Dacs, amps and speakers, but this is only one part in the equation and a very significant part that many don’t think about is the room you play your music in.
And this can be adressed by either building a optimized hearing room that few have or by using the best room correction filter and software!
In my experience this is well worth the effort and the invest!
[quote=“Mike_Treadaway, post:4, topic:19881”]
What we need in Roon (@support) is the ability to take an input from REW and send it to an endpoint - this would make it a massively easier task to implement room correction.
[/quote]that’s already there if your running Roon on the same PC or MAC you do the measurements on.
[quote=“Peter_Lobner, post:7, topic:19881”]
In my experience this is well worth the effort and the invest[/quote]
DRC is great, however, given a choice I’d do room treatment first and then DRC for fine-tuning, particularly LF. One thing DRC cannot do anything about is nulls.
"
I might be wrong - and hope that somebody can tell me that I am - but I thought that Roon could not accept input from an external source i.e. Roon can only play things in the library or from Tidal or from Internet Radio.
In this thread (below), coefficients from REW are fed in to Roon’s Parametric Equalizer. An alternative is to export file to use in Roon’s Convolver. Seems to me that this involves using REW to play the test signal by connecting computer running REW to speakers, generating coefficients in REW and then inputting / importing them into Roon.
What I would like to be able to do is to (a) Play REW audio sweep and test tones through Roon so that, after generating filters, their impact can be measured using REW.
Of course basic room treatment is a good start. Do you own a LEDE-room? If you don’t as I suppose then good room correction filters always helps on improving the sound!
The filter I created with Acourate has over 131000 filter taps, this is way more accurate than anything you can build manually using a simple parametric eq! Overmore this parametric eq doesn’t compensate for Time Domain alterations as a good convolution filter does!
So a simple parametric eq can’t be compared to a way more sophisticated convolution approach!
You need a calibrated microphone, a Audio interface (good ADC/DAC) to connect with your speaker system and a good Software tool ( like Acourate) for creating the filter (and of course some knowledge how to work with the software to achieve good results -> See my links for tutorials above).
Roon then is just the “execution system” that runs the created filter and convolutes the audio data while playing music.
I fully understand that a filter generated using Acourate (or equivalent - I’m thinking here of FIR designer which is one of the few apps that runs under OSX) will be significantly better than using parametric equalisation.
What I was trying to raise was the question of how the impact of the filter being applied by Roon can be measured. As far as I am aware the filter generating software (Acourate, REW …) generates a series of test signals which are then analysed. Unless the test signals can be played VIA Roon then I can’t see any way in which the impact of the filter being applied in Roon can be measured. i.e. Software Test Tones -> Roon -> Apply Convolution -> Speakers -> Microphone -> Software Measurement.
Perhaps I’m being a bit thick so glad if anyone can tell me how to do this.
Measure (test tones generated by the measurement software)
Analyse measurements
Generate a correction filter of some description (be that convolution or series of parametric EQ)
Load these into your playback software in the most appropriate way
Play music back with and without DSP enabled and compare.
Repeat and experiment. Possibly you can skip measurements depending on how well you did them first time round.
@DuckSoup I think I see what you mean - you want to measure the effectiveness of the correction? I think you’d have to find a frequency sweep cd or download or Tidal (I have one from there) and play it back as an audio track from Roon. Failing that you’d just have to use software that allows you to test the full workflow - I’m guessing Acourare would let you do that, but I’m not sure.
I’ve looked at FIR designer. It seems very capable but you need a lot of manual input and fairly detailed understanding of the issues you’re solving, and interpreting what it’s giving you back. Great for experts in DRC, not so good for the rest of us.
Thanks - and yes, I was hoping to test the effectiveness of the correction in Roon.
Maybe the best way (as an initial step) would be to record the output from REW test sweeps as a file (that’s easy) and then play that back in Room whilst running measurement in REW. I’ll have a play tomorrow (watching rugby union here in UK today) and report back.
That is quite interesting. What is the format of the filter you use for convolving with time domain compensation? Is it an Acourate generated wav file? Or a set of IIR filters that need to be read into Roon DSP?
My current convolution wav from REW uses 33k filter taps, looking at the signal path in Roon. Will Roon adopt your 131k filter taps?
If that al works, it would justify a purchase of Acourate imo.
You can definitely get perfect sweeps as wav file or streamable that you can just add to your Roon library and play. Better that they’re perfect I think.
Overmore this parametric eq doesn’t compensate for Time Domain alterations as a good convolution filter does!