Does anyone else prefer a dipped midrange?

[sigbergaudio]

You are Norwegian so I assume you understand Swedish.

faktiskt.io • Visa tråd - Mätningar för ekvalisering

Yep, I'm even half Swedish actually. I will have a look.

 

[sigbergaudio]

You are Norwegian so I assume you understand Swedish.

faktiskt.io • Visa tråd - Mätningar för ekvalisering

I tried to effectively skim the first entry here, and if I understand it correctly this is a suggestion for how to approach the job of equalizing out variations in the response, I didn't catch a direct reference to the fact that there should be no slope above 1khz? Do you have a more direct reference so I don't have to read four pages?

 

[abdo123]

I tried to effectively skim the first entry here, and if I understand it correctly this is a suggestion for how to approach the job of equalizing out variations in the response, I didn't catch a direct reference to the fact that there should be no slope above 1khz? Do you have a more direct reference so I don't have to read four pages?

I think they're pointing to the fact that the guide is equalizing to flat anechoic on-axis above 500Hz. Actually flat Everything till 50Hz if google translate is working as intended.

 

[goat76]

I tried to effectively skim the first entry here, and if I understand it correctly this is a suggestion for how to approach the job of equalizing out variations in the response, I didn't catch a direct reference to the fact that there should be no slope above 1khz? Do you have a more direct reference so I don't have to read four pages?

The conclusion is that you can’t look at a regular house curve measurement, made with a long measuring window, and use that to determine if you like a dipped midrange or not. The room itself can have added a lot of reflection to what the microphone has picked up, depending on the particular room's acoustics, and it may not be representative of what you hear.

It’s the same principle with the use of EQ. You should never base your EQ adjustments on a non-windowed measurement over about 500 Hz, you need to know how the direct sound response is above that range.

 

[sigbergaudio]

Ah found it: "Ekvalisera till konstant frekvensgång upp till 500 Hz med ett litet lyft om ca 3 dB under 50 Hz" - didn't catch that the first time around. The way I understand the approach described they're measuring 1-2 m directly in front of the speaker, not in the listening position. I agree that should basically be flat above 500hz.

Flat to 50hz and maximum +3dB below 50hz in the listening position is harder to agree with. That will imo sound bright to most people.

 

[goat76]

Flat to 50hz and maximum +3dB below 50hz in the listening position is harder to agree with. That will imo sound bright to most people.

I don't agree with that particular part, I was just referring to the above 500 Hz thing.

I think the frequency range under 500 Hz is way more subjective, and you can go for the Harman curve or whatever you prefer.

 

[Chromatischism]

The conclusion is that you can’t look at a regular house curve measurement, made with a long measuring window, and use that to determine if you like a dipped midrange or not. The room itself can have added a lot of reflection to what the microphone has picked up, depending on the particular room's acoustics, and it may not be representative of what you hear.

It’s the same principle with the use of EQ. You should never base your EQ adjustments on a non-windowed measurement over about 500 Hz, you need to know how the direct sound response is above that range.

If you are doing just speaker EQ, sure. But it sounds like you are saying that it's wrong for room correction to reduce sound power to tame a reflection due to choices in the domestic living environment. Of course it's best if you don't have to do that (lose the coffee table, move speakers, etc), but that's the basis for room correction.

 

[sigbergaudio]

I don't agree with that particular part, I was just referring to the above 500 Hz thing.

I think the frequency range under 500 Hz is way more subjective, and you can go for the Harman curve or whatever you prefer.

But also note that a speaker measuring flat on-axis 1m from the speaker doesn't necessarily measure flat in the listening position. My speakers measure flat from 500hz and out in such a measurement (and anechoically), but in the listening position I have a steadily falling slope. How straight it is from 1khz-10khz varies from room to room, and also microphone placement.

 

[terryforsythe]

Have you done any measurements with a short window above 500 Hz to make sure how your speakers direct sound performs above that point?

According to the “experts”, you should aim for a flat direct sound for your speakers from around 500 Hz and up (no matter how that looks like in a normal room curve measurement), because the direct sound will be the dominating thing you actually will hear at the listening position.

The "expert" advice to which you refer sounds more like a rule of thumb. In some rooms, a speaker with a flat frequency response can sound great without need for equalization above 500 Hz. But in other rooms, proper filtering will improve the sound.

My speakers have a fairly flat response (Elac Uni-Fi Reference UBR62). My room is a little tricky, though - tiled wall behind the speakers, tiled floor, large glass windows, etc. Thus, room reflections are not insignificant. Using Dirac Live to generate the room correction filters, which extend through audio frequency spectrum and are based on measurements made at and around the listening position, noticeably improved the sound of my system. Still, it is far from perfect.

 

[Thomas_A]

I think there is some misunderstanding. The flat on-axis measured anechoically or at close speaker distance should also be flat at LP when considering only the direct path from speaker to microphone. The only factor affecting the frequency response is the small drop in HF due to distance in air. This is however very small at normal listening distance. If you include reflected sound and the absorption in the measurement there will for sure be a gradual drop as the Harman room curve.

 

[goat76]

If you are doing just speaker EQ, sure. But it sounds like you are saying that it's wrong for room correction to reduce sound power to tame a reflection due to choices in the domestic living environment. Of course it's best if you don't have to do that (lose the coffee table, move speakers, etc), but that's the basis for room correction.

1. I like the idea of physical room acoustic treatment.

2. I like the idea of digital room correction in the bass region.

3. I like the idea of speaker correction above 500 Hz.

 

[goat76]

But also note that a speaker measuring flat on-axis 1m from the speaker doesn't necessarily measure flat in the listening position. My speakers measure flat from 500hz and out in such a measurement (and anechoically), but in the listening position I have a steadily falling slope. How straight it is from 1khz-10khz varies from room to room, and also microphone placement.

I’m not saying a flat speaker will measure flat in the listening position.

What I’m saying is that no matter how the measurements looks like from the listening position in a regular house curve measurement, you are sure your speakers measures flat in an anarchic chamber. Then you should never make any digital correction above 500 Hz because from around that frequency, and up, you will already perceive the flat response of the speakers/the direct sound (at a normal listening distance).

 

[terryforsythe]

I think there is some misunderstanding. The flat on-axis measured anechoically or at close speaker distance should also be flat at LP when considering only the direct path from speaker to microphone. The only factor affecting the frequency response is the small drop in HF due to distance in air. This is however very small at normal listening distance. If you include reflected sound and the absorption in the measurement there will for sure be a gradual drop as the Harman room curve.

I agree with respect to the direct path, and the gradual drop due to the room reflections. But in my room, for example, the interference from the reflected sound at the listening position is not perfectly linear with respect to frequency - the speakers do not have a perfectly linear dispersion pattern, and the room reflections are not perfectly linear. Indeed, my speakers measure fairly flat in the near field, and are not too far off from each other. But, where they are placed in the room means they are at different distances from the staircase on one side and the windows on the other side. Moreover, the reflection coefficients of the two sides of the room are not the same. In consequence, the frequency responses of the respective speakers at the listening position differ greater than their near-field frequency responses. With the filters generated by Dirac Live, their frequency responses as measured at the listening position better correlate. (Dirac Live generates filters for each channel).

I can toggle between having Dirac Live on and off. With it on, I perceive a fuller sound stage and better imaging when seated at the listening position. How much of this is due to the bass correction and how much is due to correction in higher frequencies I don't know, though. Maybe it is something I will investigate.

 

[goat76]

The "expert" advice to which you refer sounds more like a rule of thumb. In some rooms, a speaker with a flat frequency response can sound great without need for equalization above 500 Hz. But in other rooms, proper filtering will improve the sound.

My speakers have a fairly flat response (Elac Uni-Fi Reference UBR62). My room is a little tricky, though - tiled wall behind the speakers, tiled floor, large glass windows, etc. Thus, room reflections are not insignificant. Using Dirac Live to generate the room correction filters, which extend through audio frequency spectrum and are based on measurements made at and around the listening position, noticeably improved the sound of my system. Still, it is far from perfect.

Are you sure the corrections above 500 Hz are the real reason it sounds better? All the credits should maybe go to the corrections you made to the bass region.

Even Toole himself says that a room curve should not be used as a target curve.

 

[terryforsythe]

Are you sure the corrections above 500 Hz are the real reason it sounds better? All the credits should maybe go to the corrections you made to the bass region.

Even Toole himself says that a room curve should not be used as a target curve.

As noted, using Dirac Live I perceive better sound staging and better imaging. I suspect the bass filtering probably is the bigger contributor to the improved sound staging I perceive, but I suspect that getting the frequency response of the speakers in better correlation with each other as measured at the listening position might be the bigger contributor to the better imaging I perceive. But, I don't know for sure. I might investigate this. I can manually generate E.Q. settings to filter just the bass, and then compare that to the filters generated by Dirac Live.

 

[Thomas_A]

I agree with respect to the direct path, and the gradual drop due to the room reflections. But in my room, for example, the interference from the reflected sound at the listening position is not perfectly linear with respect to frequency - the speakers do not have a perfectly linear dispersion pattern, and the room reflections are not perfectly linear. Indeed, my speakers measure fairly flat in the near field, and are not too far off from each other. But, where they are placed in the room means they are at different distances from the staircase on one side and the windows on the other side. Moreover, the reflection coefficients of the two sides of the room are not the same. In consequence, the frequency responses of the respective speakers at the listening position differ greater than their near-field frequency responses. With the filters generated by Dirac Live, their frequency responses as measured at the listening position better correlate. (Dirac Live generates filters for each channel).

I can toggle between having Dirac Live on and off. With it on, I perceive a fuller sound stage and better imaging when seated at the listening position. How much of this is due to the bass correction and how much is due to correction in higher frequencies I don't know, though. Maybe it is something I will investigate.

If you have reflections that disturb the direct response above 500 Hz and within 2 ms you should act to remove or dampen those acoustically. This is especially if the disturbances arrive from the direction of the speakers. Did you do the measurements at LP with short time window or investigate the impulse window?

 

[terryforsythe]

If you have reflections that disturb the direct response above 500 Hz and within 2 ms you should act to remove or dampen those acoustically. This is especially if the disturbances arrive from the direction of the speakers. Did you do the measurements at LP with short time window or investigate the impulse window?

The boss (a.k.a., my wife) is not about to let me remove the coffee table. When listening late at night after she goes to bed, though, I throw a blanket over it. It helps. Nonetheless, the impulse response is not very good at the listening position. Dirac Live uses a swept sine wave. I don't recall whether it has a setting to set the measurement window. So with it I used the default settings.

I have measured the speakers with a short window using REW, but not at the listening position.

 

[Thomas_A]

The boss (a.k.a., my wife) is not about to let me remove the coffee table. When listening late at night after she goes to bed, though, I throw a blanket over it. It helps. Nonetheless, the impulse response is not very good at the listening position. Dirac Live uses a swept sine wave. I don't recall whether it has a setting to set the measurement window. So with it I used the default settings.

I have measured the speakers with a short window using REW, but not at the listening position.

Coffe table is not bad as it usually blocks the floor reflection. Can you see the first reflection, ie the speaker drivers, if you put a mirror on the table?

 

[terryforsythe]

Coffe table is not bad as it usually blocks the floor reflection. Can you see the first reflection, ie the speaker drivers, if you put a mirror on the table?

On right speaker, not the left. The coffee table is glass.

 

[Thomas_A]

On right speaker, not the left. The coffee table is glass.

Then you have a bit of asymmetry in the setup. If possible and with "permission" I would look if there is any way to alter position of things such that the coffee table blocks the floor reflection only and not act as first reflection from the speaker to ear. It seems that the right speaker is rather close to the coffee table to get such reflection?