Using measurements to improve soundstage/imaging?

[klettermann]

There's lot out there about imaging/soundstage and speaker positioning, room design, acoustic treatments, recording techniques, etc. But I've found next to nothing on measurements that might be used to improve matters. Is the reason that whatever contributes is outside the realm of measurement? Or that the specific setup governs how good it's going to be on a particular system? To be clear, my setup is already pretty excellent IMO (streamer -> MiniDSP -> DACs -> uber-capable amp -> rebuilt Magnepan MGIIIa's). It can be very holographic depending on source. Still, there's plenty to fool around with - toe-in, tweeters outside or inside, exact positioning (cm's matter), etc. Is there was some kind of measurement process that informs such stuff? Comments, ideas and discussion would be most welcome. Thanks, as always and cheers,

 

[Keith_W]

The measurements that are relevant are:

- ICCC or IACC (depending on your software). This measures how close the left and right speaker measurements are to each other with 1 as a perfect score and 0 showing no correlation whatsoever. 0.9 and above is excellent, 0.8 - 0.9 is good, and 0.6 - 0.8 is typical. DSP can improve the early component of ICCC but not the late component.
- Step response. Overlay left and right speakers and look to see how similar they are - not only the early part, but also the tail.
- ETC. Early and loud reflections smear the sound and ruin precision and imaging. The rule is - any reflection arriving <20ms to the main impulse needs to be a minimum of -15dB. Some say -20dB.
- Critical distance and Direct to Reflected Ratio (DRR). Sitting inside the critical distance means higher DRR, meaning more direct sound from the loudspeaker in proportion to the reflections. This is a matter of preference, but I have found that sitting inside the critical distance seems to improve stereo imaging.

There may be more. That's all I can think of for now.

 

[JeremyFife]

@Keith_W makes much more sense!

 

[Blumlein 88]

Once you confirm each channel matches as well as possible along the lines of what has already been suggested, I find lots of "soundstage" to be from small frequency response differences. A small bubble in FR from 2-5 khz brings things forward in the soundstage. A bit of dip in the same range moves them back in the soundstage. A shelving up below 200 hz can widen it some, and the opposite for shelving down. With the interactions between those you can to some minor extent sculpt the soundstage to your liking. Vocalists alone need slightly different FR ranges adjusted and can be moved forward or back some.

 

[Daverz]

Here's some info on how to use REW to measure the ETC:

Unpacking ETC: Time-domain measurements & early reflections -

Testing is a very useful tool to understand why rooms sound the way they do and provides information on how to improve the sound we experience in our rooms.

www.gikacoustics.com

 

[kemmler3D]

@Keith_W already shared some important measurements that could be used to predict the quality of soundstage in-room.

What to do with those measurements is another question - correcting the speakers to have better L/R matching, moving things around in the room, adding room treatments... all might help.

Soundstage is a subjective experience, so it's hard to say how it will change based on changes in the measurements.

 

[Keith_W]

By the way, since you mentioned toe-in, we should quickly discuss it. If you point your speakers straight ahead, you are listening to the off-axis response. This has a few effects:

1. It functions as a kind of tone control. High frequencies will roll off earlier since they usually have more directivity.
2. Because you are getting less direct sound, the DRR goes down and the critical distance becomes shorter (as you may have already guessed, the critical distance is frequency-dependent).
3. Side wall early reflections become louder and can potentially smear notes and therefore ruin imaging if the speakers are placed too close to the side walls. This is particularly pronounced in asymmetric listening rooms, especially rooms which are open on one side with a wall on the other side.

Some of these things can not be easily seen in measurements. For example, there is no way to tell which direction a reflection came from when you read your ETC. At best, you can calculate the delay, then look around your room and try to guess which surface was responsible. Also, you need to do special measurement techniques to find your critical distance, you won't find it with a normal sine wave sweep from the MLP. Actually, I suspect there may be a way, but I need to run my proposal past someone very clever on ASR first. Maybe @NTK or @youngho.

 

[klettermann]

Thanks much to all, very very helpful. The ETC thing and further explanation on the GIK site is perfect for me at this point. I haven't attacked reflections, decay or other room effects at all, apart from having big (and very old!) ASC tube traps in corners. Otherwise it's bare walls, carpet and soft couch. Fortunately the room is mostly a shoebox, though there's a column that gives me acoustic heartburn.

This also leads to some possibilities on the toe-in question. Given that the speakers are highly directional dipoles, might toe-in change the magnitude of early reflections? Also, unlike most speakers, the tweeter and midrange driver are not vertically aligned. So, when you mess with toe-in, you're also introducing tiny time alignment changes. People seem to jump through fiery hoops to time align drivers. Aiming directly towards the MLP would minimize that and maybe there are associated time or phase effects. It will be interesting to try. I fiddled with it in the past, but only looked at FR, not any of this stuff.

In conclusion, this gives me some better insights and tools to really start looking at the room and eventually see what happens to imaging with more tweaks and serious room treatment. The rabbit hole keeps getting deeper....
Thanks again and cheers,

 

[klettermann]

Once you confirm each channel matches as well as possible along the lines of what has already been suggested, I find lots of "soundstage" to be from small frequency response differences. A small bubble in FR from 2-5 khz brings things forward in the soundstage. A bit of dip in the same range moves them back in the soundstage. A shelving up below 200 hz can widen it some, and the opposite for shelving down. With the interactions between those you can to some minor extent sculpt the soundstage to your liking. Vocalists alone need slightly different FR ranges adjusted and can be moved forward or back some.

Ah ha! By coincidence I've got a wide band EQ device (Schiit Loki Max). It's useless for speaker EQing but perfect for what you suggest. More to play with! Cheers,

 

[kemmler3D]

might toe-in change the magnitude of early reflections?

I think you're still going to have ERs at comparable levels but the tonality of the ERs will tend to change more with toe-in with narrow radiation. As such I think you should expect some impact on soundstage. Have a look at ERs but also waterfall to see how that changes.

 

[klettermann]

I think you're still going to have ERs at comparable levels but the tonality of the ERs will tend to change more with toe-in with narrow radiation. As such I think you should expect some impact on soundstage. Have a look at ERs but also waterfall to see how that changes.

Yes! I'll report back on my findings. Might be little while though. Thanks and cheers,

 

[youngho]

Some of these things can not be easily seen in measurements. For example, there is no way to tell which direction a reflection came from when you read your ETC. At best, you can calculate the delay, then look around your room and try to guess which surface was responsible. Also, you need to do special measurement techniques to find your critical distance, you won't find it with a normal sine wave sweep from the MLP. Actually, I suspect there may be a way, but I need to run my proposal past someone very clever on ASR first. Maybe @NTK or @youngho.

Sorry, I'm not clever at all!

However, concepts like critical distance don't apply to small rooms since there is no true diffuse or reverberant field (this is discussed in Sound Reproduction). Can still try to approximate with various measurement parameters.

Also, in typical room setups, the first several ipsilateral reflections are usually relatively predictable.

Since you already pointed out issues with ETC, I wonder if you may be considering an iterative process of measuring DRR at different listening positions relative to the loudspeakers, as well as with different loudspeaker toe-ins.

It's also possible that you're trying to come up with a new metric along the lines of Griesinger's LOC (integrating loudness in a certain frequency range over one time frame relative to another) but in a much shorter timeframe, since early reflections in smaller rooms can occur within the first several milliseconds? There would be clear issues in such an approach, at least to my amateur eye...

Or possibly something along the lines of Genelec illustrates in their GRADE reports: https://www.genelec.com/-/blog/how-to-analyse-frequency-and-temporal-responses (under the Temporal Response section)? But you seem to be concerned about direct, as opposed to perceived direct (for which I'm assuming effects of auditory compression after ~1ms with frequency-dependent release, if I am referencing @j_j correctly from his 2021 PNW AES talk), so the time frames could potentially vary with frequency, which would be one of the issues above.

 

[dualazmak]

Many of the important and critical factors have been already nicely pointed by @Keith_W and other ASR colleagues.

Even though it is obvious and well understood, let me emphasize that all of the factors are interdependent with each other and much dependent on your own/individual actual room acoustic mode/environments including size/shape/furniture, etc.

Although I assume the point of "time alignment among the SP drivers" would be somewhat overlapped and/or related with some of the factors pointed by @Keith_W, I would like to suggest the importance of "0.1 msec precision time alignment among all the SP drivers" at your listening position.

I did rather primitive but intensive "time alignment" tunings in my multichannel multi-SP-driver multi-amplifier fully active audio setup as follows under the below spoiler cover;

Spoiler

- Precision measurement and adjustment of time alignment for speaker (SP) units:
Part-1_ Precision pulse wave matching method: #493
Part-2_ Energy peak matching method: #494
Part-3_ Precision single sine wave matching method in 0.1 msec accuracy: #504, #507
- Measurement of transient characteristics of Yamaha 30 cm woofer JA-3058 in sealed cabinet and Yamaha active sub-woofer YST-SW1000: #495, #497, #503, #507 (for time alignment and optimization of XO Fq between them.)

My post here would be also of your reference and interest;
- Not only the precision (0.1 msec level) time alignment over all the SP drivers but also SP facing directions and sound-deadening space behind the SPs plus behind our listening position would be critically important for effective (perfect?) disappearance of speakers: #687

In some setups (like my system), the wide-3D dispersion of high-Fq sound given by rather-narrow directivity tweeters and/or super-tweeters would effectively contribute to better soundstage/imaging (in my case, metal horn super-tweeter sound);

Spoiler

- A new series of audio experiments on reflective wide-3D dispersion of super-tweeter sound using random-surface hard-heavy material:
Part-1_ Background, experimental settings, initial preliminary listening tests: #912
Part-2_ Comparison of catalogue specifications of metal horn super-tweeter (ST) FOSTEX T925A and YAMAHA Beryllium dome tweeter (TW) JA-0513; start of intensive listening sessions with wide-3D reflective dispersion of ST sound: #921
Part-3_ Listening evaluation of sound stage (sound image) using excellent-recording-quality lute duet tracks: #926
Part-3.1_ Listening evaluation of sound stage (sound image) using excellent-recording-quality jazz trio album: #927
Part-4_Provisional conclusion to use Case-2 reverse reflective dispersion setting in default daily music listening: #929

All of these/those points have been implemented in my latest system setup, as you may find in this post, only if you would be interested in;
- The latest system setup of my DSP-based multichannel multi-SP-driver multi-amplifier fully active audio rig, including updated startup/ignition sequences and shutdown sequences: as of June 26, 2024: #931

 

[klettermann]

Many of the important and critical factors have been already nicely pointed by @Keith_W and other ASR colleagues.

Even though it is obvious and well understood, let me emphasize that all of the factors are interdependent with each other and much dependent on your own/individual actual room acoustic mode/environments including size/shape/furniture, etc.

Although I assume the point of "time alignment among the SP drivers" would be somewhat overlapped and/or related with some of the factors pointed by @Keith_W, I would like to suggest the importance of "0.1 msec precision time alignment among all the SP drivers" at your listening position.

I did rather primitive but intensive "time alignment" tunings in my multichannel multi-SP-driver multi-amplifier fully active audio setup as follows under the below spoiler cover;

Spoiler

- Precision measurement and adjustment of time alignment for speaker (SP) units:
Part-1_ Precision pulse wave matching method: #493
Part-2_ Energy peak matching method: #494
Part-3_ Precision single sine wave matching method in 0.1 msec accuracy: #504, #507
- Measurement of transient characteristics of Yamaha 30 cm woofer JA-3058 in sealed cabinet and Yamaha active sub-woofer YST-SW1000: #495, #497, #503, #507 (for time alignment and optimization of XO Fq between them.)

My post here would be also of your reference and interest;
- Not only the precision (0.1 msec level) time alignment over all the SP drivers but also SP facing directions and sound-deadening space behind the SPs plus behind our listening position would be critically important for effective (perfect?) disappearance of speakers: #687

In some setups (like my system), the wide-3D dispersion of high-Fq sound given by rather-narrow directivity tweeters and/or super-tweeters would effectively contribute to better soundstage/imaging (in my case, metal horn super-tweeter sound);

Spoiler

- A new series of audio experiments on reflective wide-3D dispersion of super-tweeter sound using random-surface hard-heavy material:
Part-1_ Background, experimental settings, initial preliminary listening tests: #912
Part-2_ Comparison of catalogue specifications of metal horn super-tweeter (ST) FOSTEX T925A and YAMAHA Beryllium dome tweeter (TW) JA-0513; start of intensive listening sessions with wide-3D reflective dispersion of ST sound: #921
Part-3_ Listening evaluation of sound stage (sound image) using excellent-recording-quality lute duet tracks: #926
Part-3.1_ Listening evaluation of sound stage (sound image) using excellent-recording-quality jazz trio album: #927
Part-4_Provisional conclusion to use Case-2 reverse reflective dispersion setting in default daily music listening: #929

All of these/those points have been implemented in my latest system setup, as you may find in this post, only if you would be interested in;
- The latest system setup of my DSP-based multichannel multi-SP-driver multi-amplifier fully active audio rig, including updated startup/ignition sequences and shutdown sequences: as of June 26, 2024: #931

Wow. This is a lot to digest. But by coincidence it's along the lines of what I was doing yesterday, specifically getting time alignment between respective speakers as good as I possibly could. I did so using REW, which reports delay down to 4 significant figures (0.000X ms). First I got all the speakers down to <0.1 ms delay. Then I kept adjusting in 0.01ms increments. My experience was that 0.01 increments gave pretty reproducible results, but numbers below that were noise. Then there's the question of driver alignment. I've already got the toe-in angled such that the speaker panel midpoints are a precisely perpendicular to the MLPS. EQ and XO weren't changed.

So guess what? It actually did make enough difference that my partner noticed a difference. Clarity and crispness were better, soundstage was deeper and more 3 dimensional. On Madonna's song Vogue the opening voices moved dramatically from where they were before, to the point that they were coming directly from the sides of the MLP like surround speakers. Actually kind of unnerving! I haven't yet scrutinized the scans or looked at anything other than delay. But I can say that precision time alignment moved the needle. I hope I'm not in the realm of cable lifters or audiophile ethernet switches with this.

Look like I've got some reading to do. Thanks and cheers,

 

[dualazmak]

Thank you for your above positive response!

Your visits and participation on my project thread will be much welcome.


I assume you would be interested also in these other threads which I have been hosting...

- Let's share diagrams (and photos) of our total physical audio system and the whole signal path, with a few words and/or links
- https://www.audiosciencereview.com/forum/index.php?threads/lets-share-diagrams-and-photos-of-our-total-physical-audio-system-and-the-whole-signal-path-with-a-few-words-and-or-links.56410/

- An Attempt Sharing Reference Quality Music Playlist: at least a portion and/or whole track being analyzed by 3D color spectrum of Adobe Audition
- https://www.audiosciencereview.com/...by-3d-color-spectrum-of-adobe-audition.47103/

- Music for Testing Treble (High Frequency) Sound
- https://www.audiosciencereview.com/...or-testing-treble-high-frequency-sound.50690/

- Apparent power consumption of whole audio system during daily audio listening sessions: how SDGs-friendly is it? (not the idle power, please.)
- https://www.audiosciencereview.com/...iendly-is-it-not-the-idle-power-please.51987/

- Lute Music and Other Early Music: Stunning Recordings We Love
- https://www.audiosciencereview.com/...arly-music-stunning-recordings-we-love.21247/

 

[DVinylfan]

I have an anecdote re. stereo imaging which I hope isn't out of place on this thread.
It's circa 1987, and I have recorded a mixtape in my bedroom using very basic Pioneer turntable and integrated amp bought used from a local auction, onto a compact cassette using an Aiwa ADF-xxx tape deck.
The track I'm recalling is The Cure's "Close to Me", and the specific element is the close-miked breath sounds at the beginning of the track.
One friday evening I brought my cassette to my friend's house to play it on his Dad's professionally installed living room system, which comprised a Nakamichi Dragon cassette deck, (again IIRC) Quad 33/405 amplifier, and B&W DM220 loudspeakers.
When I heard this track playing on this system I was amazed to hear the 'breath sounds' at the beginning of the track seem to hover in front of me, as if they were physically in the room. I had never heard anything approaching this level of stereo imaging and 'realism' listening to the track on the system it was recorded on, and my Mission 700 speakers (budget speakers then but now command crazy money?) on shelves against the wall.
Back then, my conclusion from this experience was that, since the signal chain recording was every component I had except the speakers, there must be something exceptional about the B&W DM220, and I coveted a pair for decades.
Recently I have begun to understand better what many here take for granted, which is that not only the quality of the speakers but their placement and the conditions of the listening room are critical concerning the quality of the stereo image.