RoomFit volume settings

[EddNog]

I hope you didn't get my comments wrong. I do not argue against EQing for pleasant low level listening. Not at all.

I am merely questioning the unrestricted and absolute usefulness of the equal loudness contour. Too many people get the meaning wrong, some seem to think that every sound event at a given average SPL would "require" the associated correction curve to sound "right". That's plain BS, of course.

Even when applied appropriately an equal loudness contour correction is not magically and naturally right. It's a valid starting point. But if you cannot have the original sound experience, why not just go with what gives you personally the best illusion?

I think that’s the alleged point of Magic Beans…supposedly analyzes your room arrangement and somehow creates a special curve that is supposed to sound best specifically for your room. No idea how that would work, and I’m not paying for it to try.

-Ed

 

[harkpabst]

I think that’s the alleged point of Magic Beans…supposedly analyzes your room arrangement and somehow creates a special curve that is supposed to sound best specifically for your room. No idea how that would work, and I’m not paying for it to try.

I'm sceptical of anything with magic in its name.

In any case, if the RoomFit results are different for different measurement levels it's certainly worth exploring the reasons.

 

[Wiimer]

It may because it sounds better to raise the bass at low volumes but going back to my original comment in #1, I do see different measurements and RoomFit results at different volumes. So which one is the one I should base my settings on?

I don't know if that is normal and if it is due to the mic, the room or the speakers?

I will make some more experiments when I get the Dayton Audio iMM-6C.

In my case, when performing MMM measurements with RoomFit, I experienced the IMM-6C picking up room noise at low volumes. The cause was the occasional noise from the fridge in my mini-kitchen.

 

[EddNog]

In my case, when performing MMM measurements with RoomFit, I experienced the IMM-6C picking up room noise at low volumes. The cause was the occasional noise from the fridge in my mini-kitchen.

This is typically why it’s recommended for the sweep to be 40dB above whatever noise floor is, so if noise floor is 35dB, you’d want the sweep to hit at least 75dB.

-Ed

 

[Wiimer]

This is typically why it’s recommended for the sweep to be 40dB above whatever noise floor is, so if noise floor is 35dB, you’d want the sweep to hit at least 75dB.

-Ed

Yeah, I'm a baby only learnt it in May this year

 

[harkpabst]

In particular, the moving microphone method is more prone to noise than sweeps.

 

[canard]

In particular, the moving microphone method is more prone to noise than sweeps.

At a reasonable level, the noise issue isn't really a problem for correction. It's for distortion measurements that it is, and the precautions regarding sweep methodology/measurement depth/level are essential, delicat.... ;-)

 

[harkpabst]

The main goal of this thread is to determine what is a reasonable level, isn't it?

 

[dominikz]

The Munson-Fletcher curves (and later corrections) might be about the same for everyone. The itch to "correct" for this is not.

Our hearing shows the same insensitivity to highs and (in particular) lows, no matter if we are listening to recoreded or live sounds. Nobody would ever suggest to "correct" the too low bass when the orchestra plays a piano section of a classical work. This is the way our hearing works and we're used to it, it's a characteristic, not a defect.

Now, if we listen to recorded music at below the original loudness (which is hardly defined at all for purely electronic music) some argue that we "must" compensate for our hearing's insensitivity for a "natural" perception. I'd argue that the result is still not "natural" at all. It's a listening experience that is impossible to have in the real world. It might sound more pleasing (which is absolutely OK) but it's by no means "natural" or "more correct". It's a matter of taste, nothing else.

Now, with independent EQ and RC on the horizon everybody can make use of different EQ settings based on mood, programme material or preferred behaviour at different volumes. RoomFit cannot and should not even attempt to bring these things together.

I hope you didn't get my comments wrong. I do not argue against EQing for pleasant low level listening. Not at all.

I am merely questioning the unrestricted and absolute usefulness of the equal loudness contour. Too many people get the meaning wrong, some seem to think that every sound event at a given average SPL would "require" the associated correction curve to sound "right". That's plain BS, of course.

Even when applied appropriately an equal loudness contour correction is not magically and naturally right. It's a valid starting point. But if you cannot have the original sound experience, why not just go with what gives you personally the best illusion?

These are really insightful and well-worded posts, I agree completely!

As @dominikz pointed out there is no technical reason to believe that room influence depends on level. The acoustics behind that are very linear (maybe unless we talk very small room volumes and very high sound pressure levels).

The microphone being non-linear is a far more likely scenario. So, how loud is loud in absolute terms? If in doubt I'd probably go with the lower level for the sweeps (unless you have something like the UMIK-2).

There's of course several reasons why multiple room measurements could produce different results, these are IMHO some of the main ones:

1. If single-point sweep measurements are used (i.e. no spatial averaging), the differences between measurement attempts could be significant due to minor changes in microphone position between takes, resulting in a different acoustical interference pattern being recorded.
2. If MMM measurements are used, ambient noise can cause significant differences in the measured response. E.g. in my system I will measure low-frequency peaks with MMM when a truck drives down the street, which aren't there when the street is silent. Other sources of narrow-band noise or hum like fans, AC, refrigerators, etc. can all cause deviations in the MMM measured response if measurement SNR is low enough.
3. Changes in room layout (e.g. opening a window vs closing it) will change the low-frequency room resonances and bass level - see some examples here.
4. Saturation/compression happening in the microphone and/or loudspeaker - though I'd only expect this to happen to a meaningful degree at pretty high levels.

As I mentioned in some other posts, note that MMM is robust to microphone placement differences, but sensitive to ambient noise. Single-point sweep is the opposite.
Multi-point sweep (e.g. like in Dirac Live) may be the best of both worlds, but takes much more time to perform.

The main goal of this thread is to determine what is a reasonable level, isn't it?

When measuring my in-room response I usually just go with a little bit louder than I'd like - that seems to be loud enough most of the time.

This is typically why it’s recommended for the sweep to be 40dB above whatever noise floor is, so if noise floor is 35dB, you’d want the sweep to hit at least 75dB.

-Ed

That is definitely ideal, but might not be realistic for everyone, depending on where they live. :/
Luckily, a somewhat lower SNR than that should still work OK in many cases (but depending on noise profile), IMHO!

 

[canard]

We often point to standardized approaches at 80-81 dB... which is already quite loud... and certainly needs to be adapted to everyday listening habits, it seems to me ;-)


Measurements at relatively quiet times are, of course, welcome... while avoiding helicopter or fighter jet overhead ;-)