New User Focus Fidelity

[Daverich4]

I've used the services of Home Audio Fidelity a number of times to help the sound of my system. A pillar was added to my listening room recently to level the kitchen floor above it and I thought I'd try and do some of the corrections myself using Focus Fidelity and REW. The front wall of the room is 22 feet across and the back wall is 19 feet away. The listening seat is 1/3 of the way into the room from the back wall. I tried a number of the Target Curves that come with Focus Fidelity and prefer the Harman Target but with the bass boosted 3dB instead of the default 6dB. I use Roon to drive the system with the Convolution filters. It sounds pretty decent to me but there are some bass notes down in volume compared to the overall bass. I have no way of knowing if those correspond to the dips in the measured curves or not. I'm new to Focus Fidelity and REW but have read most of the relevant articles here as well as several of the excellent E-books that have been posted. Would it be possible for someone here to take a look at my results to see if I'm on the right track? If I've not provided enough information or posted incorrectly please let me know. Thanks. (I tried to upload the .mdat file but the site isn't letting me. Does it need to be in some other format)?

 

[staticV3]

Would it be possible for someone here to take a look at my results to see if I'm on the right track?

The measurements aren't really useful in the way they're displayed in your screenshot.

For Sweep measurements, Var smoothing is a good compromise that doesn't hide room modes.

Before exporting, you want to click on the "20 .. 20,000" button in the bottom right corner.

Then open the Limits windows with the button top right and click on "Fit y to data".

Finally, use the Camera icon top left to export the graph, making sure that Legend is turned on.

That being said, for meaningful in-room frequency response measurements, using the Moving Microphone Method is greatly preferred over the default Sweep.

(I tried to upload the .mdat file but the site isn't letting me. Does it need to be in some other format)?

ASR doesn't support uploading .mdat directly. You'll need to Zip it first.

 

[Daverich4]

Thank you for getting back to me and helping me out. I've redone the chart with the settings you've suggested and zipped the .mdat. I did notice that using your settings the modified Harman curve looks kind of odd. As far as using the Moving Microphone Method, my room is set up for a single person (me) and I used a total of 7 measurements. One in the middle, two on either side and one front and one in back. Would it still be better to use the Moving Microphone technique? Thanks.

 

[staticV3]

I did notice that using your settings the modified Harman curve looks kind of odd.

"Fit y to data" was a quick and easy to explain method to reign in the y-axis scale a bit which was very zoomed out before, but did end up zooming in a bit much.

Generally, if the range between the highest displayed SPL and the lowest is 50dB, then that's a nice balance for in-room frequency response.

For your data, something like 45 to 95dB SPL would do well.

As far as using the Moving Microphone Method, my room is set up for a single person (me) and I used a total of 7 measurements. One in the middle, two on either side and one front and one in back. Would it still be better to use the Moving Microphone technique?

I would always use MMM to assess in-room response.

Averaging multiple sweeps is a step in the right direction, but MMM is best. This is well explained in the white paper (PDF download link): https://www.ohl.to/audio/downloads/MMM-moving-mic-measurement.pdf

One caveat is that you want to play L+R Correlated for accurate bass response, as bass in tracks is usually coherent.
Here for example my TV speakers:

You can see how the combined bass response doesn't really match either of the individual bass responses, but it's the combined response that you'll hear like 95% of the time.

In contrast, treble you want to measure either L and R separately (if you have asymmetries and can correct channels individually), or L+R Uncorrelated if you want one correction presets for both channels.

Here's an example of what can happen with correlated L+R:

Need help with in-room measurements

Can anyone please tell me how these measurements look? This was done from a stationary mic from listening position (Beyerdynamic MM-1 with 0 degree cal file.) I have a custom setup with two different size subwoofers next to each speaker and am using crossovers to make it work. I played with the...

www.audiosciencereview.com

The REW Beta has that Uncorrelated option when generating noise for MMM, not sure if it has been added to the Stable build as well.

BTW, measurements using the Moving Microphone Method do not need additional smoothing applied to them, as the resulting 1/48 Oct data is usually very readable already. E.g:

(Green: stock, Blue: with EQ)

Just make sure that in the RTA window->Cogwheel->Appearance, you turn off "Use bars on RTA".

 

[Daverich4]

Thank you, I guess the next step for me is to read through the links.

 

[Daverich4]

"Fit y to data" was a quick and easy to explain method to reign in the y-axis scale a bit which was very zoomed out before, but did end up zooming in a bit much.

Generally, if the range between the highest displayed SPL and the lowest is 50dB, then that's a nice balance for in-room frequency response.

For your data, something like 45 to 95dB SPL would do well.


I would always use MMM to assess in-room response.

Averaging multiple sweeps is a step in the right direction, but MMM is best. This is well explained in the white paper (PDF download link): https://www.ohl.to/audio/downloads/MMM-moving-mic-measurement.pdf

One caveat is that you want to play L+R Correlated for accurate bass response, as bass in tracks is usually coherent.
Here for example my TV speakers:
View attachment 517434
You can see how the combined bass response doesn't really match either of the individual bass responses, but it's the combined response that you'll hear like 95% of the time.

In contrast, treble you want to measure either L and R separately (if you have asymmetries and can correct channels individually), or L+R Uncorrelated if you want one correction presets for both channels.

Here's an example of what can happen with correlated L+R:

Need help with in-room measurements

Can anyone please tell me how these measurements look? This was done from a stationary mic from listening position (Beyerdynamic MM-1 with 0 degree cal file.) I have a custom setup with two different size subwoofers next to each speaker and am using crossovers to make it work. I played with the...

www.audiosciencereview.com

The REW Beta has that Uncorrelated option when generating noise for MMM, not sure if it has been added to the Stable build as well.

BTW, measurements using the Moving Microphone Method do not need additional smoothing applied to them, as the resulting 1/48 Oct data is usually very readable already. E.g:
View attachment 517435
(Green: stock, Blue: with EQ)

Just make sure that in the RTA window->Cogwheel->Appearance, you turn off "Use bars on RTA".

I’ve read through both of the links you supplied and it’s clear to me that I could use a little more entry level help. I could do the MMM measurement but with no comprehension of why it’s superior. The second link about the correlated is over my head at this time. Thanks for trying to point me in the right direction but I have a lot to learn before any of this is useful.

 

[kemmler3D]

I’ve read through both of the links you supplied and it’s clear to me that I could use a little more entry level help. I could do the MMM measurement but with no comprehension of why it’s superior. The second link about the correlated is over my head at this time. Thanks for trying to point me in the right direction but I have a lot to learn before any of this is useful.

The gist is that bass builds up in specific spots in the room and either reinforces or cancels out, i.e. "standing waves' or modes.

To hear it for yourself just put on a steady bass tone like 60hz and walk around the room. You will probably notice it's much louder in some spots than others. The spots in the room that are louder vary by frequency, so if you're staying in one spot, some tones will be much louder than others, which is not what we want, of course. This shows up in most room measurements as a mountain range-like shape in the graph below 200hz, but what we'd rather see is a flat line.

When you do a measurement in one spot it can tend to exaggerate this effect compared to what it actually sounds like when you listen.

So, the moving mic method ends up averaging it out over a certain area by moving the mic through some of the quieter and louder areas while the measurement is taking place. This gives you a more realistic measurement of what you'll hear, since nobody sits in the same exact spot down to the inch when they listen. When you do your room correction on the basis of MMM, you have less propensity to over-correct the sound.

Correlated noise just means both channels (L+R) are playing the same noise at the same time. Otherwise it's random on both sides. Since music tends to have the same thing going on in the bass in both channels, it makes sense to use correlated noise to correct the bass.

 

[Daverich4]

The gist is that bass builds up in specific spots in the room and either reinforces or cancels out, i.e. "standing waves' or modes.

To hear it for yourself just put on a steady bass tone like 60hz and walk around the room. You will probably notice it's much louder in some spots than others. The spots in the room that are louder vary by frequency, so if you're staying in one spot, some tones will be much louder than others, which is not what we want, of course. This shows up in most room measurements as a mountain range-like shape in the graph below 200hz, but what we'd rather see is a flat line.

When you do a measurement in one spot it can tend to exaggerate this effect compared to what it actually sounds like when you listen.

So, the moving mic method ends up averaging it out over a certain area by moving the mic through some of the quieter and louder areas while the measurement is taking place. This gives you a more realistic measurement of what you'll hear, since nobody sits in the same exact spot down to the inch when they listen. When you do your room correction on the basis of MMM, you have less propensity to over-correct the sound.

Correlated noise just means both channels (L+R) are playing the same noise at the same time. Otherwise it's random on both sides. Since music tends to have the same thing going on in the bass in both channels, it makes sense to use correlated noise to correct the bass.

Thanks for jumping in, your comments are really helpful.