Many different things regarding my setup. advice/help take a look please :)

[UKSPAWN]

Hello,

First off thanks to the REW creator and everyone who contributes. I will be donating when i can (when system is complete)

And now to my reason for posting here.

I am looking for some help here and there and have several questions about REW and measurments i have taken, how best to do certain things etc etc.
So instead of creating 5 different topics i thought id just create one and discuss each thing here. I hope thats fine.

Ok so to kick off a few simple questions about delay and accuracy.

Lets say I measure the Left Channel (L) and also use the L as the timing ref, then there should be 0 delay right? so the number that comes out just happens to be almost exactly the distnace from the speaker to the mic. is this correct? its a Negative (-) value.

So can i then use this as a timing offset and that makes it T=0 then everything else can be relative to 0? or should i not add the offset?
Secondly how accurate is it, i seem to get varying results from the same speaker over 5 or more runs. is that normal?
What decimal place would be acceptable for the L/R channels, currently its 0.0050ms or therabouts. is that good enough?
If i try to move a speaker by a mil, then do the same measure then sometimes it will tell me iv gone too far the other way.. so im guessing thats good enough?

nextingly...
Time alignment or phase alignment.. impulse alignment etc etc?

What should i be doing to get the sound of explosions etc from my subs to all kick me in the chest at the same time?
I have heard that doing a time alignment should be done in the time domain using impulses, but the way i saw it done was in a different program, is this able to be done in REW?

Is time alignment and phase alignment a seperate thing, do you choose one or the other or do both etc.

I should phase align at the crossover Freq right? but i also want the initial energy to arrive at the same time, im confused isn't that 2 differnt delays?!

For now thats all i can think off, i have more questions but i forgot them untill i go to do something again i will leave it at that.
Feel free to answer any or all or ask more questions or just say hello!



Oh, system... i have a MiniDSP UMIK1 and just ordered a MiniDPS 2x4HD

My system is a 2 channel, SMSL SU9 DAC > Rotel A12 > Monitor Audio Bronze 5g speakers an SVS SB1000Pro 12" subwoofer (from preouts) and a Gale 4070 8" subwoofer (high level speaker connection until mini DSP arrives.)
.
until the DSP arrives i have just been measuring different speaker and sub locations for the best placment for avoiding nulls/peaks and tuning levels/ subwoofer EQ etc etc.

I have been told to remove the gale from the system, but it has sentimental value and i also like the sound signature of it, and i have it so why not use it?! right!

I plan to connect the DSP this way.
PC USB > SMSL DAC (Variable level as VOLUME) > MiniDSP 2x4 > Output 1+2 = L/R > Rotel AMP (Fixed level eg as a power amp)
------------------------------------------------------------------------------------------ > Output 3 /4 = Sub 1 / 2...

Does this sound about right? I plan to cut the lower freq from the Main L/R channels, delay them to align with the sub(s) and delay the "faster" sub so they all align to the mains. then do Low pass / high pass filters on the subs (bass protection below resonance etc) / crossovers and finally DSP on the subs and *maybe* the mains upto about 500hz only if really needed.

Thanks for looking and any info.

SPAWN

 

[John Mulcahy]

Lets say I measure the Left Channel (L) and also use the L as the timing ref, then there should be 0 delay right? so the number that comes out just happens to be almost exactly the distnace from the speaker to the mic. is this correct? its a Negative (-) value.

The delay should be close to zero, but is typically slightly different as the overall delay of a speaker's impulse response can differ from the delay of the timing signal, which is reproduced mainly by the tweeter.

So can i then use this as a timing offset and that makes it T=0 then everything else can be relative to 0? or should i not add the offset?

Up to you whether you update the timing offset with the measured delay, it should be very small anyway.

Secondly how accurate is it, i seem to get varying results from the same speaker over 5 or more runs. is that normal?

Normally within about 10us

What decimal place would be acceptable for the L/R channels, currently its 0.0050ms or therabouts. is that good enough?

That corresponds to about 1.7 mm, can you keep your head that still?

What should i be doing to get the sound of explosions etc from my subs to all kick me in the chest at the same time?

Some of the sound will be from your subs, some from the mains, best option is usually to aim for a good phase alignment between the subs and between subs and mains. REW has an alignment tool in the All SPL graph controls to help with that.

 

[UKSPAWN]

thanks for the reply.

Some of the sound will be from your subs, some from the mains, best option is usually to aim for a good phase alignment between the subs and between subs and mains. REW has an alignment tool in the All SPL graph controls to help with that.

okay, but this is what i dont understand, if i use the phase alignment it gives me a delay right. so i add that delay to sub? or speaker?
But doesn't that then mean iv just upset all the timing iv done from the delay the measurments gave me?

my current delays are
12" SVS sub = 0 (its the 'slowest')
8" sub 5.9ms
front left 7.93ms
front right 7.65ms

and the nwhen i run any measurments it tells me everything is within 0.0050ms .. so thats pretty ok.

But if i then say, alighn the 8" sub to my mains, iv just added delay to the sub, meaning when i take a measurment again its all out of timing again? then i have to do the phase with the big sub, meaning thats now out of time again....

im confused. do i reset everything to 0 and then do the phase align? but then when i add delay to time align isnt that going to mess the phase again? ... seems like an infinite loop lol

Im sure im just misunderstanding something... i have been reading a lot but its getting so my head is hurting and id be greatful if someone could help simplify it for me.

 

[sm52]

Read it, maybe it will help. https://www.avnirvana.com/threads/n...someone-who-understands-a-tittle-german.8730/

 

[UKSPAWN]

Read it, maybe it will help....

argh, thank you but no that didnt help. It looks like that user is more familiar with it. i dont even know what FDW is and how to use that etc...

Iv watched a few youtube videos too and everyone seems to do things differently and some im sure are doing it wrong so i dont want to learn it all wrong, id like to start to learn the correct way..

I will try again tomorrow. ive been at it all day now and i feel it sounds nice and it measures better than before too but i dont think its optimal by any means.

All i did was measure everything seperate and do the delays, then do some EQ on each of them seperate from MLP .

i will take some sweeps and post them and see if anyone can step-by-step me and help me learn the process correctly.

im just not getting what i should be doing in a basic way and i dont understasnd some terminoligy and processes.
I know im not dumb maybe im overcomplicating it but i cant learn well from reading im very much visual.

Thanks for tyring guys ill be back another day and post some things after reading some more.

 

[sm52]

I don't have a system like you or those guys. But, as I understand it, first you need to get the measurement of each speaker separately. Then, with the help of REW, align the two subwoofers in phase and in level. Then add up the measurements of the two subs and get one. It's like you have one sub. Then calculate the delays for the other speakers and add their measurements together with the delays. And now you are left with two measurements - the sum of the subs and the sum of the other speakers. They also need to be coordinated with each other. One of those guys, jtalden, is definitely good at this. If you manage to contact him, he will explain everything as it should.

 

[UKSPAWN]

yeah im hoping someone can take a look at my measurments and help.
im starting to get the hang of putting the mic in the right position and getting consistent results even days later.

my problem now is i need to start labeling stuff well so i can tell whats what when i come back to the file in a few days or whatever and they are all called 1 2 3 4 lol.. so yeh. i think i will delete all my saved files so far and start over and label everything well.

i also have a config in my MINI DSP that i feel sounds nice to me. i mean its probably trash but it sounds better to me than before i had the device. so atleast i can start on config 2 and try stuff and just revert back to my config 1 if i don't like it.

thanks for the reply, and what is your system?

Cheers,

 

[jtalden]

I can do the REW alignment tool analysis for you if you like. As you discovered, it is not a method that is easy to learn effectively without resonable understanding and experience with acoustic analysis. There are much easier methods to determine delays that usually provide favorable delay settings. This method just provides a better understanding of what the room effects are in the XO range and shows the impact of delay changes virtually. I understand you have already determined delays, but are not sure they are favorable.

Step 1
let's first confirm your measurements are repeatable and normal when using REW acoustic timing by taking 6 full-range measurements of the right speaker using the left speaker as the reference speaker. Post the mdat so I can confirm that the measurement system is working well.

Step 2
Measurement Setup

• Mic positioned at LP
• Apply the desired XO filter settings to all speakers
• Leave all delay setting as currently set. [Any needed changes from the current settings will be determined by the analysis.]
• Use full-range sweeps, e.g., 20-20k Hz for all measurements.
• Use REW acoustic timing with the left channel as the reference channel.

Measure

1. L
2. R
3. SW8
4. SW12

You will need to post your actual XO settings so that I can understand your setup. It is not clear to me if this is a 2-way, 3-way or a hybrid of the 2. I will then confirm your settings or suggest new ones. I will also try to respond to your future questions.

 

[UKSPAWN]

I can do the REW alignment tool analysis for you if you like. As you discovered, it is not a method that is easy to learn effectively without resonable understanding and experience with acoustic analysis. There are much easier methods to determine delays that usually provide favorable delay settings. This method just provides a better understanding of what the room effects are in the XO range and shows the impact of delay changes virtually. I understand you have already determined delays, but are not sure they are favorable.

Step 1
let's first confirm your measurements are repeatable and normal when using REW acoustic timing by taking 6 full-range measurements of the right speaker using the left speaker as the reference speaker. Post the mdat so I can confirm that the measurement system is working well.

Step 2
Measurement Setup

• Mic positioned at LP
• Apply the desired XO filter settings to all speakers
• Leave all delay setting as currently set. [Any needed changes from the current settings will be determined by the analysis.]
• Use full-range sweeps, e.g., 20-20k Hz for all measurements.
• Use REW acoustic timing with the left channel as the reference channel.

Measure

1. L
2. R
3. SW8
4. SW12

You will need to post your actual XO settings so that I can understand your setup. It is not clear to me if this is a 2-way, 3-way or a hybrid of the 2. I will then confirm your settings or suggest new ones. I will also try to respond to your future questions.

Hey thanks that sounds great. i will do them over the weekend and post the results.
im pretty sure i can get all those measurments no problem.

I have been playing with it again today but i feel sometimes im messing around more than enjoying music so iv put it to bed for today. I did try using the alignment tool again but it didnt sound or measure right when i used those delays.

also I have noticed the timing delays given change depending on the sweep. if a 256k and then compare the left right timings, and then do a 1m and compare the timings it gives different results. i assume its fine as long as you stick to the same length while setting delays right?


...and iv been playing with the crossovers but im doing something maybe unconventional and maybe considered wrong but i want the smaller 8" sub to play along with the mains longer and not crossover at the same point. i know thats probably bad in some way but the extra mid-bass im getting from doing that is really nice and i dont want to cut out everything below 100hz on my mains as they play pretty good to ~50Hz.. . anyway. iv saved my config now so i can try more traditional crossovers and see what i prefer.

anyway, thans again for offering to help me, i will let you know when iv taken the measurments you asked for.

cheers! and have a nice weekend all.

 

[jtalden]

Hey thanks that sounds great. i will do them over the weekend and post the results.
im pretty sure i can get all those measurments no problem.

Please post Step 1 first to assure the measurement setup is working correctly.

also I have noticed the timing delays given change depending on the sweep. if a 256k and then compare the left right timings, and then do a 1m and compare the timings it gives different results. i assume its fine as long as you stick to the same length while setting delays right?

Use 256k sweeps. Also activate: REW preferences/analysis/"Adjust clock with acoustic ref."

...and iv been playing with the crossovers but im doing something maybe unconventional and maybe considered wrong but i want the smaller 8" sub to play along with the mains longer and not crossover at the same point. i know thats probably bad in some way but the extra mid-bass im getting from doing that is really nice and i dont want to cut out everything below 100hz on my mains as they play pretty good to ~50Hz.. . anyway. iv saved my config now so i can try more traditional crossovers and see what i prefer.

The delay changes needed are only correct for the XO settings used during the measurements. Any changes to the XO configuration requires a new REW analysis. I can help analyze 1, 2 or maybe 3 XO options at the most. You may want to consider my feedback on your initial preferred XO configuration before measuring with different XO settings.

 

[UKSPAWN]

activate: REW preferences/analysis/"Adjust clock with acoustic ref."

Ah okay, will do. thanks

You may want to consider my feedback on your initial preferred XO configuration before measuring with different XO settings.

yeah i think when i put stuff here i will go with no crossovers and advise me on what looks good to crossover at? and then just do a setup totally on what you show me how to do. then hopfully i will be able to do it myself if i do want to try other crossovers etc.

 

[jtalden]

yeah i think when i put stuff here i will go with no crossovers and advise me on what looks good to crossover at? and then just do a setup totally on what you show me how to do. then hopfully i will be able to do it myself if i do want to try other crossovers etc.

I was not offering to determine the most favorable XO filter settings. That analysis is more than I am willing to do. I was instead suggesting that you use the current XO settings that sound reasonably good. If those XO settings even with delay adjustments are still not very favorable my analysis may include a suggestion as to what XO filter changes may be helpful in improving the results.

 

[UKSPAWN]

alright thanks that sounds good. ill take the Step 1 measurments later today and upload them for you.

Thanks

EDIT. it may have to be tomorrow now due to life stuff

 

[UKSPAWN]

hey i just wanted to let you know i have been reading more and today thanks to your advice to switch on the "adjust clock with timing ref" ive managed to get the delays between L/R channel to ~0.0005ms! and the subs are now pretty spot on too. Really nice!

I tried more with the alignment tool and am steadily learning how it works. although some of the timings it says will be better create a huge null in some places so i guess i have to keep messing with it OR my timings are already quite optimal...

It certainly sounds so much better than before. sterio image has improved and things sound more "holographic"
im now down to these timings.

12" sub = 0
8" sub = +3.47ms
L/R main = +3.48ms

Im going to keep reading and see what i can do on my own but thanks so much for offering to help me.

Many thnaks all

 

[jtalden]

 

[UKSPAWN]

hello again @jtalden

hope everyone is well.
I have moved house recently and so am back to setting up the system again. This time i have some room treatment / absorbtion in the mid-late stages (testing diffusion) (all DIY)

Since i was last here i have pretty much been solid learning about all this... (between life stuff) and i *think* im in a much better position to understand the help you offered me.

I have a few questions regarding phase to make sure im doing things correctly now.


1) Should phase measurments be taken near-field or in the MLP in order to phase align a speaker to a sub? I've been doing it from MLP

2) Does the room change the phase? (related to above)
2.1) speaker/sub poRts and phase? (iv'e sealed my speaker ports and my older ported sub - seems to align better)

3) Once i have determined the delays to align phase, is this permenant even if i change the EQ and/or X-Over or are the delays only relevant to those settings? (i assume/think you have to do new timings as the changes in amplitude and crossovers etc will change phase, is that correct?)

is it possible to rotate the phase more than +/- 180 degrees as my measurments come out with thousands of degrees in unwrapped phase, but a program such as rephase cant correct more than +/-180 ... should i even bother trying to make the phase of all my drivers the same at the crossover freq, is it even poissible? or do some things just not match up no matter what you do?

Im going to attach a pic or 2 or 5 of my new room response, decays and such - ( Psy Smoothed because thats what im bothered about, percieved audable difference not absolute technical 'flatness' to the .25 dB or anything like that).

I just would like confirmation that my stuff looks ok with nothing obviously wrong. and/or tips to make my measuring better etc. there is so much to REW.

appreciate any input from anyone and feel free to ask your own related questions if you wish i dont consider it hijacking is its contributing to the thought process. hey maybe i can even help someone!

**EDIT**
Measurments are with no timing ref, 96khz output and 4m sweeps (40~ second) for accuracy. from the single MLP and 3 of them averaged in order to do the EQs. I also do the EQs with PSY smoothing as it seems to be less drastic changes and after all thats "what we hear" right?

If you need any other info let me know.
**

Thanks in advance, merry holidays n' happy new year!

SPAWN

 

[jtalden]

hello again @jtalden
I have a few questions regarding phase to make sure im doing things correctly now.

1) Should phase measurments be taken near-field or in the MLP in order to phase align a speaker to a sub? I've been doing it from MLP

Mic at LP for SW delay timing measurements.

2) Does the room change the phase? (related to above)

Yes, the room greatly influence the phase response at the LP

2.1) speaker/sub pots and phase? (iv'e sealed my speaker ports and my older ported sub - seems to align better)

It's fine to seal the ports and it does often help dissimilar SWs to better operate together.
A variable phase control on a SW can be used instead of changing the delays in a DSP but has more limitations. It can be left at zero in in case.

3) Once i have determined the delays to align phase, is this permenant even if i change the EQ and/or X-Over or are the delays only relevant to those settings? (i assume/think you have to do new timings as the changes in amplitude and crossovers etc will change phase, is that correct?)

Level changes to the SWs or the mains has no effect on the timing. Global EQ changes also does not impact the timing between the SWs and mains. Independent SW or mains EQ changes in the XO range can have a minor impact on timing but is normally not significant. XO filter or frequency setting changes do impact timing more than EQ, so it is best to measure again after changes to confirm the timing.

is it possible to rotate the phase more than +/- 180 degrees as my measurments come out with thousands of degrees in unwrapped phase, but a program such as rephase cant correct more than +/-180 ... should i even bother trying to make the phase of all my drivers the same at the crossover freq, is it even poissible? or do some things just not match up no matter what you do?

The room effects make the phase rotation much greater than it actually is for the direct sound of the speaker. We are interested in the phase rotation of the direct sound and particularly the phase tracking between the SWs and the mains. This what we refer to as timing, phase tracking, or phase coherence. If necessary, we can use the REW FDW setting options to help remove much of the room influence on the phase when we are measuring from the LP. That makes it much easier to see the phase tracking of the direct sound. It often takes some experience to sort out the relative phase response when measuring at the LP.
Typical setups have maybe 180-720° of phase rotation of the direct sound. It depends on the system design and setup. Yes, rePhase can easily help create an FIR filter to remove the system phase rotation no matter how large. You would need a convolver to use that filter it in your setup.

Im going to attach a pic or 2 or 5 of my new room response, decays and such

All your charts look very good indeed. There is a hint that delay timing has the SWs delayed more than necessary resulting in more phase rotation and group delay than needed. This does not appear to have impacted the SPL in the XO range though, so the impact on sound quality is minimal.

 

[UKSPAWN]

The room effects make the phase rotation much greater than it actually is for the direct sound of the speaker. We are interested in the phase rotation of the direct sound and particularly the phase tracking between the SWs and the mains. This what we refer to as timing, phase tracking, or phase coherence. If necessary, we can use the REW FDW setting options to help remove much of the room influence on the phase when we are measuring from the LP. That makes it much easier to see the phase tracking of the direct sound. It often takes some experience to sort out the relative phase response when measuring at the LP.


Typical setups have maybe 180-720° of phase rotation of the direct sound. It depends on the system design and setup. Yes, rePhase can easily help create an FIR filter to remove the system phase rotation no matter how large. You would need a convolver to use that filter it in your setup.
All your charts look very good indeed. There is a hint that delay timing has the SWs delayed more than necessary resulting in more phase rotation and group delay than needed. This does not appear to have impacted the SPL in the XO range though, so the impact on sound quality is minimal.

Ahh right! So i need to learn this FDW stuff next. thank you so much.
I did actually suspect i had too much delay but it summed OK, i assumed id aligned like a "wavelength" or 2 behind if that makes sense (terminology?!)


I also used the RTA and played some noise like 30-100hz and then just played with the delay in my miniDPS (flicking the mousewheel up and down in 0.05ms increments, would you say this is a good method to the the general are before doing sweeps? -

If i upload my MDAT would you take a little butchers (look) at it in your own time, im in no rush and its becoming less frustrating the more im learning.

Would a basic diagram of room layout and positions of absorbers / speakers etc help or useless? or actual physical measurments to the subs help to get this better...

As you say the Freq response looks ok and it sounds ok too (after a while its difficult to tell if its better or just different, so leaving it a day or so resets that)

As im using 2 different subs (for now, ordering my 2nd SB1000PRO soon TM) when im listening i FEEL like i can HEAR/FEEL when the subs are aligned well as they just "VIBE" together and give you that "silence between the note" - thats probably all the wrong terminology and i hope you know what i mean, but that combined with the RTA and the sweeps... thats how i got to that point.

Now if i learn this FDW stuff i can spend another day telling my kids to stop running around upstars cuz "dads testing"

Thank you so much, iv read so many of your posts and you are a very helpful member of this community.

Thanks again.

 

[UKSPAWN]

does the red and green trace in CLARITY look ok? there seems to be something of a dip about 200-500~ HZ - that is voice inteligability right? or is thjat not a bad dip?

Im starting to feel i can see what it sounds like, do you look at graphs and thing "yeah id like the sound of that setup"

Would mine sound ok to your ears? is there anything you would work on?

 

[jtalden]

Ahh right! So i need to learn this FDW stuff next. thank you so much.
I did actually suspect i had too much delay but it summed OK, i assumed id aligned like a "wavelength" or 2 behind if that makes sense (terminology?!)

Yes, that is what I was thinking also.

I also used the RTA and played some noise like 30-100hz and then just played with the delay in my miniDPS (flicking the mousewheel up and down in 0.05ms increments, would you say this is a good method to the the general are before doing sweeps?

It's hard to be sure which wavelength you are on that way, but you can get a favorable SPL result. There is no need for that if using the alignment tool.

If i upload my MDAT would you take a little butchers (look) at it in your own time, im in no rush and its becoming less frustrating the more im learning.

Yes, given the correct measurements (post-8 step 2), I can check the alignment.

Would a basic diagram of room layout and positions of absorbers / speakers etc help or useless? or actual physical measurments to the subs help to get this better...

That won't help with timing analysis directly, but we all like to see that.

 

[UKSPAWN]

ok if you need timing ref i will need to do some more measurments another day (i just need to enjoy some music and movies now im fatigued from testing). Along with some room pics

Ill still upload the MDAT for how i came to the current config for now. wow its 38mb EEK.. will it work, lets see...

No 38mb is too big ill put it on my GoogleDrive but i understand if you dont download things online from unknown sources

TEST4.mdat

drive.google.com

 

[jtalden]

does the red and green trace in CLARITY look ok? there seems to be something of a dip about 200-500~ HZ - that is voice inteligability right? or is thjat not a bad dip?

Im starting to feel i can see what it sounds like, do you look at graphs and thing "yeah id like the sound of that setup"

Would mine sound ok to your ears? is there anything you would work on?

Clarity is way outside my expertise. It looks very good based on what little I know about it.
Good measurements suggest good sound. Your measurements look very good.

 

[UKSPAWN]

Clarity is way outside my expertise. It looks very good based on what little I know about it.
Good measurements suggest good sound. Your measurements look very good.

Good measurements suggest good sound. Your measurements look very good.

Excellent, thank you


Oh and you said about a convoluter, cant i use FIR in my miniDSP?


EDIT
...Yeah, still here and reading lol...

How do i add a FDW to my measurments or must it be already added at the time of taking the measurment?

 

[UKSPAWN]

How do i add a FDW to my measurments or must it be already added at the time of taking the measurment?

BUMP:

I have enabled the FDW option in the preferences. is there anything else? what are good window settings for better seeing phase (cycles / octaves?)?

I cant find a way to add a window to already taken measurments so im assuming you can't.

Anyone help me get some better phase readings for alignment?

Thank you

OH, and if anyone wants to give me some 1 on 1 time im willing to pay a little $/£ for the time if neccessary. maybe via some voice chat back/forth or something.

 

[jtalden]

Oh and you said about a convoluter, cant i use FIR in my miniDSP?

Some of MiniDSP boxes have capability for FIR filters, but most all are limited in the number of taps available. To remove all low frequency phase rotation a large number of taps is needed. The Roon music server has a convolver that is capable for instance. Some people just use a software music server that includes a convolver function. JRiver and Foobar2000 are examples.

I have enabled the FDW option in the preferences. is there anything else? what are good window settings for better seeing phase (cycles / octaves?)?

I cant find a way to add a window to already taken measurments so im assuming you can't.

Measure without the FDW active in preferences. We can apply and remove FDW as needed on measurements for any particular analysis process.

Anyone help me get some better phase readings for alignment?

If you post the correct 4 measurements (Post-2, Step 2) I can provide some guidance.