Tuning with .WAV files and Timing Reference.......is there any procedures

[Newfiestang50]

So as I posted in another thread I plan on using my MiniDSP 2x4HD as a preamp for both my mains and subs (dual SVS SB-3000) with it being feed with a Sonos Port streamer. The mains (PSB Stratus Goldi) are powered with an Emotiva XPA 2. The connection between the Sonos and miniDSP will be via toslink. I've been informed by other forum members that with such a configuraation it is possible to use a wave file for the pink noise and sweeps with a timing reference and play these signals back from my Sonos streamer. That part I have figured out but after that im kind of lost as to the specifics of tuning with such a setup. I am familiar with performing sweeps in REW using HDMI to a reciever.

So a few questions to start:
1. It states that the timing reference must always be played through a full range speaker. Thats understandable but if am taking a measurement for a sub, wont the full range speaker affect the measurement for that sub? I think the answer has something to do with these timing options but not 100% sure.

2. Timing Offset & System Delay, can someone explain that in a more simple approach for say tuning a sub but timing reference sent through L main.

3. The Generator creates the sweeps but are the sweeps required for each main speaker, both, etc? Do I need both full range sweeps and lower freq sweeps for the subs?

As I said in the other post the main reason for trying this configuration is to basically tune the whole system but specifically apply crossover to the main speakers and limit the base output to them and apply a compressor in the MiniDSP. So basically the 4 o/p's from the MiniDSP will supply signal for L&R mains plus the two subs. I currently have everything running with my Yamaha RXV-2600 which is an older receiver and there isnt much configuration in terms of crossover for the mains. It does sound ok but I think i can make everything sound a little better using the miniDSP as the pre amp. I know there are better units for this but i already have it and would just like to try it. Im also interested in getting further education on REW as well so why not try this.

Thanks, any help appreciated.

 

[JStewart]

I’ve tried to answer your questions below.
Because of somewhat recent changes in the miniDSP GUI I think it’s going to be difficult to find an existing tutorial that is perfect for your scenario. This stuff is mostly done by enthusiasts and the tech doesn’t stand still for long.

Before you begin in earnest, decide how you want the subs configured. A right and left for stereo or dual mono configuration which most of us will likely recommend for it’s ability to smooth out the response. There’s also a strong case to be made that since we can’t tell direction on low bass stereo is pointless in those registers and you will also read that music is mostly mixed with mono bass anyway.

Next set up your channel routing and crossovers in the miniDSP and get the sub levels where you like them. (If you need a hand here perhaps @deercreekaudio can help as I’ve not used the new GUI) Once done, at this point you’ll have a working system you can listen to.

Next experiment with playing REW files through the system. You will need to be able to route a sweep to each speaker individually.

The process I will recommend is:
1. get the mains equidistant from the mic or set the delay for one of them so sound from either arrives at mic at same time.
2. measure the subs and if possible move to a better location for better response.
3. once the subs are in their final resting place, phase align the subs with each other. You can get a better idea of what this is about in this older video:

4. EQ the subs.
5. EQ the mains (below 250Hz)
6. Pick an XO frequency and take measurements for the purpose of aligning the combined subs to the mains.
7. Add shelving filters for bass and treble to taste.
8. When it comes to EQ pretty graphs are not the same as good sound. There are good practices and principles to follow, but in the end let your ears be the final judge.

There will be questions to be answered along the way for the above steps for sure.

So a few questions to start:
1. It states that the timing reference must always be played through a full range speaker. Thats understandable but if am taking a measurement for a sub, wont the full range speaker affect the measurement for that sub? I think the answer has something to do with these timing options but not 100% sure.

It would unless, for measuring purposes, you route both subs to play only from either the left or right, whichever is not the timing reference. Example use left for timing, have subs only on right.

Oh, and you have a USB mic so using acoustic timing is the only option.

2. Timing Offset & System Delay, can someone explain that in a more simple approach for say tuning a sub but timing reference sent through L main.

So I’m not understanding the question.
A timing reference is used so that the relative timing of all the speakers can be determined and delays can then be implemented so that the sound from all speakers arrives at the listening position at the same time. So you might be thinking why not just determine that as time of flight from the speaker to the mic and the reason is because there is varying system latency in your computer so it’s not always the same period of time after starting the sweep to sound emerging from the speaker. Using another speaker as reference overcomes this.

3. The Generator creates the sweeps but are the sweeps required for each main speaker, both, etc? Do I need both full range sweeps and lower freq sweeps for the subs?

You will need Left and Right measurement sweeps 20Hz to 20kHz with timing reference for each being the same, I.e. left or right. No other measurement sweeps are needed.

Periodic pink noise can be used for EQ. You will get differing opinions on only left or right vs both. I would do a left and a right to see if there are any glaring unique problems below 200 to 250Hz and both for tonal adjustments above.

And if @jtalden has time for it, will you please chime in on anything incorrect or needed clarification in my advice? Thank you!

 

[jtalden]

All @JStewart advice in both threads is good (as always). His comments provide a good overview of the process steps. For more overview information there are lots of threads here and many other videos to view. There are still setup details and REW settings that not covered in the overview of the process steps so that is what we can help with.

If you want to proceed from here, I suggest you let us know your current situation and immediate questions.
Are the subs in your preferred positions and XO configuration setup already? Mono subs or stereo sub setup? Have you created REW measurement files yet?

 

[JStewart]

Thank you @jtalden . I appreciate your getting eyes on it.

 

[Newfiestang50]

Ok so the first of a series of basic questions as i proceed through this. As i am familiar with REW in taking measurements the conventional way with the laptop connected via HDMI to AVR most of my questions will be around taking measurements with wave file sweeps as im sure you already assumed from above.

So im about to start recording the files with the generator, for the pink noise wave files, is there a way to make these longer in length, once i record them and save as wave they are really to short to be able to use for setting levels in the system before doing sweeps. Is is just a matter of increasing the repetitions here?

And do everything look ok here for the sweeps settings? Do you think the starting freq of 15Hz is low enough.

Also, i had to select PCM32 as the format as Sonos wont seem to play the default 32 bit float. Is this ok?

 

[jtalden]

The recent versions of REW file generation dialogs look as shown below:
For the sweeps, you will need one as shown for the left channel and another one for the right channel. A full range measurement file as shown is all that is needed for measuring either L, sub(s), or both. 12 to 15 Hz is a good starting point for the sweeps.

For MMM EQ a PinkPN file can be created as shown. 40 sec should be enough time for the measurement process. We can just mute one of the 2 channels in the DSP when measuring.

There is no need to set levels with the PinkPN file. Instead, set the volume a little lower on your setup than you listen to music at and make a sweep measurement. Then adjust the volume and remeasure as needed to reach about 75 dB on the REW chart.

Start with creating a right and left sweep measurement at about 75 dB and post the mdat here.

 

[Newfiestang50]

Here are the mdat for my mains however im confused as to how to perform the sweeps for the subs without playing one of the mains as those have the timing ref included. What am i missing? If i perform the sweep with one of the mains playing as well it would seem that the lower freq output of the mains would be included in measurement of the sub?

 

[Newfiestang50]

And i think i figured out a way to not have the left main with timing reference not interfere with the subs measurement. Basically I just set the high pass crossover to 500Hz. Sound ok to use this method?

 

[jtalden]

The sweep measurements look good so I can confirm there are no issues making sweep measurements. Well done. Even the sub timing and levels appears to be close to a favorable setting. I would guess the subs are currently located resonable close to the mains rather than randomly in the room.
Are you are settled on the sub positions and XO settings you intend to use? Did you settle on mono subs or stereo? If settled, what settings are being used? If you instead intend to try different sub locations and set favorable sub to sub timing and XO settings that is a different effort.

We don't know your step/objective. Different measurements are made for the different steps in the process.

Main Steps:

1. Investigate optional sub positions, find sub to sub timing needed for those positions and choose a XO frequency (optional, can leave the subs as placed and use the popular 80 Hz LR48 XO)
2. EQ prior to XO timing (optional, but a favored approach by many)
3. XO timing - Set favorable subs to mains timing.
4. EQ (can optionally just be done last after subs to mains timing)

And i think i figured out a way to not have the left main with timing reference not interfere with the subs measurement. Basically I just set the high pass crossover to 500Hz. Sound ok to use this method?

For XO timing step, I would use the R channel sweep and leave the left main unmuted so the timing chirp will sound. Mute the right main in the MiniDSP so only the sub(s) you want to measure will play. We can do that now, if you want to bypass the optional sub/XO setup step and the optional preliminary EQ step that @JStewart suggested.

If you intend to instead investigate optional sub positions, find sub to sub timing needed for those positions and choose an XO frequency to use then make another L+R full range measurement file without the timing reference. It is not needed for that process. Just measure each main alone and make measurement of one of the subs in each optional position any XO should be defeated for this step. Use the DSP to mute other speakers from playing. The analysis in done in REW.

 

[JStewart]

And i think i figured out a way to not have the left main with timing reference not interfere with the subs measurement. Basically I just set the high pass crossover to 500Hz. Sound ok to use this method?

For XO timing step, I would use the R channel sweep and leave the left main unmuted so the timing chirp will sound. Mute the right main in the MiniDSP so only the sub(s) you want to measure will play. We can do that now,

To further clarify, in the XO step, which ensures the subs are at their best alignment with the mains, sweeps are taken with the crossovers engaged for the subs together and a main OR each sub and it’s associated main depending on your setup preference.
For jtalden‘s suggestion above, only the Right channel signal is routed to each of subs. This way there is no main that plays on the timing reference channel when measuring the sub. You can change the routing back after measuring.

Edit: if you don’t want to mess with routing, and assuming the timing chirp is on the left, measure the right sub first, then unplug it from the back of the MiniDSP and put the left sub in it’s place for the measurement. The sub sweeps can then be added together in REW if your setup is dual mono.

Edit Edit: Turn the subs off or unplug their AC before moving wires.

 

[Newfiestang50]

The sweep measurements look good so I can confirm there are no issues making sweep measurements. Well done. Even the sub timing and levels appears to be close to a favorable setting. I would guess the subs are currently located resonable close to the mains rather than randomly in the room.

Im pretty sure the mdat I pasted in above were only the main speakers. will post another mdate file later with subs in there as well

Are you are settled on the sub positions and XO settings you intend to use? Did you settle on mono subs or stereo? If settled, what settings are being used? If you instead intend to try different sub locations and set favorable sub to sub timing and XO settings that is a different effort.

We don't know your step/objective. Different measurements are made for the different steps in the process.

I have the mains amp connected to o/p's one and two on the minidsp, the rear sub (side wall to rear of room) is on o/p 3 connected with a SVS triband wireless transmitter/receiver, and the front sub is on o/p 4 and its located next to the right front main. Its connected via wired RCA. The subs are set up in dual mono. The subs are limited to their current locations. Currently i just have them level matched and just threw in a X-over of 120Hz to start, high i know.

Main Steps:

1. Investigate optional sub positions, find sub to sub timing needed for those positions and choose a XO frequency (optional, can leave the subs as placed and use the popular 80 Hz LR48 XO)
2. EQ prior to XO timing (optional, but a favored approach by many)
3. XO timing - Set favorable subs to mains timing.
4. EQ (can optionally just be done last after subs to mains timing)

For XO timing step, I would use the R channel sweep and leave the left main unmuted so the timing chirp will sound. Mute the right main in the MiniDSP so only the sub(s) you want to measure will play. We can do that now, if you want to bypass the optional sub/XO setup step and the optional preliminary EQ step that @JStewart suggested.

I did something a little different so please tell me if this makes sense. So to further play with sub alignment i muted the right main speaker and the sub i didnt want to play. I used the left main for timing as you mentioned but i applied a 1200Hz low pass filter in it so there would be no low freq interfering and the timing chirp still triggers the sweep to start. This made sense to me as the measurement taken should now only be what that sub being tested is playing. Does this sound ok.

If you intend to instead investigate optional sub positions, find sub to sub timing needed for those positions and choose an XO frequency to use then make another L+R full range measurement file without the timing reference. It is not needed for that process. Just measure each main alone and make measurement of one of the subs in each optional position any XO should be defeated for this step. Use the DSP to mute other speakers from playing. The analysis in done in REW.

havent gotten this far yet but maybe tonight

 

[jtalden]

Im pretty sure the mdat I pasted in above were only the main speakers. will post another mdate file later with subs in there as well

Okay good. That explains why the timing throughout the bass range looks good. The measurements served their purpose to confirm you can make good measurements.

I have the mains amp connected to o/p's one and two on the minidsp, the rear sub (side wall to rear of room) is on o/p 3 connected with a SVS triband wireless transmitter/receiver, and the front sub is on o/p 4 and its located next to the right front main. Its connected via wired RCA. The subs are set up in dual mono. The subs are limited to their current locations. Currently i just have them level matched and just threw in a X-over of 120Hz to start, high i know.

Very helpful information - distributed mono subs a good choice and 120 Hz is not too high in my book.

I did something a little different so please tell me if this makes sense. So to further play with sub alignment i muted the right main speaker and the sub i didnt want to play. I used the left main for timing as you mentioned but i applied a 1200Hz low pass filter in it so there would be no low freq interfering and the timing chirp still triggers the sweep to start. This made sense to me as the measurement taken should now only be what that sub being tested is playing. Does this sound ok.

Yes, it is good to experiment with various measurements to understand how it all works. Your 1000 Hz high pass filter idea on the mains could work okay for step 1, but I don't know why you would want to do it that way. I would remove all DSP filters and mute the right main and the rear sub in the DSP. After the front sub is measured then mute it and unmute the rear sub and measure it. That way we have a measurement of just the 2 subs and, if we want, we can also look at sub step response and other characteristics without the interference from the right main.

 

[Newfiestang50]

but I don't know why you would want to do it that way.

Because I dont know any other way to perform the sub measurements, i need to keep the timing reference playing from the left main as this is where ive chosen to play it for every other measurements. How do i play the sweep with timing reference and not have the low freq not play from the main speaker and interfere the measurement taken of one of the subs.

Maybe im not seeing something but the timing ref sweep needs to play from a main speaker correct?

And if i allow that sweep to play at full range from 12 to 20Khz then the low freq (< 250 Hz) will be recorded with the sub thus not giving a true representation of the measurement of the sub alone......correct?

The more i read your guys instructions the more i think there was a way to record the timing chirp without recording the full freq sweep of 12 to 20Khz.

 

[jtalden]

From Post 12

I would remove all DSP filters and mute the right main and the rear sub in the DSP. After the front sub is measured then mute it and unmute the rear sub and measure it. That way we have a measurement of just the 2 subs and, if we want, we can also look at sub step response and other characteristics without the interference from the right main.

There are usually several way to accomplish a task. I suggested an easy way to do it. You don't yet see the logic, but why not try it. If I made a mistake, or you do, the poor mdat measurement result will provide us clues as to the problem.

This process uses the right channel measurement sweep you already have created and tested above.
The timing chirp will play in the unmuted left main speaker. The measurement signal will pass through the right channel only, but the right main speaker is muted and will not sound. The DSP should be setup to direct both the right and left channel information to the 2 subs as well so the unmuted sub will play the sweep arriving from the right channel.

@JStewart, am I missing something?

 

[JStewart]

The DSP should be setup to direct both the right and left channel information to the 2 subs as well so the unmuted sub will play the sweep arriving from the right channel.

As you ay there are different ways to accomplish the same. I’d feed both subs from the right channel only with a left channel timing reference. Take the measurements. Change the routing back. Done.

And sometimes a picture can help when words fail…. This is what the miniDSP routing GUI would look like.

(apologies for crude cut and paste. My Windows machine away for a week.)

This chart shows the settings for each measurement when routing is as above.

Maybe im not seeing something but the timing ref sweep needs to play from a main speaker correct?

Yes.

 

[Newfiestang50]

From Post 12

There are usually several way to accomplish a task. I suggested an easy way to do it. You don't yet see the logic, but why not try it. If I made a mistake, or you do, the poor mdat measurement result will provide us clues as to the problem.

LOL yes you are correct, i do not see it, no fault of yours, i just havent been able to understand how the sweep with timing reference dont interfere with other speaker measurements when that sweep actually plays full range along with the speaker i intend to measure. I will study your response above again and see if i can get this through my thick skull lol. Pleas have patience with me lol.

 

[Newfiestang50]

As you ay there are different ways to accomplish the same. I’d feed both subs from the right channel only with a left channel timing reference. Take the measurements. Change the routing back. Done.

And sometimes a picture can help when words fail…. This is what the miniDSP routing GUI would look like.
View attachment 67800
(apologies for crude cut and paste. My Windows machine away for a week.)

This chart shows the settings for each measurement when routing is as above.
View attachment 67801



Yes.

And I understand the chart above perfectly but again my question is, when playing the left channel sweep with timing chirp, wont the low freq from that left channel sweep give inaccurate results of the channel being measured (say right sub) as it will measure the left channel low freq and the sub frequencies being measured? If not then im totally missing something as the left channel sweep as i recorded is a full spectrum sweep, 12 Hz to 20 KHz. Maybe i done this reference sweep wrong and thats where all the confusion is here.

 

[jtalden]

One of the 2 measurement files you already created has the left channel timing chirp and the right channel measurement sweep. I confirmed that in the mdat you posted. You used it to create the "Main R Full Sweep" measurement.

We are just using that same file again.

The timing chirp is on the left channel. There is no measurement sweep on the left channel.

The measurement sweep is on the right channel. There is no timing chirp on the right channel.

When we are making the measurement the timing chirp plays first on the left channel. When it completes then REW starts the measurement capture process, and the measurement sweep is played on the right channel. Only the measurement sweep is captured not the timing chirp.

 

[JStewart]

And I understand the chart above perfectly but again my question is, when playing the left channel sweep with timing chirp, wont the low freq from that left channel sweep give inaccurate results of the channel being measured (say right sub) as it will measure the left channel low freq and the sub frequencies being measured?

No. In the routing chart above it clearly shows the left input channel is not output to any sub!
If you set the routing as in the example
and mute outputs according to the chart for what you’re measuring
and both recorded sweeps have timing reference as left then
it will work. Please try it and see.

 

[Newfiestang50]

No. In the routing chart above it clearly shows the left input channel is not output to any sub!
If you set the routing as in the example
and mute outputs according to the chart for what you’re measuring
and both recorded sweeps have timing reference as left then
it will work. Please try it and see.

Ok followed your instructions and routing chart. Attached is the mdat. Let me know if it looks ok and we will proceed with next step. The below mdat sweeps have no EQ or time alingment done yet. The are base sweeps following your method.

 

[Newfiestang50]

And this is the routing for dual mono correct?

 

[JStewart]

And this is the routing for dual mono correct?

View attachment 67823

Yes. The routing is correct.

I’m thinking each input gain (the purple block) for each sub should be decreased by 6dB in order to keep the input from clipping in the digital realm, but I’m 100% on that. Perhaps someone else can confirm?
Regardless, making that change will keep the sub outputs equal to the mains outputs, electrically speaking.

EDIT: miniDSP doesn’t suggest the input gain adjustment as I have in this post so I’m likely in error on this point, but don’t understand why

https://www.minidsp.com/applications/subwoofer-tuning/sub-integration-rew

 

[Newfiestang50]

Yes. The routing is correct.

I’m thinking each input gain (the purple block) for each sub should be decreased by 6dB in order to keep the input from clipping in the digital realm, but I’m 100% on that. Perhaps someone else can confirm?
Regardless, making that change will keep the sub outputs equal to the mains outputs, electrically speaking.

EDIT: miniDSP doesn’t suggest the input gain adjustment as I have in this post so I’m likely in error on this point, but don’t understand why

https://www.minidsp.com/applications/subwoofer-tuning/sub-integration-rew

Hmmm ok, so why would the subs output only need to be decreased, are you inferring that from the latest mdat. With that said the front sub came with a distortion warning when playing the sweep.

 

[JStewart]

Hmmm ok, so why would the subs output only need to be decreased, are you inferring that from the latest mdat. With that said the front sub came with a distortion warning when playing the sweep.

Not inferring from the mdat. I don’t have a REW capable machine at the house this weekend.

When looking at the routing photo you posted the thought occurred to me. My reasoning for reducing the left and right mains input to each sub was because a full left signal and a full right signal are added together and sent to each sub, so each sub will be twice as loud as a main (electrically speaking as the sub’s amplifier volume adjustment comes after). The 2nd reason is because I don’t know what the result in the digital realm is when the left and right signals are combined for each sub.

Hopefully someone else can provide clarity.

By the way, was this the specific warning you saw or something different?

If it was you can find more info in REW help

Making Measurements

 

[jtalden]

Ok followed your instructions and routing chart. Attached is the mdat. Let me know if it looks ok and we will proceed with next step. The below mdat sweeps have no EQ or time alingment done yet. The are base sweeps following your method.

The file you posted was not suitable for use in step one of process. The single measurement that was labeled "All Subs & Mains". We need the 4 speakers measured individually.

@JStewart posted this to show the 4 measurements that are needed:

This chart shows the settings for each measurement when routing is as above.

All 4 of these should be measured using the existing file that measures R with timing chirp L.