Thread Rating:
  • 1 Vote(s) - 5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Sweet Room
#21
Thanks SonicLife and Ron.

I ordered a miniDSP mic but it hasn’t arrived yet.

So I guess the high level steps are:

1. connect my MacBook to Devialet(Devialet play in two channels stereo mode)
2. connect the mic to MacBook
3. move the mic like the video shown and click on record button in the REW software until the line settled down
4. then I get a graphic result in REW
5. then I look at those graphic result, if there is a drop let say at 60Hz for about 20db, then In the text file somewhere I do a +20db, if I see there is a up in the graphic somewhere for 10db, then in the text file I should do a -10db to pull it down?
6. Save the txt file into SD card and then Devialet expert pro will use it.

Am I correct in above steps? Do I need to do left channel and right channel individually or it’s better let two channel run together to get a overall result?

In the video they use a black box equipment, I guess I don’t need that?

And also in the initial setup in REW, he chooses 48KHz, should I choose the same option?

Thanks a lot for helping me to understand how this works.
Reply
#22
That should be correct.


Gesendet von iPad mit Tapatalk
Reply
#23
Re.: your point 5. You should be careful with positive gains. It will lead to digital clipping. Look for peaks (room modes) in the frequency plot (under 200Hz) and reduce these with Sweet Room. If you find you have to raise an area (i.e. use + gain) the <"dB_GAIN_>  for both channels should have the same, but negative value.
*
Devialetless!
Roon, ROCK/Audiolense XO/Music on NAS/EtherRegen/RoPieee/USPCB/ISORegen/USPCB/Sound Devices USBPre2/Tannoy GOLD 8
250 Pro CI, MicroRendu(1.4), Mutec MC-3+USB
Reply
#24
(17-Oct-2020, 12:56)ogs Wrote: Re.: your point 5. You should be careful with positive gains. It will lead to digital clipping. Look for peaks (room modes) in the frequency plot (under 200Hz) and reduce these with Sweet Room. If you find you have to raise an area (i.e. use + gain) the <"dB_GAIN_>  for both channels should have the same, but negative value.


Hi What do you mean by : If you find you have to raise an area (i.e. use + gain) the <"dB_GAIN_>  for both channels should have the same, but negative value.

I am not quite getting this.

So in the text file we try to use negative number rather than positive to avoid clipping?
Reply
#25
Since the Devialet does all its signal processing using floating point, digital clipping is not the real issue.

However boosts shall be handled with really extreme precautions because you will rapidly hit the maximum power of your amplifier and then trigger soft clipping.

With a boost of 20dB like mentioned by @PeppaPig , it would mean that if you are listening with the volume above -20dB you could enter in soft clipping.

So the safer approach is the one recommended by @ogs to only reduce and avoid boosting.
That means that instead of flattening the curve around the middle you flatten it around the bottom.
You will need to listen with the volume knob turned higher, but that will be much safer.

Jean-Marie
MacBook Air M2 -> RAAT/Air -> WiFi -> PLC -> Ethernet -> Devialet 220pro with Core Infinity (upgraded from 120) -> AperturA Armonia
France
Reply
#26
(17-Oct-2020, 01:36)PeppaPig Wrote: Thanks SonicLife and Ron.

I ordered a miniDSP mic but it hasn’t arrived yet.

So I guess the high level steps are:

1. connect my MacBook to Devialet(Devialet play in two channels stereo mode)
2. connect the mic to MacBook
3. move the mic like the video shown and click on record button in the REW software until the line settled down
4. then I get a graphic result in REW
5. then I look at those graphic result, if there is a drop let say at 60Hz for about 20db, then In the text file somewhere I do a +20db, if I see there is a up in the graphic somewhere for 10db, then in the text file I should do a -10db to pull it down?
6. Save the txt file into SD card and then Devialet expert pro will use it.

Am I correct in above steps? Do I need to do left channel and right channel individually or it’s better let two channel run together to get a overall result?

In the video they use a black box equipment, I guess I don’t need that?

And also in the initial setup in REW, he chooses 48KHz, should I choose the same option?

Thanks a lot for helping me to understand how this works.

Seems about right. It's worth having a play using the mic in your laptop, it won't be perfect but will give you a change to play with the software, which is the hard part, there is a lot to get your head around.
Roon, Rega P9 + Dynavector XX2Mk2 > 440 Pro > Sonus Faber Guarneri Evolution
Reply
#27
(17-Oct-2020, 15:05)Jean-Marie Wrote: Since the Devialet does all its signal processing using floating point, digital clipping is not the real issue.

However boosts shall be handled with really extreme precautions because you will rapidly hit the maximum power of your amplifier and then trigger soft clipping.

With a boost of 20dB like mentioned by @PeppaPig , it would mean that if you are listening with the volume above -20dB you could enter in soft clipping.

So the safer approach is the one recommended by @ogs to only reduce and avoid boosting.
That means that instead of flattening the curve around the middle you flatten it around the bottom.
You will need to listen with the volume knob turned higher, but that will be much safer.

Jean-Marie

I agree with the above but there is another reason to avoid trying to correct for dips in the room response in the bass frequencies.

Dips in the room response at bass frequencies occur in locations which are a null for a room mode ("standing wave"). It doesn't matter how much  EQ you apply to the signal at that frequency, there will always be a null in the room response at that position of the room. It's effectively impossible to fill a null because it's caused by the sound waves cancelling each other at that location in the room. If you have a problem null in your room response at the listening position, the only ways to effectively deal with it are to move the speaker positions and/or the listening position, or to physically deal with the room mode. You can eliminate peaks in the response with cuts in EQ at those frequencies but you cannot eliminate dips in the response which result from room modes.

While the nulls will always be found in the same positions in a room, moving the speaker positions can change how strongly the speakers activate load the room mode so if you can move the speakers to a location where they do not activate the mode as strongly, then the dip in the bass at the null points in the room will not be as severe. If you move the listening position to an area outside the location of a null point, the reduction in level at that frequency will also not be as severe.

The alternative is physical treatments in the form of bass traps.

Correction techniques such as Sweet Room are often called "room correction" but  they do not actually correct the room, in fact they do not change the room or its behaviour in any way. What they do is to compensate for the behaviour of the room by applying an opposite alteration to the musical signal so if you have a peak in the bass they apply a cut and if you have a dip in the bass they correct with a boost. In my view what they do should be called "room compensation" rather than "room correction". On the other hand physical room treatments actually do change the room response, they effectively become a part of the room and the changes they make in the room response can actually smooth out the peaks and dips in the room response at bass frequencies as well as doing other things. Physical treatments actually can correct the room. That doesn't mean that physical corrections are automatically better than electronic corrections, they have different capabilities and strengths. Dips in the room's response at bass frequencies is one problem where physical solutions (moving the speakers and/or listening position, or using bass traps) are better than electronic corrections. On the other hand, if you have a dip in the response at bass frequencies which is due to the speaker rather than the room, physical corrections won't help but electronic correction can help. You need to pick the right solution for the problem and there is no one single approach, whether that be electronic correction or physical correction. that can solve every problem.
Roon Nucleus+, Devilalet Expert 140 Pro CI, Focal Sopra 2, PS Audio P12, Keces P8 LPS, Uptone Audio EtherREGEN with optical fibre link to my router, Shunyata Alpha NR and Sigma NR power cables, Shunyata Sigma ethernet cables, Shunyata Alpha V2 speaker cables, Grand Prix Audio Monaco rack, RealTRAPS acoustic treatment.

Brisbane, Qld, Australia
Reply
#28
I cannot agree more with @David A !

Jean-Marie
MacBook Air M2 -> RAAT/Air -> WiFi -> PLC -> Ethernet -> Devialet 220pro with Core Infinity (upgraded from 120) -> AperturA Armonia
France
Reply
#29
(17-Oct-2020, 21:12)David A Wrote:
(17-Oct-2020, 15:05)Jean-Marie Wrote: Since the Devialet does all its signal processing using floating point, digital clipping is not the real issue.

However boosts shall be handled with really extreme precautions because you will rapidly hit the maximum power of your amplifier and then trigger soft clipping.

With a boost of 20dB like mentioned by @PeppaPig , it would mean that if you are listening with the volume above -20dB you could enter in soft clipping.

So the safer approach is the one recommended by @ogs to only reduce and avoid boosting.
That means that instead of flattening the curve around the middle you flatten it around the bottom.
You will need to listen with the volume knob turned higher, but that will be much safer.

Jean-Marie

I agree with the above but there is another reason to avoid trying to correct for dips in the room response in the bass frequencies.

Dips in the room response at bass frequencies occur in locations which are a null for a room mode ("standing wave"). It doesn't matter how much  EQ you apply to the signal at that frequency, there will always be a null in the room response at that position of the room. It's effectively impossible to fill a null because it's caused by the sound waves cancelling each other at that location in the room. If you have a problem null in your room response at the listening position, the only ways to effectively deal with it are to move the speaker positions and/or the listening position, or to physically deal with the room mode. You can eliminate peaks in the response with cuts in EQ at those frequencies but you cannot eliminate dips in the response which result from room modes.

While the nulls will always be found in the same positions in a room, moving the speaker positions can change how strongly the speakers activate load the room mode so if you can move the speakers to a location where they do not activate the mode as strongly, then the dip in the bass at the null points in the room will not be as severe. If you move the listening position to an area outside the location of a null point, the reduction in level at that frequency will also not be as severe.

The alternative is physical treatments in the form of bass traps.

Correction techniques such as Sweet Room are often called "room correction" but  they do not actually correct the room, in fact they do not change the room or its behaviour in any way. What they do is to compensate for the behaviour of the room by applying an opposite alteration to the musical signal so if you have a peak in the bass they apply a cut and if you have a dip in the bass they correct with a boost. In my view what they do should be called "room compensation" rather than "room correction". On the other hand physical room treatments actually do change the room response, they effectively become a part of the room and the changes they make in the room response can actually smooth out the peaks and dips in the room response at bass frequencies as well as doing other things. Physical treatments actually can correct the room. That doesn't mean that physical corrections are automatically better than electronic corrections, they have different capabilities and strengths. Dips in the room's response at bass frequencies is one problem where physical solutions (moving the speakers and/or listening position, or using bass traps) are better than electronic corrections. On the other hand, if you have a dip in the response at bass frequencies which is due to the speaker rather than the room, physical corrections won't help but electronic correction can help. You need to pick the right solution for the problem and there is no one single approach, whether that be electronic correction or physical correction. that can solve every problem.


Very Good comment!
Reply
#30
I am very satisfied with Sweet room. Maybe the manual is not so custom friendly but it works!
For measurement I used: 
XTZ Room Analyzer II/II Pro
I only corrected for room resonances below 250 HZ. Found 3 resonaces in each channel which are corrected. Now I have a a straight (+- 2.5 dB) respons from 31 to 20 000 Hz. Very impressive result for such a small speaker as the Dynaudio C1/II. I have tried SAM but did not like the
.txt   eq.txt (Size: 669 bytes / Downloads: 37)  result. Impressive first but not in the long run.
I attach my  eq.txt file.
Expert Pro 220. Dynaudio C1 II. Rega P8. Lyra Delos. Primare BD 32. 
Reply


Forum Jump:


Users browsing this thread: 2 Guest(s)