REW questions

[Boxerfan88]

This does look really high, how loud was the sweep done, and what measurement distance?

 

Sent from my X9009 using Tapatalk

 

Sweep level @ -20dBFS. Mic placed at about 2m from speaker front (MLP).

 

Yeah, I went to read up the REW distortion FAQ, and I will have to re-measure it. The FAQ suggest using longer sweeps (1M) or stepped sine (I have no clue what this is). Going to experiment this weekend.

 

Might be hardware related too...

[Boxerfan88]

4% is high and unusual for frequencies from the main speakers. Especially if the distortion level is across the whole audio spectrum.

 

It should way down to 0.1% or less, even with room acoustics. You will be able to hear 1%, and 4% is quite bad.

 

There is a high chance of improper setup of REW. What hardware are you using to measure?

 

Yeah, I think so too. I went to read up the REW distortion FAQ, and I will have to re-measure it. The FAQ suggest using longer sweeps (1M) or stepped sine (I have no clue what this is). Going to experiment this weekend.

 

Might be hardware related too...I have not yet purchased the UMIK-1, instead I use the Yamaha AVR mic to PC for measurement.

 

Still learning REW...it's a very powerful tool...

[Jag]

AVR mic to pc will result in noisy and distorted measurements.

 

I can loan you a mic if you are interested. Get a umik instead.

[Boxerfan88]

AVR mic to pc will result in noisy and distorted measurements.

 

I can loan you a mic if you are interested. Get a umik instead.

 

Yes, I am. I'll PM you. Thank you.

[ronildoq]

For my room, I’ve never seen the RT60 graph look this good.

 

Before

 

 

After

 

 

 

Sent from my iPhone using Tapatalk

 

Anything between 200-500 ms decay times are good, 300-350 is a good indication, ideally if they have even decay times across the range in a flat target. 1/3 octave view will tell u more. Ideally for bigger living rooms , 350-400ms decay time is acceptable range.having said that, rt60 doesn’t really tell u the quality of the sound in the room. These type of measurements are more suited for auditoriums

 

I would instead, look at the energy time curves, look for the early reflection points and solve those. Then I’ll look at all other resonances in the room. Far more meaningful and fruitful exercise

[Boxerfan88]

Anything between 200-500 ms decay times are good, 300-350 is a good indication, ideally if they have even decay times across the range in a flat target. 1/3 octave view will tell u more. Ideally for bigger living rooms , 350-400ms decay time is acceptable range.having said that, rt60 doesn’t really tell u the quality of the sound in the room. These type of measurements are more suited for auditoriums

 

I would instead, look at the energy time curves, look for the early reflection points and solve those. Then I’ll look at all other resonances in the room. Far more meaningful and fruitful exercise

 

Thank you for sharing. I went to re-read the RT60 FAQ, and my “after EQ” plot is along the lines described in the FAQ. [emoji3]

 

For domestic listening rooms and recording studios with volumes of less than 50 cubic metres (1,800 cubic feet) the recommended RT60 value is 0.3 s. For larger rooms, up to 200 cubic metres (7,000 cubic feet) the recommendation is 0.4 to 0.6 s. In both cases the value should be fairly uniform across the frequency range, though it will typically tend to increase at lower frequencies.

 

I’ll post the ETC curves later for comments. Thanks again.

 

Sent from my iPhone using Tapatalk

[Boxerfan88]

AVR mic to pc will result in noisy and distorted measurements.

 

I can loan you a mic if you are interested. Get a umik instead.

 

Hi Jag,

 

Thanks for loaning me the UMM6. You are absolutely correct, the AVR mic is noisy and gives incorrect readings.

 

Using your UMM6, the measured THD is below 1%.

 

=> Light red (UMM6)

=> Dark red (AVR mic)

 

 

 

Sent from my iPhone using Tapatalk

[Boxerfan88]

Noise floor (no signal) as measured by UMM6. I wonder why those side bands (1k, 2k, 3k, etc..) are showing up? Is it caused by the internal ADC?

 

 

 

 

Sent from my iPhone using Tapatalk

[Boxerfan88]

THD(2..9) < 1% sweet!

 

 

 

Sent from my iPhone using Tapatalk

[Boxerfan88]

I would instead, look at the energy time curves, look for the early reflection points and solve those. Then I’ll look at all other resonances in the room. Far more meaningful and fruitful exercise

 

Analyze it like this?

 

 

Sent from my iPhone using Tapatalk

[ronildoq]

I use REW to analyse it like this

 

 

This is my right speaker only measurement, before any EQ. My main goal is to look at the early reflection points.

 

You can see that there isn’t a smooth decay of spikes especially at 8.47ms, there is one particular reflection in the room that is causing this early reflection energy to MLP. This reflective area is exactly 2.931 meters + 3meter (direct speaker distance to MLP ) =6.931m . I believe it’s the reflection off from ceiling to the MLP

 

Other areas that need attention for my right speaker is 4ms, 5.5ms , 13ms, 14.75ms and 26ms. U will notice these are all above -20db threshold. A good sounding room will have spikes that are down by 18-20db within the first 20-40ms window. Look out for those and treat those area.

 

Remember it’s not so important that is down by 20db, what’s more important is eliminating the peaks. The decay pattern may or may not be continuous. If u use diffusor, the spectral energy from the diffusors will show a non continuous decay time, that is perfectly alright , we still have energy going into the room

 

Next I’ll look at uniformity between the graphs of individual speakers. Both speakers including the centre channel should be as close as possible when you overlay the graphs. The impulse should look uniform when overlay them together, and not like a ripple effect up down. So even if there is coloration, they are all identical ... so u will still get that timbral balance from all speakers

 

From there you will know it will sound very good.

 

I’m still waiting for the SMT wings... very excited

 

 

Sent from my iPhone using Tapatalk

[ronildoq]

Hi Jag,

 

Thanks for loaning me the UMM6. You are absolutely correct, the AVR mic is noisy and gives incorrect readings.

 

Using your UMM6, the measured THD is below 1%.

 

=> Light red (UMM6)

=> Dark red (AVR mic)

 

 

 

 

Sent from my iPhone using Tapatalk

 

Glad you have sorted it out. In another case scenario, we discovered that sometimes the sensitivity file from the mic could be a problem as well. U might want to check that you have the correct sens factor file loaded for the microphone , that could be the cause

 

In any case, the USD75 UMiK 1 mic is one the best tools / investment I’ve ever made. It’s almost a must have equipment for me. Helped me a great deal in understanding my system and equipment, and room. The best price to performance ratio tool ever !

[Boxerfan88]

Glad you have sorted it out. In another case scenario, we discovered that sometimes the sensitivity file from the mic could be a problem as well. U might want to check that you have the correct sens factor file loaded for the microphone , that could be the cause

 

In any case, the USD75 UMiK 1 mic is one the best tools / investment I’ve ever made. It’s almost a must have equipment for me. Helped me a great deal in understanding my system and equipment, and room. The best price to performance ratio tool ever !

 

I used the cal file that Jag sent to me, I think it should be ok. Yes I agree, I need the UMIK-1, and plan to get it very soon. Last night I did another run at distortion (to measure L/R separately), lo and behold, distortion levels are even lower!! I was hearing nice clear sound but my original measurement made me doubt my own hearing...  :)

 

I see THD(2..9) of <0.3% for individual channels, so the learning for me was that L+R sound will interact in the room and cause more distortion (common sense to me). Image sequence: L, R, L+R.

 

 

BTW, I have 2pcs MOAB placed at front-side wall corners. It’s a hifi/HT room cum study therefore have work-desk & bookshelves as diffusers. Only untreated area is my ceiling. No more space for bulky stuff that’s why I am going after EQ for my stereo music.

[Boxerfan88]

I use REW to analyse it like this

 

 

This is my right speaker only measurement, before any EQ. My main goal is to look at the early reflection points.

 

You can see that there isn’t a smooth decay of spikes especially at 8.47ms, there is one particular reflection in the room that is causing this early reflection energy to MLP. This reflective area is exactly 2.931 meters + 3meter (direct speaker distance to MLP ) =6.931m . I believe it’s the reflection off from ceiling to the MLP

 

Other areas that need attention for my right speaker is 4ms, 5.5ms , 13ms, 14.75ms and 26ms. U will notice these are all above -20db threshold. A good sounding room will have spikes that are down by 18-20db within the first 20-40ms window. Look out for those and treat those area.

 

Remember it’s not so important that is down by 20db, what’s more important is eliminating the peaks. The decay pattern may or may not be continuous. If u use diffusor, the spectral energy from the diffusors will show a non continuous decay time, that is perfectly alright , we still have energy going into the room

 

Next I’ll look at uniformity between the graphs of individual speakers. Both speakers including the centre channel should be as close as possible when you overlay the graphs. The impulse should look uniform when overlay them together, and not like a ripple effect up down. So even if there is coloration, they are all identical ... so u will still get that timbral balance from all speakers

 

From there you will know it will sound very good.

 

I’m still waiting for the SMT wings... very excited

 

 

Sent from my iPhone using Tapatalk

 

 

 

Thank you for sharing. I am learning lots from you. You brought up a very good point, look at L/R separately as they may interact with the room differently (based on where they are placed). I will do so, and share my observations. Thanks once again!

[ronildoq]

 

 

Thank you for sharing. I am learning lots from you. You brought up a very good point, look at L/R separately as they may interact with the room differently (based on where they are placed). I will do so, and share my observations. Thanks once again!

 

I still have a lot of work to do as well. Now just capturing data before the smt wings arrive, then I can compare the before and after effect. I prefer to use physics and tools behind why and how I place them. Then I’ll proceed to tweak to preference.

 

The most significant placement for diffusors I find is at the back of MLP. That would be the no 1 spot I’ll go for.

 

Ceiling will help, but not significant enough, because our ears decipher sound better on the horizontal plane. Mounting them on the wall instead, brings far more significant impact

 

For bass trapping, dihedral mount is good, leaving air gap at the back to trap longer waves

 

Same like you, i still have lotsa work to do to improve the system further

[Boxerfan88]

Just sharing my analysis... I was wondering about those peaks @ 8.2m, 9.0m (my room is only 4m * 4m), those peaks must be due to complex tangential and oblique modes.

 

0.55m = L speaker, front wall

0.8m = R speaker, side wall

1.5m = L speaker, ? ?

2.1m = L+R speaker, floor bounce

4.6m = L+R speaker, rear wall bounce

5.1m = ? ?

8.2m = ? ?

9.0m = ? ?

 

[ronildoq]

Just sharing my analysis... I was wondering about those peaks @ 8.2m, 9.0m (my room is only 4m * 4m), those peaks must be due to complex tangential and oblique modes.

 

0.55m = L speaker, front wall

0.8m = R speaker, side wall

1.5m = L speaker, ? ?

2.1m = L+R speaker, floor bounce

4.6m = L+R speaker, rear wall bounce

5.1m = ? ?

8.2m = ? ?

9.0m = ? ?

 

 

Those farthest away could be the reflections behind MLP , 9m + the distance from speaker to mic, add it up. Use a string this length, paste it near mid driver of speaker, then see which reflected area is causing it. Temporary move your Moab to hat spot and measure again see if it’s gone

[Boxerfan88]

Hmm... about the 9m peak, I still don’t really understand where it is coming from.

 

Let me explain, the room is 4m length, 4m width. MLP is 1m from back-wall (or 3m from front-wall).

 

For a 9m peak, it would have bounced around the room, not bounced off one reflection point. No?

 

 

Sent from my iPhone using Tapatalk

[ronildoq]

The ones that are less than 2ms could be those near your mic area, or MLP, the direct energy hitting a chair then bouncing off and being captured by the nearby mic

 

The ones that are further, could be energy radiating off the front wall, reflected to back wall, then to MLP .

 

When you kill the reflection energy  from the front wall, there is nothing to reflect to the back wall.

 

So energy from direct speakers to back wall, behind your MLP , place the panels there. Then try measure again see if the reflections are gone

 

So front walls behind mains, back wall behind MLP , place the panels there , measure again

 

The ones 6m so could be side walls

[ronildoq]

Hmm... about the 9m peak, I still don’t really understand where it is coming from.

 

Let me explain, the room is 4m length, 4m width. MLP is 1m from back-wall (or 3m from front-wall).

 

For a 9m peak, it would have bounced around the room, not bounced off one reflection point. No?

 

 

Sent from my iPhone using Tapatalk

 

It is definitely bouncing off a second reflective area. That’s why the goal is to kill off the first reflection point. When u kill off the first, there will not be a second

 

Toe in of the speakers will affect that as well

[ronildoq]

Remember the goal is to tame down the peaks and notches , all related to first reflection points. Once u have achieved that,you will find that the system and sound quality in the room will be much much more satisfying, it will bring about quality to sound in that room ...

 

That is why we need tools to help us. Otherwise you could be spending years and still don’t know what to do or how to improve the system

[Boxerfan88]

It is definitely bouncing off a second reflective area. That’s why the goal is to kill off the first reflection point. When u kill off the first, there will not be a second

 

Toe in of the speakers will affect that as well

 

Excellent point! Kill off the first reflection, and the secondary reflections will sort themselves out. [emoji106]

 

About toe, my L/R speakers are toed in a bit (maybe about 15*). Can help explain a bit more what effect does it have?

 

 

Sent from my iPhone using Tapatalk

[ronildoq]

Excellent point! Kill off the first reflection, and the secondary reflections will sort themselves out. [emoji106]

 

About toe, my L/R speakers are toed in a bit (maybe about 15*). Can help explain a bit more what effect does it have?

 

 

Sent from my iPhone using Tapatalk

 

The point with toe in is that you will need to know the degree of the toe in, and because the angle of incidence will always = angle of reflection, you will be able to determine the first reflection point.

 

Put it this way, if you have a sun light shining on a mirror, when u angle the mirror up down left right, u will see it’s reflections of sunlight on the wall according to how u angle that mirror.... it’s eccatly the same concept....  use this concept to find that first relflection point.... another way to look at it is billiard or snooker. Before taking your shot, if u angle the cue stick 45 degrees and hit it, then the ball will bounce off 45 degrees off the table.... something like that....

 

For rew, use the string method, establish the distance of the reflection, the area where it bounces off a solid surface, is the critical area that causes the reflection, put your panels there and voila !

[Guest BadEnglish]

...

About toe, my L/R speakers are toed in a bit (maybe about 15*). Can help explain a bit more what effect does it have?

 

 

Sent from my iPhone using Tapatalk

 

Are your speakers designed to toe-in ?

[Boxerfan88]

Are your speakers designed to toe-in ?

 

Honestly I donno. Mine are Usher MD2.

 

I didn’t like straight fire, coz the soundstage was wide but a tad fuzzy. I toed in a little to lock in the center image better, and with it came along more precise imaging.

 

 

Sent from my iPhone using Tapatalk