• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Polk LSiM 703 Speaker Review (@ObjectAudio)

ObjectAudio

Member
Forum Donor
Joined
Jan 22, 2021
Messages
95
Likes
448
The speaker
Its scores and numbers, as in the report



*Updated 6 Nov 2026:
Here is an update to the review report. I remeasured the speaker in a slightly more spacious, yet relatively tight, space compared to my office, updated the layout style, and added a conclusion with an objective performance comparison against the Spinorama.org database and some fixes based on thread feedback.

For anyone curious about the accuracy and validity of this measurement method, you can see a direct comparison of my system measuring a KEF R3 against Erin's and ASR's Klippel NFS measurements here:
https://www.audiosciencereview.com/forum/index.php?threads/how-good-can-a-home-spinorama-in-a-tight-space-be-compared-to-klippel-—-kef-r3-study.74303/#post-2715476




Here is a review and detailed measurements of the Polk LSiM 703, which I purchased in the crazy Adorama sale like 10 years ago along with its siblings, the LSiM 705 and the LSiM 706C center channel.

I think they are great speakers—very smooth, never in your face, and truly enjoyable speakers that look great.

All measurements, data processing, and plotting were performed using LWaves Speaker Lab, a custom measurement software suite I am currently developing. The resulting measurements are very promising, and when paired with a bit of EQ and a subwoofer, the predicted preference score ranks among the highest you will find.

As we can see compared to my previous review of the Philips SoundSphere speakers, completely different speaker architectures can achieve with some EQ a very similar ideal with subwoofer preference scores, yet they will sound totally different. This is mainly due to their distinct radiation patterns and beaming behaviors: while the Polk’s controlled directivity produces sharp, precise imaging, the wider off-axis dispersion of the SoundSphere creates a much more diffuse and spacious soundstage.


The raw speaker​



Spinorama​


CTA-2034 spinorama​


CTA-2034 spinorama



On-axis, listening window, early reflections and sound power, with the two directivity indices below. Flat, parallel curves with smooth directivity indices are what listeners prefer.


Early reflections​


Early reflections



The early reflections by wall - floor, ceiling, front, side and rear - and their total.


Predicted in-room response​


Predicted in-room response



What a typical room makes of the speaker. A gentle downward tilt of about 1 dB per octave tends to sound balanced; dashed: Harman’s in-room target.


Frequency response​


Response linearity​


Level = 0 dB · 85.0 dB · on-axis, 0.3 - 3 kHzLinearity · −2.4 / +2.8 dB · 80 Hz - 16 kHzWithin ±1.5 dB · 87 % · of that rangeWithin ±3 dB · 100 % · of that rangeF3 · 46 HzF10 · 38 Hz

Response linearity



Each curve against the speaker’s own level: green ±1.5 dB, amber ±3 dB.


Driver components - near field​


Driver components - near field



Each driver and the port measured close up. * = scaled to the woofer’s area.


Directivity​


Horizontal off-axis responses​


Horizontal off-axis responses



Every 10° from 0° to 90° to one side.


Vertical off-axis responses​


Vertical off-axis responses



Every 10° up to ±40° above and below the listening axis (below dashed).


Horizontal directivity​


Horizontal directivity



Level against angle and frequency. Even, gradual narrowing is the goal.


Vertical directivity​


Vertical directivity



The same in the vertical plane.


Horizontal polar​


Horizontal polar



The horizontal pattern at a set of frequencies.


Vertical polar​


Vertical polar



The vertical pattern at a set of frequencies.


Directivity balloons​


Directivity balloons



The sound in every direction at three frequencies: round = all directions, a narrow lobe = beaming.


Horizontal beamwidth​


Horizontal beamwidth



The angle each side where the level falls 6, 12, 18, 24 and 30 dB below the on-axis. Lines close together and level = a steady beam. The −6 dB beam does not hold one angle (±15°) up to 10 kHz: it spans 94° from 1 to 10 kHz.


Vertical beamwidth​


Vertical beamwidth



The same in the vertical plane (+ up, − down). The −6 dB beam does not hold one angle (±15°) up to 10 kHz: it spans 96° from 1 to 10 kHz.


Horizontal off-axis waterfall​


Horizontal off-axis waterfall



Every horizontal angle: 0° (on axis) in the middle, ±180° (behind) at the ends.


Impedance​


Impedance and phase​


Impedance and phase



The load the amplifier sees: magnitude (Ω) and phase.


Time domain​


Phase and group delay​


Phase and group delay



The speaker’s phase and its group delay.


Step response​


Step response



How the drivers arrive in time: one smooth rising step means the drivers are time-aligned.


Cumulative spectral decay​


Cumulative spectral decay



How the sound dies away after the impulse; a ridge that lasts is a resonance.


Burst decay​


Burst decay



How fast each frequency dies away, in periods; a long dark ridge is a resonance.


Distortion​


THD at every level​


THD at every level



Total harmonic distortion at every level, at 1 m - lightest the quietest, black the loudest.


Harmonic distortion at 86 dB @ 1 m​


Harmonic distortion at 86 dB @ 1 m



THD, H2 and H3 at 86 dB SPL at 1 m.


Harmonic distortion at 96 dB @ 1 m​


Harmonic distortion at 96 dB @ 1 m



The same at 96 dB SPL - where the drivers’ limits begin to show.


Dynamic range​


Compression at rising levels​


Compression at rising levels



Each level against the quietest, the drive difference taken off: 0 dB = no change, a dip = compression. Levels: 76 / 86 / 96 dB SPL at 1 m. Grey: the quietest level repeated (±0.04 dB typical) - only a change bigger than the grey is the speaker’s.


Multitone​


Multitone distortion (20 Hz - 20 kHz)​


Multitone distortion (20 Hz - 20 kHz)



32 tones played together; everything between them is distortion, in dB below the tones. Levels: 76 / 86 / 92 dB SPL at 1 m.


Compression from multitone (20 Hz - 20 kHz)​


Compression from multitone (20 Hz - 20 kHz)



Each tone’s level change against the quietest level. Grey: the quietest level repeated (±0.32 dB typical) - only a change bigger than the grey is the speaker’s. Read 20 cm from the cone; below 100 Hz at 1 m (the port is on the back).


Maximum SPL​


Maximum SPL - the speaker with its amplifier​


Maximum SPL - the speaker with its amplifier



The loudest it can play at each frequency. Black: the limit - the amplifier (Fosi Audio V3 · 48 V 10 A, clips at 30.5 V) or the speaker’s own, whichever comes first. Green: the speaker’s own limit (3 dB compression or 10 % THD), 20-48 Hz. Shaded: still clean at 96 dB. Yellow: music (12 dB crest) reaches 96.7 dB SPL at 1 m - first limited by the amplifier’s clip. Excursion is not modelled (no T/S data).



With its EQ​



CTA-2034 spinorama - with its EQ​


CTA-2034 spinorama - with its EQ



As above, with its EQ.


Predicted in-room response - with its EQ​


Predicted in-room response - with its EQ



As above, with its EQ.


Response linearity - with its EQ​


Level = 0 dB · 85.1 dB · on-axis, 0.3 - 3 kHzLinearity · −3.2 / +1.1 dB · 80 Hz - 16 kHzWithin ±1.5 dB · 86 % · of that rangeWithin ±3 dB · 99 % · of that rangeF3 · 46 HzF10 · 38 Hz

Response linearity - with its EQ



As above, with its EQ.


Maximum SPL - the speaker with its amplifier - with its EQ​


Maximum SPL - the speaker with its amplifier - with its EQ



As above, with its EQ. Yellow: music (12 dB crest) reaches 97.3 dB SPL at 1 m - first limited by the amplifier’s clip.


The EQ used​


EQTypeFrequencyGainQ
The EQpeak65 Hz-1.0 dB8.00
peak110 Hz-1.9 dB2.10
peak432 Hz+1.3 dB6.78
peak845 Hz+1.3 dB8.00
peak2.44 kHz+1.3 dB1.60
peak2.95 kHz-2.3 dB1.48
peak4.37 kHz-1.4 dB2.70
peak6.62 kHz-2.7 dB2.05
peak7.58 kHz+5.7 dB1.36
peak9.6 kHz-4.5 dB1.06

The EQ​


The EQ



The EQ’s response (minimum phase, as a DSP plays it).


Conclusion - How it ranks​


Conclusion - How it ranks


Against every speaker spinorama.org holds measured with a Klippel NFS, all scored with the same formulas.

Data: spinorama.org measurement data - measurements by Audio Science Review, Erin’s Audio Corner and others, each credited on spinorama.org - licensed CC BY-NC-SA 4.0. Recomputed with this app’s own formulas. This ranking is shared under the same licence.


How it was measured
  • Microphone: OmniMic (calibrated), at 1.0 m.
  • The spin: 36 horizontal and 36 vertical angles.
  • In-room, gated at 4.8 ms - room-free above about 210 Hz.
  • The bass below 215 Hz from near-field close-ups (the woofer, the port and the mid), joined to the far field.
  • Distortion and compression: sweeps at 76 / 86 / 96 dB SPL at 1 m, close-ups below 607-623 Hz (each level its own join).
  • Multitone (32 tones): 76 / 86 / 92 dB SPL at 1 m, also read 20 cm from the cone.
  • Maximum SPL: calculated for the Fosi Audio V3 · 48 V 10 A amplifier.
  • Limits: one unit, measured in a normal room with a gated spin.
  • The spinorama follows ANSI/CTA-2034-A; the preference rating is the Olive model.
 
Last edited:
Most of us (ageing audiophile men) can barely hear above 12-15K. AFAIC - results up there are largely academic.
Absolutely agree on the first, the strong beaming though starts rather already around 7-8 kHz and I had similar Vifa tweeters where I would definitely notice this behaviour in my listening sessions.
 
i have 4 of these,
have been thinking about upgrading but i wouldn't know with what,
Kef r3 /q7 ? philharmonic BMR? mofi sp8? Wharfedale Linton? Wharfedale EVO 4.3?
any recommendations on a 3 way bookshelf or compact tower that has good LFX and wide directivity on the HF?
measures well frequency response and distortion limits?
 
I loved the mid-range on these speakers for percussion but I found the treble very dull in my room at the time. The bass was solid. They are heavy little suckers.
 
I enjoyed a pair of these for about 10 years. No wonder I liked them so much. The beaming of the tweeter did make them picky about listening angle, but you only set that once, so not a big deal.
 
Nice review and did not know this model )

Teardown and alternative measurements


 
I used to have these.
At one point they were on clearance in the USA at Adorama for many months on and off at insanely low prices. ($400-450 per pair)
Very high build quality in the hands.
Very high quality drivers for the commercial price point.
Very high power handling but I did not not care for their sound when driven hard to very hard.
There was no notable distortion, I attribute the subjective dissatisfaction to the narrowing highs.
I've noticed this trait in few speakers and I don't care for it.
There is something weird about it that makes things sound 'hardened' to me along with a simultaneously forward midrange and a weird 'dullness' to me. This is very unnatural sounding to me in my room(s).
At lower and medium SPL I did not notice it as much.
I stay away from these type of designs now after not liking several with this polar pattern. (often they use a ring radiator tweeter)
I was able to double my money when I sold them used though due to the initial sale price and then selling well after the sale ended, was a fun trial.
 
SPL by deg looks it doesn't match as right degree. There are several glitches in contour plot and line chart. Same problem in your previous review too.
Maybe if you are using something like auto coding machine there might be some errors. (My personal experience)
You should check if the SPL/Deg file names are matched correctly.
Anyway, great job!

스크린샷 2026-10-27 오전 8.52.27.png
스크린샷 2026-10-27 오전 8.50.53.png

스크린샷 2026-10-27 오전 8.54.09.png
 
Level against angle and frequency, horizontal plane (contour map). Even, gradual narrowing with frequency is the goal; sudden widening or narrowing shows a crossover or a baffle effect.

I guess ±60 degrees at that frequency is too wide for most tweeters to handle.

Generating wide angle at high frequencies for (pseudo-)omnidirectional speakers has research potential
 
SPL by deg looks it doesn't match as right degree. There are several glitches in contour plot and line chart. Same problem in your previous review too.
Maybe if you are using something like auto coding machine there might be some errors. (My personal experience)
You should check if the SPL/Deg file names are matched correctly.
Anyway, great job!

View attachment 567233View attachment 567234
View attachment 567236
Thanks for catching that — you were right that something was off, though it wasn't the file names: those are matched correctly (the turntable stamps every angle automatically). Two other things caused the glitches:

1. A software bug: a low-frequency correction step (used to handle room reflections below about 1 kHz) set the side and rear angles several dB too low, creating the artificial stripes and drops around ±90° to 120°. I've already fixed it.
2. Room reflections: my office is a less-than-ideal acoustic space while I'm developing the program, and some side and rear measurements caught a reflection from something close behind the speaker. That caused the high-frequency ripples and the thin lines in the contour plots.

The overall score barely changes, because those off-axis angles contribute very little, but I'll re-measure the side and rear angles with a clearer space and repost clean data. Thanks again for the close look!
 
Thanks for catching that — you were right that something was off, though it wasn't the file names: those are matched correctly (the turntable stamps every angle automatically). Two other things caused the glitches:

1. A software bug: a low-frequency correction step (used to handle room reflections below about 1 kHz) set the side and rear angles several dB too low, creating the artificial stripes and drops around ±90° to 120°. I've already fixed it.
2. Room reflections: my office is a less-than-ideal acoustic space while I'm developing the program, and some side and rear measurements caught a reflection from something close behind the speaker. That caused the high-frequency ripples and the thin lines in the contour plots.

The overall score barely changes, because those off-axis angles contribute very little, but I'll re-measure the side and rear angles with a clearer space and repost clean data. Thanks again for the close look!
Actually I don't see it is just a merging issue. Seeing 80,90 deg in this measurement, two responses look not having continuous change.

1793062915021.png

1793063088337.png


See -120 deg. That kind of vivid glitch means that the measurement to deg is matched incorrectly.
 
i have 4 of these,
have been thinking about upgrading but i wouldn't know with what,
Kef r3 /q7 ? philharmonic BMR? mofi sp8? Wharfedale Linton? Wharfedale EVO 4.3?
any recommendations on a 3 way bookshelf or compact tower that has good LFX and wide directivity on the HF?
measures well frequency response and distortion limits?

If you want to try a wide directivity and a spacious sound the BMR is very good. I owned them and they are a type of 10/10 given the price and result, but you will have to listen yourself and decide if you prefer the dispersion characteristics to the LSiM703s. They have nice bass reach but limited total maximum output due to the smaller woofers being asked to dig extra deep. There is no replacement for displacement.
Some of this will be room and over-all set-up dependent as well.

I also really like some of the JBL designs using some of the newer HDI waveguides (4309 and 4329)
The beam width is shockingly even up into the high highs with these and they sound so good to me.
The level of subjective power and realism just makes some speakers sound weak/ not realistic even if they are smooth and pleasant.
The 4309 speakers are light in the bass and require sub(s). With a tweak in the mids for PEQ and high passed at about 80-90hrz to subs they sound incredible to me.
You could also try the Pro line with the 708p (not '708i), but they look 'Pro' not lux wood grainy.

The KEF R3 is a SOTA concentric design for the price point. That said I find them slightly boring, I just do not tap my toes so to speak. With the JBLs I want to dance around and get fired up go into track after track. FWIIW I have owned and tested the R3s 2 times. The bass also suffers from some strong harmonic distortions that I do not like so I would high pass them and not run them full range at higher SPLs. They do not sound subjectively clean in lower bass to me full range above 'medium high'.

A darker, warmer, smoother, more conservatively rolled off design I do like is the old Boston Acoustics VS260. It has a euphonic quality that I seem to consistently experience with later Boston designs. It is wonderful but not for everyone and not a daily driver for me. (oddly enough they also use a ring radiator design, but I like these speakers)

I have had the Linton's in house and never got into them, wanted to but just did not like them very much. They have a slight 'dirty' & slightly irritating sound to me subjectively and yet many folks swear by them. Try a pair with a good return policy available. I got mine super cheap and so could sell them to recoup easily.

Who knows, you will have to try some stuff out. (or skip the crazy rabbit hole and just keep your Polks) It will all be sighted so expect to have some biases (and make some things up) unless you happen to be one of the very few who don't have many. For me some speakers sound really good and yet I can turn it off without missing a beat and others for some reason cause me to put track after track on. I keep those. Probably all my imagination but what the hell.
 
Last edited:
Actually I don't see it is just a merging issue. Seeing 80,90 deg in this measurement, two responses look not having continuous change.

View attachment 567246
View attachment 567247

See -120 deg. That kind of vivid glitch means that the measurement to deg is matched incorrectly.

Actually I don't see it is just a merging issue. Seeing 80,90 deg in this measurement, two responses look not having continuous change.

View attachment 567246
View attachment 567247

See -120 deg. That kind of vivid glitch means that the measurement to deg is matched incorrectly.
This is the Philips Speakers not the Polk and you're right—the 80° and 90° curves don't match the others, and those lines in the plot are due to bad measurements, not the speaker. Sound bounced off me and objects in my office behind the speaker during the spin, which spoiled several measurements on one side (causing that sudden change above +70°) and one at −120°. My software now spots these automatically.

This speaker is symmetrical on both sides, so I replaced the bad measurements with the clean ones from the opposite side, and now the curves change smoothly. However, the −120° measurement was bad on both sides, so I'll measure it again. I'll post new plots when I have a chance to measure them in a larger space.and again overall score barely changes. Thanks for spotting it!
 
Great review, thank you. Just a sugestion. A little more basic info at the beginning. Driver size (cant tell from pics), active/passive (for a second I thought they were active), sensitivity, price.
 
Back
Top Bottom