Which frequency response is better?

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
I use a 0-degree calibration file from UMIK-1 which is advised for stereo measurements. Are you saying that 90 degree is a better and that UMIK-1 should point at the ceiling?
best regards, Miro

Personally I would stick with the UMIK advice. Obviously 90 degree vertical plus the 90 deg Cal File is the only way to measure surrounds. If the Mic and Cal Files are decent there will be very little difference either way.
Imperfect Omni mics, i.e. most of them, suffer from a HF boost when pointed directly at source. By turning away say 70 degrees this HF boost is softened to a pretty much flat response. The exact angle is different for different types of mic. But of course the Cal File overrides this. A 90 degree Mic will deliver flattest response with a 90 degree Cal file. There are finer points but in context this is not of importance IMO
 

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
Miro, I would like to refocus on your situation. I think all will concur on the distilled facts and perhaps the recommendations.
1 Your room is lively. This causes a lack of clarity. Absorptive treatment will help. Curtains. Cloud (can you do a cloud?) Carpet or rugs. Listening close to the speakers would also help but I doubt that possible?
2 Your LF would probably have lesser dips and peaks if the speakers were very far away from your Front Wall. But you have said the speakers have to remain pretty much where they are. As I said I would encourage you to play with them within the range available. This includes double actions such as kissing the Front Wall AND plugging the reflex ports with cotton wool or even a bung. Try it. As you said it can be hard to tell which is 'better'. For optimising LF response by tweaking locations of speakers and listener, I would use REW to give clear answers. We hope for a rising LF to say 30Hz or so, with as little peaking and nulling as possible. Note the wider bumps are much more audible than narrow ones.
3 I would like to add another tool to your kit here, which could address both issues, of Clarity and LF. Height. Try raising the speakers. I think recent versions of REW can actually measure Clarity, but for now, for you, I think ears will work fine to assess increases in Clarity and Imaging due to Treatment, and being closer to the speakers, and angling them in towards your ears, and raising height.
 
Last edited:

Miro

Registered
Thread Starter
Joined
Jul 2, 2020
Messages
17
hi all,

Thank you for helping. I did some measurements today and here is my current situation with pictures of my lsitening situation and frequency response (with 1/6 and no smoothing). The left speaker is 70cm from the front wall and 102cm from the side wall.

I am going to experiment and I am thinking to buy miniDSP DDRC-24 with Dirac Live, but the big question is, is it worth €500?

best regards, Miro
 

Attachments

  • Knipsel 2.JPG
    Knipsel 2.JPG
    165.8 KB · Views: 40
  • Knipsel.JPG
    Knipsel.JPG
    222.9 KB · Views: 48
  • 20200904_155457 (002).jpg
    20200904_155457 (002).jpg
    375.7 KB · Views: 48
  • 20200904_155438 (002).jpg
    20200904_155438 (002).jpg
    292.5 KB · Views: 42

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
You are welcome. Do experiment. Test Trumps Theory. Is that measurement of one speaker or both together? The rise of HF is a little bit alarming in any case. Continue using the Cal file which matches the way you are pointing the mic.
This has way more influence than your choice of 0 or 90. Cal files work!
I would always measure at least L, R, L+R. To be honest when fine tuning I sometimes take at least two sets of measurements, with the the microphone placed at ear positions.
REW can do arithmetic later, e.g. average the Left Ear pair, or average the Left and Right Ear responses measured with both speakers running.....
Don't get bogged down in detail, you are looking for clear indications A vs B. Minimise dips in your LF response by positioning.
A bit late in the day but your pictures are very informative. From the photos my immediate instinct says those speakers look close together. Generally optimum stereo is heard when an imaginary triangle is formed between the two speakers and a third apex some distance BEHIND your head, say 250mm or whatever sounds best. Obviously Width is now tied in with listening distance. But of course Width also affects LF.
Consider this process as a crash course in Air Traffic Control! Pun intended....

I can certainly steer your plane in to land in terms of optimising speaker and listener spots. But everything will improve dramatically if you install a Cloud, Big Curtain, Rug.
I would do those things first, then decide on DSP. I am a big fan of Dirac Live, and consider it the best in class. If your sound sources and path are digital, I would chose a simple Minidsp box with Digital In/Out.
In my studio and domestic streaming etc. I use the standalone DL app. No hardware.
If you are analogue of high quality I would be very slow to introduce the two extra stages of A/D and D/A.
500 bucks would go far with GIK Home Store and More and a GIK or RealTraps cloud. The benefits will apply across the board, no matter what equipment you may change to. It will also improve the tonality of conversation and other ambience. The 'clatter' and lack of clarity of a modern visual aesthetic is quite stressful. The opposite of a quiet outdoors ambience.
 
Last edited:

Miro

Registered
Thread Starter
Joined
Jul 2, 2020
Messages
17
You are welcome. Do experiment. Test Trumps Theory. Is that measurement of one speaker or both together? The rise of HF is a little bit alarming in any case. Continue using the Cal file which matches the way you are pointing the mic.
This has way more influence than your choice of 0 or 90. Cal files work!
I would always measure at least L, R, L+R. To be honest when fine tuning I sometimes take at least two sets of measurements, with the the microphone placed at ear positions.
REW can do arithmetic later, e.g. average the Left Ear pair, or average the Left and Right Ear responses measured with both speakers running.....
Don't get bogged down in detail, you are looking for clear indications A vs B. Minimise dips in your LF response by positioning.
A bit late in the day but your pictures are very informative. From the photos my immediate instinct says those speakers look close together. Generally optimum stereo is heard when an imaginary triangle is formed between the two speakers and a third apex some distance BEHIND your head, say 250mm or whatever sounds best. Obviously Width is now tied in with listening distance. But of course Width also affects LF.
Consider this process as a crash course in Air Traffic Control! Pun intended....

I can certainly steer your plane in to land in terms of optimising speaker and listener spots. But everything will improve dramatically if you install a Cloud, Big Curtain, Rug.
I would do those things first, then decide on DSP. I am a big fan of Dirac Live, and consider it the best in class. If your sound sources and path are digital, I would chose a simple Minidsp box with Digital In/Out.
In my studio and domestic streaming etc. I use the standalone DL app. No hardware.
If you are analogue of high quality I would be very slow to introduce the two extra stages of A/D and D/A.
500 bucks would go far with GIK Home Store and More and a GIK or RealTraps cloud. The benefits will apply across the board, no matter what equipment you may change to. It will also improve the tonality of conversation and other ambience. The 'clatter' and lack of clarity of a modern visual aesthetic is quite stressful. The opposite of a quiet outdoors ambience.

Hi DanDan,

What do you mean with Cloud?

I have added these 3 measurements for the same speaker placement as in the previous pictures. I do not understand these values, but I expect that they are important because of the name Clarity and reverbation time, but what are "normal" values? I expect that these values are bad.

best regards, Miro
Spectrogram.JPG
Clarity.JPG
RT60.JPG
 

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
You could post the REW file so that anyone here can open it and analyse it for you. In general though I recommend using measurement to help decide if A or B is better.

If you add treatment as everybody is recommending, then your RT will decrease, Clarity will increase. Here's a nice big cloud in a Classroom.
H
Cloud.JPG
 

JStewart

Senior AV Addict
Supporter
Joined
Dec 5, 2017
Messages
2,040
Location
Central FL
I am going to experiment and I am thinking to buy miniDSP DDRC-24 with Dirac Live, but the big question is, is it worth €500?

DanDan has given you good advice as to what to possibly do first, but yes Dirac Live is worth it. There is likely a free trial on their website. WARNING: Don't do the free trial if you're not prepared to make the purchase. Otherwise you may become dis-satisfied with your listening having heard it but not having it. :)

You could post the REW file so that anyone here can open it and analyse it for you.

Miro, To add to this tip there is a camera icon on every REW graph screen. Try it instead of taking screenshots. You'll see the benefits when you do.
 

JStewart

Senior AV Addict
Supporter
Joined
Dec 5, 2017
Messages
2,040
Location
Central FL
@Miro , do have access to a rug anywhere you can put on floor between you and the speakers just to get a listen to the change without spending money? Even a couple of blankets thrown down will give you a good idea of the effect.
 

Miro

Registered
Thread Starter
Joined
Jul 2, 2020
Messages
17
@Miro , do have access to a rug anywhere you can put on floor between you and the speakers just to get a listen to the change without spending money? Even a couple of blankets thrown down will give you a good idea of the effect.
Yes, I'll try, thank you.
 

Miro

Registered
Thread Starter
Joined
Jul 2, 2020
Messages
17
hi all,

Thank you for advise. here I have added the REW file with the data so If anybody could look at it, I'll be thankfull.

best regards, Miro
 

Attachments

  • Left speaker at 70cm from front wall and 102cm from side wall.mdat
    1.9 MB · Views: 16

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
I will leave that there hoping that others will have a look at it for you. But as I said, you have rising HF above 10K That looks wrong to me and suggests you might be using the wrong Cal File, or some other measurement error.
 

Miro

Registered
Thread Starter
Joined
Jul 2, 2020
Messages
17
I will leave that there hoping that others will have a look at it for you. But as I said, you have rising HF above 10K That looks wrong to me and suggests you might be using the wrong Cal File, or some other measurement error.
Hi DanDan,
Thank you very much for your help. I understand it much better now thanks to your advise.
Best regards, Miro
 

DanDan

Moderator
Staff member
Joined
Mar 10, 2018
Messages
743
You are welcome Miro. REW has many ways to view the same data. The BBC used to use third octave RT60 to 'qualify' their Control Rooms.
They required every 1/3 band be within 10% of it's neighbours. Their studios would be somewhat shorter, but they assumed the average domestic listener had about 400-500mS of room tone.
You are drowning in way more than that, plus you have a harsh peak
Room Decay.png
 
Last edited:
Top Bottom