The Kinera Celest Yaksha ranks #598 out of 1061 total In-Ear Headphones for sound quality. See all Headphones rankings for every top pick.
Brand: Kinera
Prices on this page were last checked October 15, 2026 at 5:41 PM UTC. Product prices and availability are accurate as of that date/time and are subject to change; the price and availability shown on Amazon at the time of purchase will apply to the purchase of this product.
As an eBay Partner, we may be compensated if you make a purchase.
#598 Kinera Celest Yaksha: $139.00 Amazon $139.00 eBay
Pros: A four driver hybrid with two dynamic drivers, a Knowles balanced armature and a micro planar. Measurements show full sub-bass to 20 Hz (score 69 across ten measurements).
Cons: Treble sits about 3.0 dB below target, a dark, relaxed balance.
Is the Kinera Celest Yaksha right for you?
Buy it if
- you are sensitive to bright or sharp treble (treble sits 3.0 dB below target).
- you want sub-bass rumble that reaches down to 20 Hz.
- you are happy to use EQ: our preset lifts its Harman match by about 19 points.
Skip it if
- you want the best value: the higher-rated EPZ G10 costs about $115 less.
| Back | Closed |
|---|---|
| Connection | Wired |
| Deviation From Target | 2.18 dB standard deviation |
| Driver Type | Hybrid |
| Drivers | 2 x dynamic + 1 x balanced armature + 1 x micro planar |
| Form Factor | In-ear |
| Harman Preference Score | 69 (median of 10 squig.link measurements, calibrated to AutoEQ) |
| Sub-bass Extension | Within 6 dB of target down to 20 Hz (+0.0 dB at 30 Hz) |
| Tonal Tilt | -0.71 (negative is warmer, positive brighter) |
EQ Presets
Harman target
Set a preamp of -5.2 dB, then add these filters in any parametric equalizer (Wavelet, Equalizer APO, Peace, or a player with parametric EQ). Calculated by Speaker Ranking from the median of 10 squig.link measurements, in the filter layout AutoEQ uses. Predicted Harman match with this EQ: 88 out of 100, up from 69.
| Filter | Frequency | Gain | Q |
|---|---|---|---|
| Low shelf | 105 Hz | -0.5 dB | 0.70 |
| High shelf | 10,000 Hz | +1.2 dB | 0.70 |
| Peak | 174 Hz | -3.1 dB | 0.81 |
| Peak | 4,905 Hz | +5.1 dB | 1.60 |
Neutral target
Set a preamp of -5.2 dB, then add these filters in any parametric equalizer (Wavelet, Equalizer APO, Peace, or a player with parametric EQ). Calculated by Speaker Ranking from the median of 10 squig.link measurements, in the filter layout AutoEQ uses. Predicted Neutral match with this EQ: 89 out of 100, up from 33.
| Filter | Frequency | Gain | Q |
|---|---|---|---|
| Low shelf | 105 Hz | -9.0 dB | 0.70 |
| High shelf | 10,000 Hz | +1.2 dB | 0.70 |
| Peak | 174 Hz | -3.0 dB | 0.68 |
| Peak | 4,905 Hz | +5.1 dB | 1.59 |
Studio flat target
Set a preamp of -4.1 dB, then add these filters in any parametric equalizer (Wavelet, Equalizer APO, Peace, or a player with parametric EQ). Calculated by Speaker Ranking from the median of 10 squig.link measurements, in the filter layout AutoEQ uses. Predicted Studio flat match with this EQ: 88 out of 100, up from 14.
| Filter | Frequency | Gain | Q |
|---|---|---|---|
| Low shelf | 105 Hz | -11.6 dB | 0.70 |
| High shelf | 10,000 Hz | +2.0 dB | 0.70 |
| Peak | 172 Hz | -4.8 dB | 0.72 |
| Peak | 4,138 Hz | +4.0 dB | 0.92 |
Warm target
Set a preamp of -3.2 dB, then add these filters in any parametric equalizer (Wavelet, Equalizer APO, Peace, or a player with parametric EQ). Calculated by Speaker Ranking from the median of 10 squig.link measurements, in the filter layout AutoEQ uses. Predicted Warm match with this EQ: 89 out of 100, up from 80.
| Filter | Frequency | Gain | Q |
|---|---|---|---|
| High shelf | 10,000 Hz | -1.2 dB | 0.70 |
| Peak | 174 Hz | -2.9 dB | 0.79 |
| Peak | 4,905 Hz | +3.0 dB | 2.09 |
| Peak | 6,062 Hz | +1.8 dB | 6.00 |
| Peak | 7,872 Hz | -2.4 dB | 6.00 |
Bass target
Set a preamp of -5.2 dB, then add these filters in any parametric equalizer (Wavelet, Equalizer APO, Peace, or a player with parametric EQ). Calculated by Speaker Ranking from the median of 10 squig.link measurements, in the filter layout AutoEQ uses. Predicted Bass match with this EQ: 89 out of 100, up from 74.
| Filter | Frequency | Gain | Q |
|---|---|---|---|
| Low shelf | 105 Hz | +3.3 dB | 0.70 |
| High shelf | 10,000 Hz | +1.2 dB | 0.70 |
| Peak | 174 Hz | -3.0 dB | 0.85 |
| Peak | 4,905 Hz | +5.1 dB | 1.60 |
Bright target
Set a preamp of -6.4 dB, then add these filters in any parametric equalizer (Wavelet, Equalizer APO, Peace, or a player with parametric EQ). Calculated by Speaker Ranking from the median of 10 squig.link measurements, in the filter layout AutoEQ uses. Predicted Bright match with this EQ: 84 out of 100, up from 21.
| Filter | Frequency | Gain | Q |
|---|---|---|---|
| Low shelf | 105 Hz | -9.1 dB | 0.70 |
| High shelf | 10,000 Hz | +2.0 dB | 0.70 |
| Peak | 174 Hz | -3.0 dB | 0.68 |
| Peak | 684 Hz | +1.1 dB | 1.81 |
| Peak | 1,511 Hz | -1.3 dB | 4.00 |
| Peak | 4,604 Hz | +6.0 dB | 1.45 |
| Peak | 5,752 Hz | +1.8 dB | 6.00 |
V-shaped target
Set a preamp of -6.2 dB, then add these filters in any parametric equalizer (Wavelet, Equalizer APO, Peace, or a player with parametric EQ). Calculated by Speaker Ranking from the median of 10 squig.link measurements, in the filter layout AutoEQ uses. Predicted V-shaped match with this EQ: 75 out of 100, up from 63.
| Filter | Frequency | Gain | Q |
|---|---|---|---|
| Low shelf | 105 Hz | +4.0 dB | 0.70 |
| High shelf | 10,000 Hz | +2.0 dB | 0.70 |
| Peak | 4,604 Hz | +6.0 dB | 1.29 |
- Reviews, Comparisons and Sound Samples
Frequency Response
How the Kinera Celest Yaksha measures across the audible range, against the Harman in-ear target that most listeners prefer in blind tests. Where the line sits above the target that range sounds louder (more bass, brighter treble), below it quieter. The curve is the median of 10 independent measurements, so one odd sample or rig cannot skew it, and both lines are shifted so their 500 Hz to 2 kHz average sits at 0 dB.
Measurement data: median of 10 measurements from the squig.link databases of arn, atechreviews, audioamigo, dl, gadgetgenerations, hbb, jaytiss, pw, soundcheck39 and tonedeafmonk. Thanks to each of them for publishing their work.