2.1 Stereo Calibration - what am I missing?

[sm52]

The biggest breakthroughs for me were discovering how to best align the step response

I also believe that the step response is more important than the impulse graph. But also impulses from different frequency bands should not be arranged randomly. Can you articulate what step response you are aiming for?

 

[moedra]

I also believe that the step response is more important than the impulse graph. But also impulses from different frequency bands should not be arranged randomly. Can you articulate what step response you are aiming for?

The target is the ideal response, usually. However, every set of speakers is going to measure differently, so the resulting step response will look different from configuration to configuration. The goal is more about the alignment of the left and right step responses. The less different they are, the tighter the bass response is. I align them by equalizing the measurements such that the left and right group delays match when smoothed. The step responses line themselves up in stride.

 

[serko70]

Here's my two cents on the subject...

I agree that group delay similarity between left and right speakers has major influence on sound stage improvement. When done properly you will hear the sound stage widen, become more focused and even rise up! I use excess phase inversion to equalize group delays but almost any attempt to remove excess group delay below 200Hz will cause ringing in most rooms as in fact what you're doing is trying to digitally move the walls in the room. One way to deal with this is to accept a certain group delay and invert each speakers excess phase just enough to reach the average. This is easy to do in the new REW with a series of TA averaging, inversion and magn division.

Another method I am working on is to simulate a healthy group delay for the system in rePhase with the help of a rotated sharp roll off crossover filter linearization around 100Hz:

The filter above creates a base GD for my system (purple line below) and the excess phase inversion required to equalize both speakers (blue and green lines) to that base drops down to around 30ms rather than 70ms otherwise and pre-ringing can be avoided:

Also note that you will have minor magnitude changes in the freq. domain (improvements) in the 200-300Hz area when you correct GD there and final FR EQ should be done only afterwards.

 

[moedra]

Also note that you will have minor magnitude changes in the freq. domain (improvements) in the 200-300Hz area when you correct GD there and final FR EQ should be done only afterwards.

When the group delays line up, so do the step responses. I don't try to remove the delays, however. Doing that kills the sound. I just match them up using some smoothing to make the task easier.

I do the frequency magnitude correction and phase optimization in two separate layers, isolated from each other. This way, adjusting the group delays doesn't really touch the magnitude correction passes, which happen on a different level and are later combined. However, the magnitude correction does in fact help align the group delays, so I process the frequency correction first.

I'm interested specifically in how to invert the excess phase and apply it to my process, if possible. I wonder how much that would improve things, if it can be integrated. My only concern is that it would break my procedure!

 

[serko70]

When the group delays line up, so do the step responses. I don't try to remove the delays, however. Doing that kills the sound. I just match them up using some smoothing to make the task easier.

I do the frequency magnitude correction and phase optimization in two separate layers, isolated from each other. This way, adjusting the group delays doesn't really touch the magnitude correction passes, which happen on a different level and are later combined. However, the magnitude correction does in fact help align the group delays, so I process the frequency correction first.

I'm interested specifically in how to invert the excess phase and apply it to my process, if possible. I wonder how much that would improve things, if it can be integrated. My only concern is that it would break my procedure!

Just to experiment, I sometimes fully invert excess phase without even any smoothing from 100Hz to 1000Hz (this is the range we're truly sensitive to phase shifts) for each speaker. The sound is literally perfect after that. Both speakers are in full harmony, I can locate every single instrument in the sound stage, etc. But then I play one of my pre-echo test songs (Billy Jean is a very good one btw) and hear the enormous ringing. If you will never listen to any song with pre-ringing potential, you will hear all other songs at their best with full inversion

 

[serko70]

When the group delays line up, so do the step responses. I don't try to remove the delays, however. Doing that kills the sound. I just match them up using some smoothing to make the task easier.

I do the frequency magnitude correction and phase optimization in two separate layers, isolated from each other. This way, adjusting the group delays doesn't really touch the magnitude correction passes, which happen on a different level and are later combined. However, the magnitude correction does in fact help align the group delays, so I process the frequency correction first.

I'm interested specifically in how to invert the excess phase and apply it to my process, if possible. I wonder how much that would improve things, if it can be integrated. My only concern is that it would break my procedure!

Are you applying linear and/or min phase filters as necessary to shift the "response" to the step to remove left and right speaker differences in their step response? This is brilliant but it would take a lot of manual trial & error work?

 

[moedra]

Just to experiment, I sometimes fully invert excess phase without even any smoothing from 100Hz to 1000Hz (this is the range we're truly sensitive to phase shifts) for each speaker. The sound is literally perfect after that. Both speakers are in full harmony, I can locate every single instrument in the sound stage, etc. But then I play one of my pre-echo test songs (Billy Jean is a very good one btw) and hear the enormous ringing. If you will never listen to any song with pre-ringing potential, you will hear all other songs at their best with full inversion

I'm going to try to figure out how to integrate your ep inversion technique and see what happens. Should be interesting.

 

[moedra]

Are you applying linear and/or min phase filters as necessary to shift the "response" to the step to remove left and right speaker differences in their step response? This is a brilliant but it would take a lot of manual trial & error work?

Yes. It is labor-intensive, and it involves repeatedly switching back and forth between the EQ window, the overlays window (to visually align the group delays), and the main interface (to delete excessive iterations) to pull it off. I figured out that I can manipulate the group delays by adjusting the EQ of the predicted result/measurement of the frequency correction (minphase) smoothed to something like 1/6 or psychoacoustic. Now if only John could add a group delay overlay into the EQ window, I wouldn't have to do all that switching...

Anyway, I have found that focusing on the group delay below 200Hz yields the best results... anything higher doesn't seem to make much difference because it typically isn't that far off to begin with. But as the sub-200Hz group delays are lined up (I typically try to average them between each other), the step response coherence cleans up.

 

[sm52]

The goal is more about the alignment of the left and right step responses

Thanks for the answer. Speaking of the perfect Step Response, can you show two Step Responses, one before correction, one corrected, similar to perfect? From real measurements.

 

[moedra]

Thanks for the answer. Speaking of the perfect Step Response, can you show two Step Responses, one before correction, one corrected, similar to perfect? From real measurements.

Sure. Here's one I did recently. This one's results were particularly good. Good acoustic space, plus the actual measured filter looks better than the simulation, which may be because the volume was slightly lower. That's a possibility...

Raw (before):

Simulation of result (multiply filter by raw measurement):

Measured filter:

 

[moedra]

Thanks for the answer. Speaking of the perfect Step Response, can you show two Step Responses, one before correction, one corrected, similar to perfect? From real measurements.

As another example, here is Dave's (thread starter) session. His original measurements are the gray traces, and the refined step responses are in color. This isn't the simulation, however. This is just what I get after aligning the group delays. This is the target for his measurements.

He'd have to measure his filtered response from the same spot as before for us to see what actually happens. It will not be that perfect, unfortunately.

 

[Dave the Rave]

@moedra We finally got the awning screen installed behind the TV. Ran another measurement as attached (with acoustic timing).
The speakers and sub positions moved a little and feel the music has gone slightly more shrill than before. I applied a filter on the first R&L measures which sounded good but not to the level of your filters.
Appreciate your help to update the convolver. Considering the sound change, I am thinking the curve slope at 3db on both ends might help.
Thanks again.

 

[serko70]

@moedra We finally got the awning screen installed behind the TV. Ran another measurement as attached (with acoustic timing).
The speakers and sub positions moved a little and feel the music has gone slightly more shrill than before. I applied a filter on the first R&L measures which sounded good but not to the level of your filters.
Appreciate your help to update the convolver. Considering the sound change, I am thinking the curve slope at 3db on both ends might help.
Thanks again.




View attachment 58609

Did you move the microphone during these measurements? Some of them are 2 to 4 meters off?

 

[serko70]

DAC5L1 and DAC5 R2 seem to be the only consistent measurements timing-wise and they reveal a HF mismatch possibly due to left speaker toe-in adjustment (or a tweeter level adjustment in your Yamaha speakers' specific case):

You must be changing something during measurements which is causing massive time delays unless your equipment is faulty and generating random delays on its own. Also remember it's good practice to measure 0-24kHz range with 48kHz sampling rate. Your measurements end at 20kHz.

 

[John Mulcahy]

The ones with large apparent delays have big clock adjustments. Possibly dropouts during measurement.

 

[Dave the Rave]

Thanks. I left the microphone intact for all measurements - so unsure on the behaviour and which equipment may be causing the mismatch. I did dial down the mid and high speaker squawker volume to the same level this time. Also the toe-in on the right was more than the left. So I will re-measure again with the adjustments.

 

[moedra]

Thanks. I left the microphone intact for all measurements - so unsure on the behaviour and which equipment may be causing the mismatch. I did dial down the mid and high speaker squawker volume to the same level this time. Also the toe-in on the right was more than the left. So I will re-measure again with the adjustments.

If possible, put something along those side walls so as to at least soften your first reflection points. It doesn't have to be anything huge or bulky, just something to soften the reflections. The bare walls are not ideal nor are they helping you get the best sound from your system. If there is anything you can do there, do it before you measure again.

 

[ddude003]

A few stands of potted bamboo or live/dried potted plants along the side walls works well... Wooden planters filled with sand and/or charcoal or fish aquarium carbon works too... Try and keep the left and right sides symmetrical...

 

[Dave the Rave]

Hi Folks,
I just wanted to share my real appreciation of the REW founders and the community in here. Over the years, I have been investing in all types of audio upgrades - speakers, cables, DAC, sources, power feed, etc....Though these made some positive changes, DRC made the MOST significant change in my system (and at little cost - just bought a UMIK mic for the measurements). So few things I learnt:
- Room correction is a MUST - else I am always sucked into the spending/upgrade rabbit hole
- Leant a lot from senior members here who are the real deals - Really appreciate the help and support from @John Mulcahy @moedra @serko70.
- Speaker positioning is critical.... ideal speaker placement and toe-in through REW measurement identified the optimum positioning where sound became much cleaner even without filters.
- The bit I have work to do is room treatment which I am told is another must-do - i do have constraints in this dept which I will address over time
- I found keeping a simpler music chain brings the most out my system - overcomplicating/ adding software layers (e.g upscaling) etc.... just created placebo effects with no real impacts

Again I wanted to say a MASSIVE thanks to @moedra and @serko70 for taking their time (with no question asked) providing filters which I personally would never be able to come up with (I am a newbie). Respect to you.
I am currently re-discovering all my tracks enjoying the enriched details, clarity, transient I thought I would never get though the speakers (but I could hear through a headphone). Really happy I got to know REW and this community.

 

[moedra]

Hi Folks,
I just wanted to share my real appreciation of the REW founders and the community in here. Over the years, I have been investing in all types of audio upgrades - speakers, cables, DAC, sources, power feed, etc....Though these made some positive changes, DRC made the MOST significant change in my system (and at little cost - just bought a UMIK mic for the measurements). So few things I learnt:
- Room correction is a MUST - else I am always sucked into the spending/upgrade rabbit hole
- Leant a lot from senior members here who are the real deals - Really appreciate the help and support from @John Mulcahy @moedra @serko70.
- Speaker positioning is critical.... ideal speaker placement and toe-in through REW measurement identified the optimum positioning where sound became much cleaner even without filters.
- The bit I have work to do is room treatment which I am told is another must-do - i do have constraints in this dept which I will address over time
- I found keeping a simpler music chain brings the most out my system - overcomplicating/ adding software layers (e.g upscaling) etc.... just created placebo effects with no real impacts

Again I wanted to say a MASSIVE thanks to @moedra and @serko70 for taking their time (with no question asked) providing filters which I personally would never be able to come up with (I am a newbie). Respect to you.
I am currently re-discovering all my tracks enjoying the enriched details, clarity, transient I thought I would never get though the speakers (but I could hear through a headphone). Really happy I got to know REW and this community.

You're most welcome!

 

[Jojo10]

Hello Forum

I'm coming here to ask some questions after watching the tutorials of @Moadra on Youtube. They are great, thanks for that.
Best description of myself would be "All the gear, but no idea". DYI Speakers as a FAST of a Vifa fullrange Speaker and a Visaton TIW200XS with an crossover at 500Hz.
MM-1 with individual calibration file, UMC-22 and REW as measurement setup. After using a IMG Stage Line DSM-240 for the last 10 years, I switched to a MiniDSP Flex for optimization purposes.
I would like to try to setup the crossover and the room correction with FIR filters for my speakers following the instructions of @Moadra.
It provides 4x 2048 taps on the outputs and in addition PEQs for the two inputs.
1. Does it make sense to use FIR filters for crossover (LR4) and linearization of magnitude optimized with an measurement at 1m distance between the drivers and than use the PEQ of the inputs to correct the room over all at hearing position?
Should be the goal here to use minimum phase or zero phase approach (I do not understand, what each implies ;-))?
Should I adjust the GD of each driver (BB and Bass) between left and right individually or as a sum at the hearing position?
2. Using FIR for crossover and room correction: As above corrected at 1m only in the area of crossover and baffle step plus using the second banks on all outputs to correct room modes and modeling room curve at the top? This would mean to use the room correction filter twice on bank2 for the bass and the fullrange speaker to cover complete frequency spectrum.
3. Using PEQs for crossovers and FIR for room correction.

So much for today. Help would be highly appreciated.

Thanks

Jojo

 

[John Mulcahy]

I would like to try to setup the crossover and the room correction with FIR filters for my speakers

Why do you want to use FIR? Seems like the wrong choice to me.

 

[Jojo10]

Hello John
Exactly, why? I understood that with FIR filters I can optimize not only in frequency domain, but also in time domain. Doesn't this bring benefits in terms of imaging for example? Isn't the adjustment of GD, like David showed it, only possible with FIR filtering?
Your opinion and guidance would be highly appreciated.

Thanks

 

[John Mulcahy]

FIR filtering is a very inefficient way to implement a crossover filter, better using the crossover filters the Flex offers. Room effects at low frequencies are minimum phase phenomena that are properly countered by parametric EQ filters. FIR filtering can be used to change the phase response of a speaker independently of its magnitude response though there's little consensus on the audibility of the result.

 

[serko70]

How about a minimum phase FIR filter: