Loopback Method Trying To Get It

[jtalden]

Thanks for the data. The delay timings are ideal for this setup.

Summary
Summing of the TW and MW is shown in charts below. Phase crosses at 2 kHz and it has the closest phase tracking possible given the choice of polarities. This results in very favorable SPL support across the XO range.

Summing of the TWMW and MW is shown in charts below. Phase crosses at 450 Hz and it has very close phase tracking resulting in very favorable SPL support across the XO range.

You were interested in the step responses. They are shown the charts below.

The overall step response is in the upper chart. This is a pretty typical step response for a well-timed 3-way speaker. Much of the irregularity will be reduced when EQ is applied.

The relative timing between the 3 drivers can be seen in the lower step response chart. The TW and MW start at the same time and the W is delayed only about 0.15 ms from them. It is common for there to be even more variation between them as the objective is to have favorable phase tracking and thus favorable SPL. We just don't want to see large differences between them as it indicates that we haven't found the most favorable timing.

The charts below show the overall SPL and phase responses as measured vs calculated just to assure no errors were made. The only difference is that the TW level was reduced by 3 dB after the TW measurement was made. This has no effect on any of these results - all looks good.

 

[Trdat]

That is great news. I will stick to this particular delay with the inverted MW and TW. With some basic eq ing it is sounding decent, I just have to keep trying different biqauding techniques and keep learning.

Your gonna hate me but I really would like to set the delays for the subs as well. Of coure only if you can be bothered continuing. Hopefully it could be a simple step.

I do use Audiolense for a final correction but that is only for music plus Audiolense won't recognize the centre speaker seperately, I will have to combine the centre with the either left or right. On top of this there is a delay with the sound to the video when wathcing youtube while using Audiolense so I don't always use it.

So essentially it would be great to get everything set up for movies/music without turning on Audiolense. I might not even use Audiolense at all. And that leaves the subs. I have 3 subs woofers and I am aware I cannot get individual timing till i get something like a minidsp. But at least if i can match the mains with the sub that would be awesome. It is through AVR and crossover point will be around 80hz or 100hz

I read through this thread https://www.avnirvana.com/threads/understanding-sub-phase-group-delay-for-placement.5976/

but some questions still come to mind.

1. Is the instructions on post 8 the way to go?Or do we have similar measuring techniques as we have used so far...?
2. Do I need to take this measurement in the similar place as the others?
3. Are we going to basically delay the whole mains to match the subs or are we again going to isolate the woofer with the sub and phase match that?( I only ask this to get a picture)

 

[jtalden]

I have 3 subs woofers and I am aware I cannot get individual timing till i get something like a minidsp. But at least if i can match the mains with the sub that would be awesome. It is through AVR and crossover point will be around 80hz or 100hz

Pick your XO point (I will help with just one of them).

1. Is the instructions on post 8 the way to go?Or do we have similar measuring techniques as we have used so far...?

Yes, it's similar. I assume the 3 mains are set to small so that the redirected bass is sent to the 3 SWs.

Setup

• Mic placed at the MLP
• Sweep range must cover at least 10-1000 Hz for all measurements. [I usually sweep full range for convenience and confirmation of measurement integrity.]
• Loopback timing on
• XOs active

4 measurements

1. FL alone
2. FR alone
3. CC alone
4. SWs alone [measure using the L channel with FL disconnected so the redirected bass is measured.]

3. Are we going to basically delay the whole mains to match the subs or are we again going to isolate the woofer with the sub and phase match that?( I only ask this to get a picture)

We will determine if any changes are need from the current timing/distance settings to achieve:
Favorable timing for the 3 front speakers
Favorable timing between FL+FR and the SWs
Favorable timing between CC and the SWs.

I thought the FL and FR 3-way XOs were handled by Audiolense so I am confused as to how this all fits together. I don't need to understand however, so long as you do and know how to take the needed measurements and make the adjustments to timing/distances that we find.

 

[Trdat]

Pick your XO point (I will help with just one of them).

That is fine, I really appreciate that. UNless what you mean by one is one crossover point? Yes, one crossover point is fine.

If you mean by one sub, I was under the impression that all 3 subs will be merged and timed together to that of the mains? What is the real practise? To measure each sub seperately to that of the mains or all 3 at the same time?

Yes, it's similar. I assume the 3 mains are set to small so that the redirected bass is sent to the 3 SWs.

Yes they are, through the AVR.

Setup

• Mic placed at the MLP
• Sweep range must cover at least 10-1000 Hz for all measurements. [I usually sweep full range for convenience and confirmation of measurement integrity.]
• Loopback timing on
• XOs active

4 measurements

1. FL alone
2. FR alone
3. CC alone
4. SWs alone [measure using the L channel with FL disconnected so the redirected bass is measured.]

Measure using left channel? Left channel of the sub or from the audio interface?

We will determine if any changes are need from the current timing/distance settings to achieve:
Favorable timing for the 3 front speakers
Favorable timing between FL+FR and the SWs
Favorable timing between CC and the SWs

Great.

I thought the FL and FR 3-way XOs were handled by Audiolense so I am confused as to how this all fits together. I don't need to understand however, so long as you do and know how to take the needed measurements and make the adjustments to timing/distances that we find.

The FL and FR are passive crossover and Audiolense will do a final correction. But yeh, you don't need to understand it all good. Just pretend Audiolense is totall out of the equation.

Just remember that the centre speaker configuration is through the Hypex plate amp DSP while the sub configuration will be done on the Marantz AVR through the distance setting. The mains are passive crossover standard speakers connected to the AVR. The centre is also connected to the centre channel of the AVR but with its settings in the Hypex DSP. My understanding is that all this is fine.

 

[jtalden]

That is fine, I really appreciate that. UNless what you mean by one is one crossover point? Yes, one crossover point is fine.

Yes, 1 XO frequency.

If you mean by one sub, I was under the impression that all 3 subs will be merged and timed together to that of the mains? What is the real practise? To measure each sub seperately to that of the mains or all 3 at the same time?

Normal practice is to first choose the 3 individual SW delays to find the most favorable SPL and phase for the SWgroup. We can't do that without an additional box. For now, all 3 SW are used with their current timing so that is how we will measure them. By 'SWs alone' I meant 'SWgroup alone'. There are only 4 measurements listed so I thought that would be clear.

Measure using left channel? Left channel of the sub or from the audio interface?

Send the REW measurement sweep output to the AVR's left channel only and disconnect the FL speaker. The SWgroup will thus be measured with the AVR XO active. [Do not sent the REW sweep to the AVR LFE channel because the AVR XO filter would be bypassed.]

Note: I am assuming that the AVR SWgroup XO will be set to the same frequency for all 3 front main speakers. Let me know if that is not the case.

 

[Trdat]

Note: I am assuming that the AVR SWgroup XO will be set to the same frequency for all 3 front main speakers. Let me know if that is not the case.

Yes, all the same XO frequency for all speakers.

I tried measuring using the analog inputs on the AVR, my centre speaker had some anamolies plus not sure if the XO was enabled through this process, it should but the measurment showed a response from 30hz with the subs disconnected. I also had to switch the centre speaker to one of the left and right channels XLR outputs to measure it as there is only left right input on the rear for these analog inputs. There might be a way to do this with the 7.1 channel analog inputs but I couldn't get it to work.

Anyway, I will be recieving my new audio interface with SPDIF out and can use that to connect it to the coaxial in on the AVR which should make things easier.

I will get back to you soon once I get a chance to work out the best way to do this. I think either should work to be honest with the SPDIF out option from the Focusrite a little simpler.

 

[jtalden]

I don't think the 7.1 inputs will work for this purpose.

We can connect the Focusrite analog output to one of the stereo inputs on the AVR (like the CD input). If we use a splitter at the AVR we can connect to both the L and R CD inputs. Set the AVR speaker configuration to just the front 3 (no surround channels). Set the AVR mode to either 'mono' or 'all channel stereo'. Disconnect the FL and FR speakers and choose the CD input for the measurement. I think that should result in the correct SPL level relative to the FL and FR measurements.

This should also work if only L or R CD input is used (no splitter). In that case the CC SPL level will be 3 dB lower than measuring with the splitter. We can just adjust the CC measurement level +3 dB in REW to match the FL, FR and SWgroup measurements.

I have no idea how you intend use a SPDIF signal from a Focusrite Solo or retain the needed loopback. If you are using an audio interface with those options, using the SPDIF AVR input should work the same as using the AVR analog inputs.

My setup is very different and I'm unclear about your setup, so I'm not certain this advice works as intended.

 

[Trdat]

Okay so I hope the measurements are good.

But I have to tell you that due there being left and right analogue in, I had to switch the centre XLR into one of the left rights to measure it. Does that work?

I checked and the crossovers were in place at 100hz. I measured each one alone without subs and measured the subs with the left channel but with the FL not connected.

 

[jtalden]

The measurements are suspect in several ways.

• There is a lot of low frequency SPL in the 3 mains suggesting that the 100 Hz XO low pass filter may not have been active in the AVR.
• There is a leading 20 Hz low level noise preceding some of the measurements.
• Distortion is particularly high in the FR main around that 330 Hz SPL sag.
• The CC phase rotation indicates that the polarity of the CC woofer is the reversed relative to the FL and FR.
• There is a very large sag and null centered at 330 Hz. This may well be a room effect rather than a measuring problem, but it suggests a change to LP and or speaker positions should be considered.

Questions

• Are all SWs connected to be used as a SWgroup for the 3 front mains?
• Were all SW disabled during the 3 mains measurements?
• Are your measurements reliably repeatable?
• Are your SWs all ported?
• What is the difference between the 'Sub' and the 'Subs' measurements. Is it just a repeat?

Suggestions

• Reconfirm to yourself that measurements repeated several times will provide the same result for SPL, phase, impulse, step, distortion and timing. Some of the data hints that it may not be consistent.
• The AVR XO operation can be confirmed by measuring the front speakers with it set on and off to confirm it is working as expected.
• Decide if you want to reverse the polarity of the CC woofer to match the FL and FR. If so, then we should first adjust its delay timing accordingly. I can do that using the Post-50 measurements, but we will still need a new set of LP measurements for this SW timing effort.
• Just FYI, it is better practice to start your sweeps about 15 Hz rather than 0 Hz as your SWs have no usable SPL below that frequency.

I can still use this data to try to determine favorable SWgroup XO timing changes if you like, but I that won't be of any value if you will be making any changes based on the above comments.

 

[jtalden]

But I have to tell you that due there being left and right analogue in, I had to switch the centre XLR into one of the left rights to measure it. Does that work?

Yes, that works.

I checked and the crossovers were in place at 100hz. I measured each one alone without subs and measured the subs with the left channel but with the FL not connected.

Okay, I forgot you already mentioned this. It seems unlikely but, it's possible that the woofer SPL outputs are that elevated even with the XO filter active. A 12 dB/octave filter at 100 Hz should have made a bigger dent in the low-end response.

 

[Trdat]

The measurements are suspect in several ways.

• There is a lot of low frequency SPL in the 3 mains suggesting that the 100 Hz XO low pass filter may not have been active in the AVR.

I noticed that, maybe from room gain. The speaker is set in the corner.

• There is a leading 20 Hz low level noise preceding some of the measurements.

Well, if it was for one measurement it could be the metro but if its on more than one than could be a hum or something or ground loop.

• Distortion is particularly high in the FR main around that 330 Hz SPL sag.

Not sure about this one.

• The CC phase rotation indicates that the polarity of the CC woofer is the reversed relative to the FL and FR.

I've checked the polarity of the left and right as I only recently(before our measurements) changed the tweeter and mid woofer polarities to inverted. The centre speaker polarities are set from the hypex design filter I check the box and I doubt I have got this wrong, as the woofer is unchecked and MW and TW is checked. But, it could be that the FL and FR have there polarities switched from the passive crossover somehow from a mistake.

Also, I apologize but the crossover points are different. So, going on from what you have mentioned I think the phase wont be perfect anyway right?

• There is a very large sag and null centered at 330 Hz. This may well be a room effect rather than a measuring problem, but it suggests a change to LP and or speaker positions should be considered.

I am guessing this is SBIR, the speakers are very close to the walls. And my Audiolense measurement that I have taken a while back also has that sag in the middle.

Questions

• Are all SWs connected to be used as a SWgroup for the 3 front mains?

Yes, the AVR Marantz 7001 doesn't have individiual channels of subs so the subs are connected left and right with one side having a splitter for the extra sub.

• Were all SW disabled during the 3 mains measurements?

Yes, I just unplugged the RCA.

• Are your measurements reliably repeatable?

I will give it another try.

• Are your SWs all ported?

No, they are all closed boxes super sealed. 2 are the same while the 3rd one is DIY with the same amplifier but different driver.

• What is the difference between the 'Sub' and the 'Subs' measurements. Is it just a repeat?

Yes just a repeat for repeateability.

Suggestions

• Reconfirm to yourself that measurements repeated several times will provide the same result for SPL, phase, impulse, step, distortion and timing. Some of the data hints that it may not be consistent.
• The AVR XO operation can be confirmed by measuring the front speakers with it set on and off to confirm it is working as expected.

I didn't think of this, I will try. I just played noise to see if the subs were working.

• Decide if you want to reverse the polarity of the CC woofer to match the FL and FR. If so, then we should first adjust its delay timing accordingly. I can do that using the Post-50 measurements, but we will still need a new set of LP measurements for this SW timing effort.

Let me try one more time and see if the measurments are repeatable and then I can decide.

• Just FYI, it is better practice to start your sweeps about 15 Hz rather than 0 Hz as your SWs have no usable SPL below that frequency.

Okay will do.

I can still use this data to try to determine favorable SWgroup XO timing changes if you like, but I that won't be of any value if you will be making any changes based on the above comments.

Let me take one more measurement and see. If it is repeatable it would be good to switch the polarity of the centre woofer(if it would make it better considering that the crossover point is different). And then will go from there.

Okay, I forgot you already mentioned this. It seems unlikely but, it's possible that the woofer SPL outputs are that elevated even with the XO filter active. A 12 dB/octave filter at 100 Hz should have made a bigger dent in the low-end response.

This is my take on it as it happens all the time, I always get heaps of SPL on the mains through AVR's crossover. But I wonder if it has something to do with choosing LFE-mains or just LFE in the AVR options? If not then, the 12db crossover isn't making the dent intended to.

Lastly, I totally know that the speaker positioning is bad but I am limited with my 40square meter apartement, my gear takes up 20%. Lol!! There my secondary speakers so they take the back seat.

Let me get back to you after I have taken more measurements and ensured the crossover on the AVR is on.

 

[jtalden]

This is my take on it as it happens all the time, I always get heaps of SPL on the mains through AVR's crossover. But I wonder if it has something to do with choosing LFE-mains or just LFE in the AVR options? If not then, the 12db crossover isn't making the dent intended to.

Yes, there is a difference. The LFE-mains disables the XO high pass filter so that the mains are not rolled off. Phase analysis with LFE-mains selected can be a more difficult using phase analysis, but that doesn't necessarily mean the SPL result will be less favorable. While I normally like the idea of a typical XO handoff from SWs to mains, I have also seen good results using SW-mains.

FYI, to check measurement repeatability, it is only necessary to make several measurements (maybe 5 to 10) of one channel; the FL with subs active is a typical choice. If you question the results, I can take a look at the file also. This probably is not a real problem given your comments and the small room.

FYI, below is a chart showing how I know the woofer is different polarity in the CC vs the FL and FR. The phase tracks well from 20 kHz down to about 500 Hz. Below that it is clear the phase rotation angle changes greatly. Differing XO frequency and filter slopes have some impact on phase rotation also, but no matter the relative source of the difference, it is pretty clear there will be closer phase tracking below about 500 Hz with a woofer polarity change. This is not likely to make a noticeable difference to the sound quality. It is more a technical observation. Many CC speakers are designed differently to the L, R even from high end companies. I only mentioned it here because the 3 speakers employ the same drivers so they could have a more similar phase rotation if desired.

I am fine with using the data you have already provided but will wait to hear if you prefer to make any changes.

 

[Trdat]

Hello,

I never understood LFE-Mains and LFE, now I know. I think I would prefer a nice handoff to the subs from the mains, I mean thats the whole idea of having subs but I also would like to measure it to see the differences between the two on the low end FR. Perhaps LFE-mains produces a better bass response. I don't think it makes a huge difference but I will check it out.

Regarding changing the polarity for the centre woofer if you mention it won't make that much difference I am happy for it to stay this way. I also can't imagine it having any audible difference.

I will take repeatable measurements just to double check and also play around with LFE-Mains and LFE and make a decision. Chances are the measurements are okay but if I decide to go with the LFE then I will send you the new measurements anyway.

Just one question I read that Focusrtie 1st and 2nd generation don't have loopback but I am under the impression that is some other feature connected with the DAW or recording but not 100% sure. I don't think there would be a problem with performing the loopback you mentioned on the solo with a 1st generation focusrite yes? Connect mic input 2 to monitor output to 2 same right? Just checking maybe stupid question.

 

[jtalden]

Perhaps LFE-mains produces a better bass response. I don't think it makes a huge difference but I will check it out.

Regarding changing the polarity for the centre woofer if you mention it won't make that much difference I am happy for it to stay this way. I also can't imagine it having any audible difference.

I will take repeatable measurements just to double check and also play around with LFE-Mains and LFE and make a decision. Chances are the measurements are okay but if I decide to go with the LFE then I will send you the new measurements anyway.

Good Plan!

Just one question I read that Focusrtie 1st and 2nd generation don't have loopback but I am under the impression that is some other feature connected with the DAW or recording but not 100% sure. I don't think there would be a problem with performing the loopback you mentioned on the solo with a 1st generation focusrite yes? Connect mic input 2 to monitor output to 2 same right? Just checking maybe stupid question.

You are already using loopback timing and that is working well. You are likely referring to the built loopback feature that some other audio interfaces that can implement internally to make that same loopback connection. This feature is selected within it's DAW interface. If so, that is no different than just using your current setup using the external cable. The internal DAW connection only saves the price and clutter of having an external cable make that same connection.

You could also be referring to the REW loopback calibration of the soundcard, i.e., Solo audio Interface in the case? You can do that one-time calibration if you like, but it is not necessary because the Solo frequency and phase responses are very flat throughout the audio frequency range. Creating and applying a soundcard calibration will not significantly impact your results. It is a good idea to do this with a new first-time setup as it helps to assure REW and the audio interface are correctly setup for measurements. Implementing a mic calibration file is a good idea if you have one.

 

[Trdat]

After a lot of deliberating I do want to use just LFE, as there is too much bass with both LFE-Mains during movie/TV watching.

That means I have to measure again, so hopefully I can get to it soon. I will probably lower my crossover point to 80hz which won't make much difference. Also, will stick to the above polarities so all I need to do is measure and send.

It seems there is consistency in the measurements so the above is fairly correct but I did come across some anomalies with my last measruement so if I get the same as above(the previous one I sent) I'll send them through.

Thanks again for your patience.

 

[Trdat]

The measurements are suspect in several ways.

• There is a lot of low frequency SPL in the 3 mains suggesting that the 100 Hz XO low pass filter may not have been active in the AVR.
• There is a leading 20 Hz low level noise preceding some of the measurements.
• Distortion is particularly high in the FR main around that 330 Hz SPL sag.
• The CC phase rotation indicates that the polarity of the CC woofer is the reversed relative to the FL and FR.
• There is a very large sag and null centered at 330 Hz. This may well be a room effect rather than a measuring problem, but it suggests a change to LP and or speaker positions should be considered.

So Jtalden, I have been experimenting and it seems that the sound system is better with LFE+mains the bass coming out of closed box is tight and punchy can't do without it.

I revisted the above issues and like we both agreed they seem more like room issues than actual measurment anomalies. And unfortunately I can't change around the speakers as my 120 litre cabinets take priority.

The only things that I couldn't figure out is the 20hz low level noise. If this doesnt really effect the measurement then we can go with the measurement in post 58. I will obviously ensure the crossover is at 100hz unless I can choose the crossover and wont make a difference if the measurement was taken when it was at 100hz?(I have it at 80hz now) Let me know about this?

Laslty, you also mentioned about reversing polarity so lets keep the CC speaker polarity essentially keeping everything the same and work on a favourable timing between mains and subs. I think I remembered everything right?

 

[jtalden]

Okay, I started to analysis post 58 measurements again. The chart below shows the current predicted SPL based on these measurements.

I see now why I questioned this data originally. It seems unlikely that this is representative of the actual situation. If we use this data, I could suggest a sub level change of maybe -15dB and some AVR distance changes but suspect that if implemented you would report the resulting sound quality would likely be much worse (unless all you hear is bass now).

I first suspected the sub group was active during the FR, FL, and maybe centre measurements because their SPL levels in the low frequency range seems to suggest that, but the step response does not support that idea. So, adding in the sub group, we get that result shown above with the bass range maybe 15 dB louder than the high frequencies. That, coupled with the midrange sag dropping well below the upper frequencies looks really bad. That sag is a big issue on its own and timing and sub level changes will not help that.

I thought ablut this for a while and decided to go ahead and see what the sub level and timing adjustments would look like. The delay timing suggests the mic was about 10 cm off center. This is not a problem. I just shifted them together for accurate timing analysis. The centre speaker needs a distance reduction of 37 cm in the AVR to align with the FR,FL. The sub group needs 470 cm increase in distance in the AVR to best align with the 3 main speakers. The chart below shows the SPL result of a -15 dB sub SPL change and these AVR distance changes.

This still doesn't look very favorable compared to any resonable house curve. I really thought that if you remeasured you would find some issues with your measurements and replace them.

 

[Trdat]

Again I really appreciate this.

Looking at the FR I do admit that I should remeasure at least to double check and see if there is any issues lke you said. But, the better option is to see if there is a way to rearrange/reposition the speakers away from the wall to see if we can get rid of that dip. Better still, I can do some research with SBIR calculator and room modes calculator and see exactly where that dip is coming from and work with that. I will take on this option, my speakers are actually very good, and with a Hypex amp they have potential. So, I will spend time on this and eventually get back to you with better measurements. That will leave us with only the delay for the subs and centre to the FL/FR, that is when and if I get to it.

In saying that, the end result from the above might be that I can't reposition and that measurement will be exactly the same so if you don't mind helping me interpret the results as I always get confused if we increase the sub distance further away to the distance of the FR/FL are we adding delay to the sub or FR/FL? Your intructions were very clear so I made the corrections, just curious how the system works.

And one final measurement just to see the result now that I have added the new distances in, I mean I can take a look at step response but I am sure there are aother markers that can help you decide if we went about it the right way. I will post soon to see what the final result brings.

And if I turn Audyssey Correction on(not Dynamic Volume) will it mess up the delays. Obvioulsy I took measurements without it on but just curious?

 

[jtalden]

I always get confused if we increase the sub distance further away to the distance of the FR/FL are we adding delay to the sub or FR/FL?

Increasing distance of the sub is the same relative effect as adding delay to all the other speakers (not just the FR/FL).

The AVRs are smart enough to reinterpret internally all distance settings entered such that the most distant speaker setting is actually set to 0 ms delay. The other speakers then delayed accordingly to retain the correct relative delays between all speakers.

And if I turn Audyssey Correction on(not Dynamic Volume) will it mess up the delays. Obvioulsy I took measurements without it on but just curious?

I don't know.