Unfortunately not on hand.
I would need to take this measurement.
Wouldn’t be until the weekend.
No worries.
Ever done the Dirac Live free trial on PC/Mac?
For sh!ts and giggles / learning / fun comparison
I know there’s isn’t seamless integration with Roon but if you really like Dirac after trialling the latest version, there are ways to get Roon playing to Dirac Live processor on Windows 10 / macOS
No not yet, but I would like to try.
I upload a new measurement. I had a few minutes.
I cant get the water falls to work.
Measurements were taken with a static mic position.
Left, Right, and combined stereo.
Right has more free space around the speaker, left is sitting proud of a corner.
I essentially have a square room (bedroom), for my audio space so the measurements could be worse.
3x3.3m
Thanks
1 Like
I do have experience with DSP in general.
I.e. Pro sound products.
Qsys DSP, Biamp, Symetrix, Powersoft etc.
I have always done correction using PEQ, or grey box FIR filters.
All good but as you found, your moving mic measurement showed nothing in time domain.
One useful measurement is group delay which (among many things) can give an indication of frequencies you maybe should avoid correcting. Sometimes it can show up easily in frequency response alone but sometimes not.
When you combine that with the potential problem @Marin_Weigel hinted at, it can result in a sub-optimal end result (maybe, maybe not).
The moving mic method has advantages but also disadvantages if you only look at those results in isolation.
Dirac attempts to factor all this in automatically and does a really good job I’v heard from users and read. And I’ve never even used Dirac so genuinely interested if you trialled it (free), if the result sounded good to you.
Note I’m not an expert. Just sharing what I’ve learnt from experts in the field.
1 Like
I’ll offer my experience here. I experimented with REW and a UMIK mic with good results, but I wasn’t blown away by my efforts. I then tried a Mini DSP box with Dirac built in, the measurement practise with REW stands you in good stead but the Dirac process is certainly easier to carry out consistently. It basically tells you to “move mic to point x and play the test tones”. You then give it a target and let it work out the correction itself. To my ears the results were clearly better than my own efforts and I still have the DSP box 2.5 years later.
2 Likes
In my system, altering higher frequency tends to remove life from music. And also, less is more.
Here is how I did last correction I am using now:
- Find a good spot to stop correction (i.e. a place where everything to the right is good enough), for me that was 800Hz
- Limit the number of filters (EQ button and disable filters), 6-10 should typically be enough
- Autocorrect in REW from 20Hz up to limit determined by 1
- Generate convolution files
If you want more control of convolution file creation, use rePhase and import REW corrections from 3 (change EQ to rePhase and save from REW). Now you can set sample rate and number of taps, try to set number of taps to give a little latency (around 0.5 seconds) because that means its high enough resolution.
2 Likes
This is still using RTA right? Not using “Measure” in REW with measurement sweep.
So you won’t get any impulse response data this way.
I appreciate your comments.
I took another set of measurements using RTA, with no luck.
Waterfalls etc will not generate.
I read a few threads of people experiencing similar issues.
So I reinstalled REW.
I took new measurements utilising the measurements sweep function rather than the RTA function.
This enabled the waterfalls etc to be generated.
I have provided left, right and stereo combined responses.
Let me know your thoughts.
All are RAW measurements.
1 Like
Yes, I just figured this out.
I just took sweeps using measure.
1 Like
1 Like
If you only look at the frequency response without looking at time domain results in parallel, then yes looks nice but see my earlier post.
For example, looking at your right speaker group delay plot… I don’t want to generalise here because this is a really complex topic and no conclusions can be made from a single plot and I’m still learning myself (aren’t we all!) but as I mentioned above, EQ’ing some of these GD peaks could be sub-optimal.
You can improve GD using RePhase but that’s complicated and probably off topic here in this thread.
If you can run Dirac Live trial on Mac/PC, one really interesting thing would be to do an REW measurement playing measurement sweeps with DIrac Live correction in the chain…
And see the resulting frequency response, step response and group delay plots.
And of course, compare if things have improved subjectively to you too (or if the result is still not great).
A quick google from REW’s author gave this explanation, for those wondering what group delay is:
Cheers for that info:
I’ll do some reading. I’m almost certain those peaks are reflections in my room, I could hear them during the sweeps. I have ringing at some frequencies. I am addressing slowly with Room treatment, but my room orientation/finishes make it difficult.
See window and cupboard door either side of speaker.
I will treat the roof next. I usually pull the blinds when I listen.
I have tried orientating the room 90 degrees, but I lose all bass response.
6 Likes
As you should be and your ears were probably right on the money with things being worse with REW’s AutoEQ.
Do give the Dirac Live free trial a spin when you get a chance ! And if you can share the raw measurement with Dirac Live in the chain, if you do try it out.
Will do thanks. I appreciate your time and comments.
I’ll try a few filters on my DAC/Roon, and put some towels up on the walls to see what affect this has on group delay/time domain issues.
Looks like you get 14 days to try…
1 Like
Do the first measurements and listening with Dirac with your room ‘as is’. Just to do an apples to apples comparison comparing with your room right now.
Then go bonkers with room treatment measurements later to compare
2 Likes
I never used it but seen this guide:
http://diracdocs.com/Roon_&_Dirac_Live_2.0.pdf
There’s also some Roon users that could probably help you if you get stuck:
1 Like
Very informative thread!
I use REW and a mini DSP 4x10 but looking to get Roon (and Dirac).
Question, can the Dirac implementation be used on a raspberry pie with Roon? I don’t think so right?
(maybe noob question but new to Roon)
Hi @Joost_Eijkenboom and welcome to the Roon forum. I’m pretty sure that there’s no version of Dirac for ARM devices (Raspberry Pis) as things stand. You could almost certainly run light FIR filters on a Pi3 or Pi4 but you’d have to make them yourself. I run a MiniDSP DDRC-22D for Dirac which I’ve had for a few years. I looked for a Pi implementation at the time to no avail, and a quick Google search suggests nothing has changed there.
1 Like