Here is a calculated curve without crossover in rephase and what miniDSP HD makes of it.
Rephase: -14dB@ ~3840Hz
MiniDSP HD: -14dB@ ~7580Hz
Everything seems shifted to a higher octave
I would think it is because you need a 96kHz sample rate filter, presumably it doesn't do a sample rate conversion for you.
Wow, thanks for the fast reply. I hadn't thought of that wrongly assuming it would be a standard 48KHz sample-rate (the rephase screenshot on the miniDSP website shows 48KHz). With the IIR filters the manual says to use a 96KHz program to build biquads, nothing about sample-rate with FIR filters. But somewhere all the way in the back of the manual there is this:
Audio resolution
24-bit input and output, 96 kHz internal sample rate (depending on plugin)
Doesn't seem like there is any way to change it. 4K taps on 4 channels isn't that much if you have to use 96KHz :-(
Audio resolution
24-bit input and output, 96 kHz internal sample rate (depending on plugin)
Doesn't seem like there is any way to change it. 4K taps on 4 channels isn't that much if you have to use 96KHz :-(
Hi everyone,
Apologies for cross-posting, but I just made a little promotional effort for REW/rePhase on the occasion of the launch of the new interface between REW and rePhase and I wanted to share it with you 😉
Using REW and rephase to generate amplitude and time domain corrections
Apologies for cross-posting, but I just made a little promotional effort for REW/rePhase on the occasion of the launch of the new interface between REW and rePhase and I wanted to share it with you 😉
Using REW and rephase to generate amplitude and time domain corrections
@SwissBear
You claim, you're able to generate filters on "Acourate" level.
Have you compared both results??
Acourate puts a lot of "benefitial" effort into pre-ringing compensation.
How is REW/rePhase coping with that?
##############################
Another topic.
I'm in the process of setting up my measurement chain. I'd need some advice.
I stepped over several issues/questions:
1. All microphones except Earthworks seem to offer just decent frequency responses
They pretty much all need correction.
2. That leads to dedicated calibration efforts
I learned that usually calibration is done at 44.1kHz only.
What has to be done for other samplerates than 44.1?
3. I also read that most (cheap) microphones except Earthworks show some kind
of weird phase behavior. What am I supposed to do with these kind of microphones?
Can rePhase correct the phase of the microphone first? How?
Thx for your patience.
You claim, you're able to generate filters on "Acourate" level.
Have you compared both results??
Acourate puts a lot of "benefitial" effort into pre-ringing compensation.
How is REW/rePhase coping with that?
##############################
Another topic.
I'm in the process of setting up my measurement chain. I'd need some advice.
I stepped over several issues/questions:
1. All microphones except Earthworks seem to offer just decent frequency responses
They pretty much all need correction.
2. That leads to dedicated calibration efforts
I learned that usually calibration is done at 44.1kHz only.
What has to be done for other samplerates than 44.1?
3. I also read that most (cheap) microphones except Earthworks show some kind
of weird phase behavior. What am I supposed to do with these kind of microphones?
Can rePhase correct the phase of the microphone first? How?
Thx for your patience.
Hi soundcheck,
Yes, I have a license of Acourate and from my POV, both solutions are on par, as far as end results are concerned.
rePhase is probably more on the DIY type of approach, but if you like digging into things, this is a fantastic tool IMHO.
In order to get clean impulses from rePhase correction, you just need to keep close to the minimal phase, ie use minimal phase amplitude corrections as well as minimal phase filters, and this should work. In case you have difficulties, you can ask Pos who always is very helpful.
As far as I understand, rePhase is not a measurement tool. You will probably get better answers asking on the forum of measurement tools like REW or HOLMImpulse.
But you can use rePhase to generate a calibration file which includes phase information. Here is how to do it: https://translate.google.com/transl...et/showthread.php?tid=6103&edit-text=&act=url
Hope this helps.
Thank you SwissBear, nice pictorial walk through.
I kook forward to trying it.
And thank you POS, for the upcoming REW interface as well as the Tools and Bypass buttons in the Para Eq tab....
All this is really outside the scope of this thread and belongs in a thread about speaker measurement.
Just briefly:
- Sample rate of the calibration does not matter except at the very top end. If the calibration ends at 20K, who cares?
- No mic is perfectly flat. Changing the angle of even an omnidirectional mic will affect it's top end response. Most of the small mic capsules used are remarkable flat.
- Phase follows frequency response. The calibration file takes care of that.
"It's good enough" 🙄
"Sample rate does not matter on digital filters" 🙄
"Out of scope" 🙄
"Phase issues of microphones doesn't really matter" 🙄
Perhaps I'm completely off track here:
That's my understanding:
rePhase will also correct the microphone phase issues in conjunction with the other phase issues.
You then measure your correction with the same microphone.
And what will you see?
A perfect response!
But probably not a perfect speaker response.
You'll get a perfect (speaker+microphone) response.
I gave up on DRC in the past because results never turned out to be good enough.
The filter associated losses were always worse then the original.
What brought me back to the subject were statements about inaudible corrections
with Acourate. And of course recent statements that REW + rePhase can compete with
Acourate nowadays.
"Sample rate does not matter on digital filters" 🙄
"Out of scope" 🙄
"Phase issues of microphones doesn't really matter" 🙄
Perhaps I'm completely off track here:
That's my understanding:
rePhase will also correct the microphone phase issues in conjunction with the other phase issues.
You then measure your correction with the same microphone.
And what will you see?
A perfect response!
But probably not a perfect speaker response.
You'll get a perfect (speaker+microphone) response.
I gave up on DRC in the past because results never turned out to be good enough.
The filter associated losses were always worse then the original.
What brought me back to the subject were statements about inaudible corrections
with Acourate. And of course recent statements that REW + rePhase can compete with
Acourate nowadays.
Hello,
Trying to évaluate the phase response of a miniDSP Umik, with rePhase "invert" fonction, gives that curve:
meaning the phase measurement / correction error, using a calibration mic file without phase informations,
may be around 60° at low and high end.
cdt
Soundcheck,
If you look closely at what Pano said: phase will closely follow frequency response, so if you use a calibration file to correct the frequency response deviations on your measurement microphone it should correct the phase as well.
To me it looks like Acourate is your desired tool. So why not use that?
I've had plenty of fun with both RePhase and DRC-FIR. DRC-FIR works as advertised for me and I love the fact it is highly user adjustable with all of its variables. I did make my own templates though, and looked at each separate variable to determine what it was that I wanted to use.
The same with RePhase, it works very well for what it was designed to do. No worries there either. Just invest the time to learn these tools. It isn't magic, you are the one in control of both these tools. So I'd start looking at what you want to accomplish with it.
If you look closely at what Pano said: phase will closely follow frequency response, so if you use a calibration file to correct the frequency response deviations on your measurement microphone it should correct the phase as well.
To me it looks like Acourate is your desired tool. So why not use that?
I've had plenty of fun with both RePhase and DRC-FIR. DRC-FIR works as advertised for me and I love the fact it is highly user adjustable with all of its variables. I did make my own templates though, and looked at each separate variable to determine what it was that I wanted to use.
The same with RePhase, it works very well for what it was designed to do. No worries there either. Just invest the time to learn these tools. It isn't magic, you are the one in control of both these tools. So I'd start looking at what you want to accomplish with it.
Last edited:
Mr Soundcheck. Please reread what I wrote. Then do some research of your own. There are threads here on the subject and entire books published.
Doing some reading and studying will help clear up your misunderstandings, some of which are fundamental. Don't let your audionervosa get the better of you . 😉
Doing some reading and studying will help clear up your misunderstandings, some of which are fundamental. Don't let your audionervosa get the better of you . 😉
Sorry Guys.
Things like
"...It should correct as well..." .
What's that supposed to mean!?!? Either it corrects or not.
"...Acourate is for me..."
😕 It was just recently announced the filter quality with REW plus rePhase is en par with Acourate nowadays. ( No word about the preringing issue though)
Beside that Acourate cannot do anything about the recording chain either.
"...DRC-FIR works as advertised for me..."
What is that supposed to mean? As limited as advertised??
DRC-Fir is very well aware of the shortcomings and limitations. It's even
mentioned somewhere on the DRC pages. ( I ran a project myself some time back)
Looking at jmbee graphs IMO clearly confirms to me what I'm questioning.
There are potentially serious ( serious for people who're not having that "good enough" attitude) implications of phase issues and that's not just the microphone, no, the entire measurement chain up to the ADC might have issues.
Basically IMO the entire measurement-chain needs to be calibrated incl. phase and that
per samplerate - if you want rather correct and more than "good enough" results!
Otherwise you might even end up with false corrections.
I'm well aware that such a calibration is not easy.
I'm also aware that it is not easy to get calibration files other than 44.1kHz.
But things shouldn't be wiped under the carpet because there is no known solution to it.
Yep, "Mr." Pano. I reread what you wrote.
Beside other stuff:
"No mic is perfectly flat" - Thx for letting me/us know.
I havn't seen any mic that's remarkably flat - except a $600 Earthworks btw.
"The calibration file takes care of it" -- Yep. Under all conditions it (the FILE with whatever quality content ) will take care of it. Thx. Thx a lot for this.
"Sample rate does not matter except..." ...a little issue here and there. Yep.
There we have it again. It's "good enough".
Folks, why do you stay so vague all the time???
Things like
"...It should correct as well..." .
What's that supposed to mean!?!? Either it corrects or not.
"...Acourate is for me..."
😕 It was just recently announced the filter quality with REW plus rePhase is en par with Acourate nowadays. ( No word about the preringing issue though)
Beside that Acourate cannot do anything about the recording chain either.
"...DRC-FIR works as advertised for me..."
What is that supposed to mean? As limited as advertised??
DRC-Fir is very well aware of the shortcomings and limitations. It's even
mentioned somewhere on the DRC pages. ( I ran a project myself some time back)
Looking at jmbee graphs IMO clearly confirms to me what I'm questioning.
There are potentially serious ( serious for people who're not having that "good enough" attitude) implications of phase issues and that's not just the microphone, no, the entire measurement chain up to the ADC might have issues.
Basically IMO the entire measurement-chain needs to be calibrated incl. phase and that
per samplerate - if you want rather correct and more than "good enough" results!
Otherwise you might even end up with false corrections.
I'm well aware that such a calibration is not easy.
I'm also aware that it is not easy to get calibration files other than 44.1kHz.
But things shouldn't be wiped under the carpet because there is no known solution to it.
Yep, "Mr." Pano. I reread what you wrote.
Beside other stuff:
"No mic is perfectly flat" - Thx for letting me/us know.
I havn't seen any mic that's remarkably flat - except a $600 Earthworks btw.
"The calibration file takes care of it" -- Yep. Under all conditions it (the FILE with whatever quality content ) will take care of it. Thx. Thx a lot for this.
"Sample rate does not matter except..." ...a little issue here and there. Yep.
There we have it again. It's "good enough".
Folks, why do you stay so vague all the time???
As far as calibration is concerned, sampling rate only affects bandwidth.
So if your calibration file covers the range you are looking for you don't need to question the sampling rate it was recorded/processed at.
So if your calibration file covers the range you are looking for you don't need to question the sampling rate it was recorded/processed at.
I'm a first time user of Rephase with a couple of questions. I did a 3-way speaker using MiniDSP 4x10 and got good response at 1m in room with excellent time alignment and reasonably flat phase. Measuring out at the LP is another story so I want to add a MiniDSP OpenDRC into the path and use it both for room correction and to fit to a room curve. The room effects I need to deal with are at least one longitudinal mode, and floor and ceiling bounce.
These are full range speakers, EQed flat to 20 Hz, in the front corners. I noticed that measuring below 100 Hz, the left and right speakers played simultaneously filled holes in each other's response. This is reasonable because the LP is 2' closer to one speaker than the other. So for the OpenDRC EQ process, I blended an LF measurement (<=100Hz) of both speakers with an HF measurement of the speaker being EQed. The HF measurement was an average of measurements taken at my two primary listening positions, the two ends of a couch. The averaging lost the phase info so I'm just EQing the gain. Does that make sense so far?
Looking at rephase, I see paragraphic gain equalization with 17 PEQs per bank. One bank is almost but not quite enough filters. Can I use multiple banks? Is it the case that RePHase creates a FIR filter defined by the collection of PEQs that I define this way?
I found it tedious to manually tune all those individual PEQs so I opened REW and used its automatic EQ to create and fine tune PEQs for a generic DSP. I then manually transferred those PEQs into RePhase. For some reason yet undiscovered, the predicted response in REW didn't match the predicted response in RePhase. There were a few bumps to smooth over but in spite of this it did save a lot of time and I got a better end result than I had gotten wholly manually. Does anybody know why there would be a mismatch between REW and RePhase?
I still have a lot to learn about RePhase and REW auto-eq process but my next step is to load the filters I just created into the OpenDRC, remeasure, and then listen. Getting these questions answered, and further hints or suggestions, will make me feel a little less trepidation about it.
Thanks in advance,
Jack
These are full range speakers, EQed flat to 20 Hz, in the front corners. I noticed that measuring below 100 Hz, the left and right speakers played simultaneously filled holes in each other's response. This is reasonable because the LP is 2' closer to one speaker than the other. So for the OpenDRC EQ process, I blended an LF measurement (<=100Hz) of both speakers with an HF measurement of the speaker being EQed. The HF measurement was an average of measurements taken at my two primary listening positions, the two ends of a couch. The averaging lost the phase info so I'm just EQing the gain. Does that make sense so far?
Looking at rephase, I see paragraphic gain equalization with 17 PEQs per bank. One bank is almost but not quite enough filters. Can I use multiple banks? Is it the case that RePHase creates a FIR filter defined by the collection of PEQs that I define this way?
I found it tedious to manually tune all those individual PEQs so I opened REW and used its automatic EQ to create and fine tune PEQs for a generic DSP. I then manually transferred those PEQs into RePhase. For some reason yet undiscovered, the predicted response in REW didn't match the predicted response in RePhase. There were a few bumps to smooth over but in spite of this it did save a lot of time and I got a better end result than I had gotten wholly manually. Does anybody know why there would be a mismatch between REW and RePhase?
I still have a lot to learn about RePhase and REW auto-eq process but my next step is to load the filters I just created into the OpenDRC, remeasure, and then listen. Getting these questions answered, and further hints or suggestions, will make me feel a little less trepidation about it.
Thanks in advance,
Jack
It's vague because you don't understand the bigger picture and some of the fundamentals. You really should do some more reading about measuring loudspeakers. Otherwise you'll just end up chasing your tail. We are trying to help you avoid that.
You want to use rePhase, or something like it, for room corrections and/or crossovers. To get a measurement to base your room corrections on you start with a measurement of your system at your listening position. But remember, move the mic a centimeters and that measurement will change. That's something you really have to understand.
Room measurement is a big subject, but not the subject of this thread.
I would suggest that you start out with a economic measurement mic, like the Dayton calibrated one. Start doing measurements and you will quickly learn how tricky and variable it can be. It takes time, patience and discipline, which I'm sure you have.
Get some good tools, start measuring and start learning! You will have all sorts of new questions that you didn't expect. Don't get caught up in the fine details right now. Learn to walk before you fly.
If you want to know what it does for me, it's all hidden in a huge thread describing my project, you can find it here: http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array.html
On the first post of that thread, I've made a couple of links to important parts. What DRC-FIR does for me is more than I can describe in a few words here. I'm pleased with it's outcome, my goals do differ from yours though. There are some links to reviews, written by fellow forum members.
If you really want to know about it, please start reading. You're not being any more clear than us by stating: "...I ran a project myself some time back..." which does not make us understand why DRC-FIR and yourself could not get along.
Clearly it did not work out for you. Any other DRC question you may have, please use this thread: http://www.diyaudio.com/forums/full-range/275730-convolution-based-alternative-electrical-loudspeaker-correction-networks.html. This thread was started to help people get a good start with DRC-FIR. It's packed with a tutorial and some custom templates.
It does not add anything useful to this thread to be talking about programs that do not have anything to do with what pos intended for RePhase.
For the rest of the points you made Pano was way more clear than I could ever be. We are trying to help, please realise that.
Last edited:
Never mind, found some pilot errors, answered my own questions, thinking there must be a better way but don't want a PC in my playback path