• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Obsessive compulsive manual DRC in REW/rePhase

I wonder if there is a way to incorporate the unechoic data magnitude corrections (since i have them from spinorama.org) to the acourate correction process and the resulting convolution filter. Thinking out loud…

If you are able to obtain the anechoic data in .WAV or .DBL format I am sure I could figure out a way to incorporate it into Acourate's correction process. This is what I would do:

- Determine which freq range you want for room correction, and which you want for speaker correction. I would overlay the obtained anechoic data over the measured data and look to see where they deviate.
- Use SplitNJoin to join the two measurements. Note that this only joins the amplitude measurements, and not the timing measurements.
- There is an Acourate procedure to use separate amplitude and timing measurements and incorporate them into the same correction. See my document for details.
 
  • Like
Reactions: CK.
If you are able to obtain the anechoic data in .WAV or .DBL format I am sure I could figure out a way to incorporate it into Acourate's correction process. This is what I would do:

- Determine which freq range you want for room correction, and which you want for speaker correction. I would overlay the obtained anechoic data over the measured data and look to see where they deviate.
- Use SplitNJoin to join the two measurements. Note that this only joins the amplitude measurements, and not the timing measurements.
- There is an Acourate procedure to use separate amplitude and timing measurements and incorporate them into the same correction. See my document for details.
Thanks! will look into it.
 
Greetings @Keith_W , everybody,

So I have been trying out the Acourate macros, and noticed the following:

1. The RT60 calculated in REW (Photo 3) vs in Acourate (photo 4) have > 100ms difference. Is this reasonable?

2. In all the cases, the ICCP after correction is less than ICCP before correction (i.e. a negative difference). Attached the PulseL/R. again is this reasonable?

3. When measuring the FR to evaluate the filters, there is a significant different in the L+R curve when my AVR is in stereo (mode used for initial measurement) vs dolby digital mode. (attached photo-1 and photo-2). Does the upmixing engine introduce some kind of delay or phase adjustment? not sure if this is to be expected.


Thanks again!

P.S. All attached measurements have been done without changing the mic position and with all other parameters being equal.
 

Attachments

  • Photo-1_L+R_upmixed_dolby_digital.png
    Photo-1_L+R_upmixed_dolby_digital.png
    251.4 KB · Views: 10
  • Photo-2_L+R_stereo.png
    Photo-2_L+R_stereo.png
    237.8 KB · Views: 12
  • Photo-3_L+R_RT60_REW.png
    Photo-3_L+R_RT60_REW.png
    147.9 KB · Views: 12
  • Photo-4_L+R_RT60_Acou.png
    Photo-4_L+R_RT60_Acou.png
    100.7 KB · Views: 13
  • Pulse48L.zip
    983.6 KB · Views: 5
1. The reason there is a discrepancy between the RT60 in Acourate and REW is because you are comparing the Topt in REW with the RT60 in Acourate. Read this to find out more about the Topt.

2. I ran your Pulse48L/R through Acourate's macros and this is the ICCC before/after. I get an improvement of 0.6% for the ICCC10. Most of the "downgrade" comes from ICCClate, which worsens from 61.2% to 11.9% (worse by 49.3%). Acourate is prioritising correcting the early ICCC over late.

1737686976487.png


I will explain. The ICCC (Interchannel Cross-Correlation) is a measure of how similar the left and right channels are. It is split into several components as shown by Acourate - early is 0-10ms and it is by far the most important. Then there is very early (0-20ms), early (0-80ms) and late (80ms to pulse end) which in your case is 1.36 seconds. ICCC10 is mostly your speakers, and ICCC20 and above is mostly your room. As you can imagine, Acourate can only manipulate the signal. It has no influence on what happens to the signal after it leaves the speakers.

The reason the ICCC is so poor is because of the way the speakers have been set up in your room. Acourate can make your speakers completely identical. It can even make your speakers behave more identical if they are poorly set up in your room. But there is a limit to what it can do, if the stereo setup is not symmetrical in your room and it is causing egregious issues, there is nothing that any DSP can do. That is akin to moving walls with DSP.

3. Please don't truncate the legends when posting graphs. What are we looking at here? If the red line is the frequency response, you can see that they are clearly not the same. As is the phase response. So your AVR is clearly doing something. I don't know what.

I will look at your curves in more detail when I have time and let you know if I see anything interesting. Can you please explain how you have your system set up? I need to know the signal chain, and how you have connected your sub. Also post a picture or draw a diagram of your listening room.
 
Last edited:
  • Like
Reactions: CK.
Anyway let's look at some curves.

1737691001556.png


This is "before" (dark/light blue = L/R, black = sum of L+R). We can see that your subwoofer is way too loud. There are no summing issues when all the speakers are added together, meaning that all your speakers are in phase before you started.

1737691157171.png


This is "after" a quick and dirty room correction that I did. Green/red = L/R, brown = sum of L+R. We can see that Acourate has fixed the bass issue and smoothed out the frequency response nicely. Note that I chose FDW 15/1 meaning that Acourate only very gently corrects the upper frequencies.

1737691546951.png


This is the "before" step response. We can see that the subwoofer impulse is higher than the tweeter/woofer impulse, and it arrives late.

1737691630021.png


The "after" step response, after a very simple correction, looks heaps better. Lowering the volume of the sub alone has removed all that horrible bass ringing (those huge deflections you see in the "before" step response).

1737692337469.png


This is the first 20ms of the Energy-Time Curve. I have marked the -15dB target with the light blue line and the 20ms on the horizontal scale. I have also marked the reflections with arrows. Here, we can see a lot of early reflections indicating a major problem with your measurement technique. The first reflection (marked 1) is 0.65ms after the main impulse, and it is pretty loud. That translates to a distance of 0.22m. Or put another way, any measurement under 1548Hz (1000/0.65) is contaminated by reflections and can not be gated out.

We can also see that the left/right reflections are not symmetrical, indicating the speakers are not set up symmetrically in the room.

This is why I want to see a drawing of how the system has been set up.
 
  • Like
Reactions: CK.
Anyway let's look at some curves.

View attachment 423425

This is "before" (dark/light blue = L/R, black = sum of L+R). We can see that your subwoofer is way too loud. There are no summing issues when all the speakers are added together, meaning that all your speakers are in phase before you started.

View attachment 423426

This is "after" a quick and dirty room correction that I did. Green/red = L/R, brown = sum of L+R. We can see that Acourate has fixed the bass issue and smoothed out the frequency response nicely. Note that I chose FDW 15/1 meaning that Acourate only very gently corrects the upper frequencies.

View attachment 423427

This is the "before" step response. We can see that the subwoofer impulse is higher than the tweeter/woofer impulse, and it arrives late.

View attachment 423428

The "after" step response, after a very simple correction, looks heaps better. Lowering the volume of the sub alone has removed all that horrible bass ringing (those huge deflections you see in the "before" step response).

View attachment 423429

This is the first 20ms of the Energy-Time Curve. I have marked the -15dB target with the light blue line and the 20ms on the horizontal scale. I have also marked the reflections with arrows. Here, we can see a lot of early reflections indicating a major problem with your measurement technique. The first reflection (marked 1) is 0.65ms after the main impulse, and it is pretty loud. That translates to a distance of 0.22m. Or put another way, any measurement under 1548Hz (1000/0.65) is contaminated by reflections and can not be gated out.

We can also see that the left/right reflections are not symmetrical, indicating the speakers are not set up symmetrically in the room.

This is why I want to see a drawing of how the system has been set up.
First of all huge thanks for taking the time, really good analysis.

1. Regarding the upmixing, yes the FR and phase are different when using the dolby surround mode and obviously the avr is doing something (room correction is turned off, of course). I was wondering if this behavior is typical with upmixing maybe or sound modes in processors in general.

2. The sub arriving late, is this an issue? Should I change the delay settings to match the main speakers. I have only phase matched the sub with mains.

3. The setup for music: two speakers connected to avr and a sub with an avr XO of 80hz.
Room correction is disabled and distances (3m) of main speakers in avr are identical since indeed they are equidinstant from mlp (smaller distance between them though maybe 2.2m). Delay of sub set for phase matching with L+R (about 2.6m).

4. Room: attached photo, its a bit old but almost nothing has changed except speakers moved back a bit. L shaped room, open in one side and mlp close to back wall with acoustic panels only in back wall.


When measuring, I use a focusrite interface to the avr analogue inputs and a dayton xlr mic.
Mic is positioned paralel to the floor but close to back wall where mlp is.
 

Attachments

  • IMG_6964.jpeg
    IMG_6964.jpeg
    930.8 KB · Views: 21
  • IMG_6963.jpeg
    IMG_6963.jpeg
    902.3 KB · Views: 20
First of all huge thanks for taking the time, really good analysis.

1. Regarding the upmixing, yes the FR and phase are different when using the dolby surround mode and obviously the avr is doing something (room correction is turned off, of course). I was wondering if this behavior is typical with upmixing maybe or sound modes in processors in general.

I don't know about AVR's. I hate them. They are all overly complex and do things without me asking. Maybe your AVR is trying to "improve" things. God only knows!! Give me a dumb amplifier any day. If I were you, I would switch of all the processing in your AVR and let Acourate do the DSP. But it's your AVR, so you figure it out ;)

2. The sub arriving late, is this an issue? Should I change the delay settings to match the main speakers. I have only phase matched the sub with mains.

Not an issue. As you can see, Acourate fixed it.

3. The setup for music: two speakers connected to avr and a sub with an avr XO of 80hz.
Room correction is disabled and distances (3m) of main speakers in avr are identical since indeed they are equidinstant from mlp (smaller distance between them though maybe 2.2m). Delay of sub set for phase matching with L+R (about 2.6m).

I see. So as far as Acourate is concerned, it thinks there are two speakers with no subwoofer.

FYI since the start of this thread, I have released my free Acourate guide. I suggest you download and read it.

4. Room: attached photo, its a bit old but almost nothing has changed except speakers moved back a bit. L shaped room, open in one side and mlp close to back wall with acoustic panels only in back wall.

When measuring, I use a focusrite interface to the avr analogue inputs and a dayton xlr mic.
Mic is positioned paralel to the floor but close to back wall where mlp is.

Yeah, that explains all the early reflections. The reason why ICCC-late is uncorrelated is because of the L-shaped room. Nothing Acourate can do about that, i'm afraid. The only solution is to rearrange the system or buy a new house. But it's easier to simply not worry about it, which is what I advise you to do.
 
I don't know about AVR's. I hate them. They are all overly complex and do things without me asking. Maybe your AVR is trying to "improve" things. God only knows!! Give me a dumb amplifier any day. If I were you, I would switch of all the processing in your AVR and let Acourate do the DSP. But it's your AVR, so you figure it out ;)
Its true you need to be very much aware of what they are doing. They are however flexible, festure rich and cost effective. Yes will look into it out of curiosity.

Not an issue. As you can see, Acourate fixed it.



I see. So as far as Acourate is concerned, it thinks there are two speakers with no subwoofer.
Yes indeed, makes the process easier.
FYI since the start of this thread, I have released my free Acourate guide. I suggest you download and read it.
Yes already read the version 2! Really good content!

Yeah, that explains all the early reflections. The reason why ICCC-late is uncorrelated is because of the L-shaped room. Nothing Acourate can do about that, i'm afraid. The only solution is to rearrange the system or buy a new house. But it's easier to simply not worry about it, which is what I advise you to do.
Indeed since this is a temp living arrangement I learnt not to worry about it, provided that I have done all that I can in terms of dsp.

Thanks again for all the help, looking forward to your book!
 
Back
Top Bottom