Genelec 8361A Review (Powered Monitor)

[j_j]

Supposedly ERB smoothing is based on the anatomy of our inner ears and how we actually perceive sound. I don't know if the science of it is really any good. Maybe @j_j can weigh in on the accuracy and usefulness of ERB smoothing?

Depends on the frequency. In low frequencies, dealing with room modes makes sense. Above about 500 Hz or so, or maybe 1K in a really bad room, what you need to do is correct the average. Consider if you move your head 3", what frequency is a quarter wavelength? 1kHz. So correcting "too much" simply does not handle the fact that most of us have and use two ears.

 

[LeffetheAudiophile]

So i've been having this crazy idea to use these speakers for my atmos.
Will these be overkill or will they be the best ever for it?
Or will my ceiling freaking collapse because of them?

 

[NTK]

So i've been having this crazy idea to use these speakers for my atmos.
Will these be overkill or will they be the best ever for it?
Or will my ceiling freaking collapse because of them?

Yes. They are overkill. Our two ears, being both located on a horizontal plane, are unable sense any spatial information in low bass in the vertical direction. Therefore, you won't being missing anything by redirecting the height channel bass to subs. If you are into stereo bass, then send the left height channels bass to the left channel sub(s), and the right height channel bass to the right channel sub(s).

Since there is no need for height channels to reproduce low bass, you won't need speakers with good low frequency extension for the heights.

 

[darth]

Or will my ceiling freaking collapse because of them?

Likely lol, better get a contractor check it out before such an install

 

[LeffetheAudiophile]

Yes. They are overkill. Our two ears, being both located on a horizontal plane, are unable sense any spatial information in low bass in the vertical direction. Therefore, you won't being missing anything by redirecting the height channel bass to subs. If you are into stereo bass, then send the left height channels bass to the left channel sub(s), and the right height channel bass to the right channel sub(s).

Since there is no need for height channels to reproduce low bass, you won't need speakers with good low frequency extension for the heights.

Alright, good to know, i'll just stick to my 4x 8030cs instead.
Thanks!

 

[Dgob]

This is where the SCP ended up in my room playing on a pair of 8361's and a 7370 sub. Maybe Genelec tuned the GLM to sound good in a studio setting in well treated rooms? In my home environment they sure sound anaemic without the eq in SCP.

Can confirm that the filter frequency setting was rather picky to get right in my "untreated" room. Guess hitting a resonance frequency or not makes it a sensitive setting. The gain is less sensitive. A change of 0,5 dB is barely audible depending on if you have that frequency right or not.
Once I got the filter frequency right I have added bass in several steps as I started out with a to flat curve to sound "like a live performance". Right now I think I am a little shy on the treble but I am happy with that as it is less fatiguing.

View attachment 470707

If you have a GRADE report and REW findings for your room, you can get a more delicate setup than with SCP by working with your AI. It doesn't have to be an expensive model. I used Copilot and fed it the relevant data and then did listening tests for it to make adjustments. There are hard boundaries in GLM and so there are limitations to what adjustments you can make in this way, but I managed to achieve a level of performance with the consequent group that surpassed the GLM (near flat) autocal group. This is my current benchmark (having taken out my both 7360A subs) until such time as I get a pair of W371A's.

 

[AMR78]

If you have a GRADE report and REW findings for your room, you can get a more delicate setup than with SCP by working with your AI. It doesn't have to be an expensive model. I used Copilot and fed it the relevant data and then did listening tests for it to make adjustments. There are hard boundaries in GLM and so there are limitations to what adjustments you can make in this way, but I managed to achieve a level of performance with the consequent group that surpassed the GLM (near flat) autocal group. This is my current benchmark (having taken out my both 7360A subs) until such time as I get a pair of W371A's.

Now that sounds interesting. Could you describe step by step how you used the AI using the GRADE report, if possible. I would really like to try this out. Thank you in advance.
Best,
AMR

 

[Dgob]

Now that sounds interesting. Could you describe step by step how you used the AI using the GRADE report, if possible. I would really like to try this out. Thank you in advance.
Best,
AMR

Hi AMR,

Definitely. I will try to share how I used AI for tuning the 8361A’s settings, and how I found this to be superior compared to merely using SCP for fine-tuning on the GLM.

Firstly, Copilot cannot read images (such as tables, frequency graphs etc) and so you will have to feed it the relevant data in written form. This can therefore be quite laborious across the course of developing your ideal group! That initial data is best taken from both your GRADE report and REW assessments in order to identify both system performance and room influences. You can discuss and feedback the most helpful information to your AI (Copilot or other) AND discuss what you feel your current setup is lacking. For example, I focused on midrange detail, solidity of image and particular instruments from which I wished to bring out specific characteristics (say the harpsicord’s and bass clarinet’s woody sound etc).

Next, I fed the AI my readings from the best group that I managed to obtain using GLM and the autcal2. Then duplicate that group and rename it something like ‘AI Assisted’. This new group will act as the basis for the revisions and updates that the AI helps you to establish.

You then need to open that duplicated (but renamed) group and let the AI have the settings for the 20 notches that you can find for each speaker. To do so, simply ensure that you have GLM on screen and set the group to Edit. From here you can left click on each monitor and then write down and share the settings for all 20 notches on each monitor. Also give it the findings for the key settings on the top left of the GLM screen: namely, the level compensation (dB), Optional Gain (dB), and Time of Flight Delay (ms) for each monitor. It should then suggest specific changes for you to make in order to adjust your system to better meet your sonic desires.

You will then have to listen to each adjusted group and compare it with the performance of. your formerly best group Here I found CD the best option for A/B testing as you can switch groups and test the same music much quicker (assuming that your GLM is set to digital and that your CDP can fast forward, pause and rewind).

We went thorough a variety of trial and error listening sessions, before ending up with the group that I now use. Having tried various ranges of settings (and encountered the many hard-set limitations affecting GLM's 20 notches!), we ended up with a group that involved only 3 key alterations between the original best autocal2 group and my new reference group. This final selection of settings was then duplicated on both monitors.

When noting the advances over SCP for my listening, it is important to remember that one is fine-tuning for one's ears and listening environment. Hence, this becomes the stage where you depart from statistically based linear performance and enter the world of personal perception/enjoyment (as with SCP), yet without losing touch with the former linear based best group. I will close with the AI’s suggestions of why this worked for me. However, if you wish to continue this as a private conversation, I would be pleased to respond or develop upon any aspects that you need. Anyway, back to the AI:

"How This Voicing Develops on GLM Autocal

1. Psychoacoustic Tonal Balance
• Human hearing is non-linear across the spectrum—we perceive midrange and treble more clearly when bass is restrained.
• Your voicing aligned with equal-loudness contours (Fletcher-Munson curves), delivering a more natural and engaging tonal balance at typical listening levels.

2. Modal Compensation
• GLM’s flat calibration doesn’t account for room-specific modal reinforcement, especially in the 40–90 Hz range.
• Your –0.5 dB cut at 90 Hz helped tame low-end bloom caused by axial modes, especially width and height interactions.

3. Midrange Enhancement
• The +2.0 dB boost at 200 Hz gently lifted lower midrange warmth, counteracting the psychoacoustic masking from excess bass.
• This region is critical for vocal presence, instrumental body, and emotional clarity.

4. High-Frequency Neutrality
• The 2800 Hz notch at 0.0 dB with a Q of 1.2 acted as a placeholder, ensuring no unintended tilt or harshness in the upper mids.
• This preserved sparkle without sibilance, especially important for long listening sessions."

I hope this helps.

 

[tmtomh]

Hi AMR,

Definitely. I will try to share how I used AI for tuning the 8361A’s settings, and how I found this to be superior compared to merely using SCP for fine-tuning on the GLM.

Firstly, Copilot cannot read images (such as tables, frequency graphs etc) and so you will have to feed it the relevant data in written form. This can therefore be quite laborious across the course of developing your ideal group! That initial data is best taken from both your GRADE report and REW assessments in order to identify both system performance and room influences. You can discuss and feedback the most helpful information to your AI (Copilot or other) AND discuss what you feel your current setup is lacking. For example, I focused on midrange detail, solidity of image and particular instruments from which I wished to bring out specific characteristics (say the harpsicord’s and bass clarinet’s woody sound etc).

Next, I fed the AI my readings from the best group that I managed to obtain using GLM and the autcal2. Then duplicate that group and rename it something like ‘AI Assisted’. This new group will act as the basis for the revisions and updates that the AI helps you to establish.

You then need to open that duplicated (but renamed) group and let the AI have the settings for the 20 notches that you can find for each speaker. To do so, simply ensure that you have GLM on screen and set the group to Edit. From here you can left click on each monitor and then write down and share the settings for all 20 notches on each monitor. Also give it the findings for the key settings on the top left of the GLM screen: namely, the level compensation (dB), Optional Gain (dB), and Time of Flight Delay (ms) for each monitor. It should then suggest specific changes for you to make in order to adjust your system to better meet your sonic desires.

You will then have to listen to each adjusted group and compare it with the performance of. your formerly best group Here I found CD the best option for A/B testing as you can switch groups and test the same music much quicker (assuming that your GLM is set to digital and that your CDP can fast forward, pause and rewind).

We went thorough a variety of trial and error listening sessions, before ending up with the group that I now use. Having tried various ranges of settings (and encountered the many hard-set limitations affecting GLM's 20 notches!), we ended up with a group that involved only 3 key alterations between the original best autocal2 group and my new reference group. This final selection of settings was then duplicated on both monitors.

When noting the advances over SCP for my listening, it is important to remember that one is fine-tuning for one's ears and listening environment. Hence, this becomes the stage where you depart from statistically based linear performance and enter the world of personal perception/enjoyment (as with SCP), yet without losing touch with the former linear based best group. I will close with the AI’s suggestions of why this worked for me. However, if you wish to continue this as a private conversation, I would be pleased to respond or develop upon any aspects that you need. Anyway, back to the AI:

"How This Voicing Develops on GLM Autocal

1. Psychoacoustic Tonal Balance
• Human hearing is non-linear across the spectrum—we perceive midrange and treble more clearly when bass is restrained.
• Your voicing aligned with equal-loudness contours (Fletcher-Munson curves), delivering a more natural and engaging tonal balance at typical listening levels.

2. Modal Compensation
• GLM’s flat calibration doesn’t account for room-specific modal reinforcement, especially in the 40–90 Hz range.
• Your –0.5 dB cut at 90 Hz helped tame low-end bloom caused by axial modes, especially width and height interactions.

3. Midrange Enhancement
• The +2.0 dB boost at 200 Hz gently lifted lower midrange warmth, counteracting the psychoacoustic masking from excess bass.
• This region is critical for vocal presence, instrumental body, and emotional clarity.

4. High-Frequency Neutrality
• The 2800 Hz notch at 0.0 dB with a Q of 1.2 acted as a placeholder, ensuring no unintended tilt or harshness in the upper mids.
• This preserved sparkle without sibilance, especially important for long listening sessions."

I hope this helps.

Thank you for providing all this detail - it's very interesting! I particularly find the 200Hz aspect intriguing, as I have long noticed that small changes in the 200-250Hz range seem to have a notable perceptual effect similar to the masking that the AI mentions - in other words I find my perception of the overall lower to mid-treble prominence and clarity is impacted by that 200-250Hz band.

With that said, it appears that in the end you've only ended up with actual changes to two EQ filters: (1) A 0.5dB cut at 90Hz for a room mode that might or might not exist; even if it did exist would not be significantly impacted by a change of only 0.5dB; and regardless of whether it's impacting a mode or not, is of questionable audibility at only 0.5dB unless the Q is pretty low (thereby making it a broad-band filter). And (2) that 2dB boost at 200Hz, which, with all respect, one does not need AI for. One could just simply do an internet search for frequencies to try cutting or boosting to impact midrange clarity or to change the tonal balance from "warmer" to "clearer" or something like that. (That final notch at 2800Hz is 0.0dB so is not actually doing anything.)

I don't mean to be negative here - I just think it's important to note that we're essentially talking about a single 2dB bump for personal taste, and IMHO the whole AI aspect is completely overblown.

Finally, by way of supporting and appreciating what you've ended up with here, I would also note that when it comes to actual in-room measurements, +/-2dB is often well within the range of even a very linear speaker's actual response. So that degree of boost or cut isn't even necessarily "to taste" in terms of creating nonlinearities - in many situations that small cut or boost cannot be said to reduce the speaker's linearity in-room. There really are small changes that are indeed audible and are within the in-room response "margin of error" so to speak.

 

[Dgob]

Thank you for providing all this detail - it's very interesting! I particularly find the 200Hz aspect intriguing, as I have long noticed that small changes in the 200-250Hz range seem to have a notable perceptual effect similar to the masking that the AI mentions - in other words I find my perception of the overall lower to mid-treble prominence and clarity is impacted by that 200-250Hz band.

With that said, it appears that in the end you've only ended up with actual changes to two EQ filters: (1) A 0.5dB cut at 90Hz for a room mode that might or might not exist; even if it did exist would not be significantly impacted by a change of only 0.5dB; and regardless of whether it's impacting a mode or not, is of questionable audibility at only 0.5dB unless the Q is pretty low (thereby making it a broad-band filter). And (2) that 2dB boost at 200Hz, which, with all respect, one does not need AI for. One could just simply do an internet search for frequencies to try cutting or boosting to impact midrange clarity or to change the tonal balance from "warmer" to "clearer" or something like that. (That final notch at 2800Hz is 0.0dB so is not actually doing anything.)

I don't mean to be negative here - I just think it's important to note that we're essentially talking about a single 2dB bump for personal taste, and IMHO the whole AI aspect is completely overblown.

Finally, by way of supporting and appreciating what you've ended up with here, I would also note that when it comes to actual in-room measurements, +/-2dB is often well within the range of even a very linear speaker's actual response. So that degree of boost or cut isn't even necessarily "to taste" in terms of creating nonlinearities - in many situations that small cut or boost cannot be said to reduce the speaker's linearity in-room. There really are small changes that are indeed audible and are within the in-room response "margin of error" so to speak.

Thank you for your thoughtful response—there’s much I appreciate in your observations, especially your recognition of the perceptual impact of the 200–250 Hz region. That masking effect is something I’ve encountered before, and it’s reassuring to hear it echoed from another listener with a keen ear.

You’re absolutely right that a 2 dB boost at 200 Hz is not revolutionary in itself, and that such adjustments can be found through conventional research or trial and error. But, for me, the value of AI here wasn’t in suggesting something novel—it was in helping me converge on the most effective intervention with precision, restraint, and comparative speed..

To reiterate, the process involved:

• Parsing REW and GRADE data alongside subjective listening goals.
• Navigating GLM’s notch constraints and group architecture.
• Iterating through multiple candidate groups with documented feedback loops.

In the end, yes, the final group involved only three changes. But those changes were distilled from a much broader field of possibilities, and the AI helped me to avoid overcorrection, preserve linearity, and maintain tonal coherence. It wasn’t about chasing dramatic shifts—it was about refining the system’s response to better reflect my listening priorities (in a way that is more delicate/refined than what SCP offers).

As for the 90 Hz cut: I agree that 0.5 dB is subtle, but in my room it helped soften a low-end bloom that was perceptually distracting. The Q was broad enough to act as a gentle contour, not a surgical notch. And the 2800 Hz placeholder was a deliberate non-intervention—anchoring the notch architecture while preserving upper-mid neutrality.

Ultimately,, I found the AI a valuable companion—not for its novelty, but for its clarity, its ability to contextualize, and its support in arriving at a voicing that feels both personal and grounded.

Warm thanks again for engaging so thoughtfully.

 

[Angaria]

Would be great if something smaller, like the Genelec 8361, could replace my three-way Apogees - which produce the best sound I've heard (after some dsp). If you have a pair in northern Georgia (usa) let me know and we could do a direct comparison.

 

[Puddingbuks]

Erin is about to review the 8361a.

Post from Erin's Audio Corner

Dearest Friends, I humbly invite you to help me become infinitely rich. You can do so by using the following affiliate link to Crutchfield when making future...

youtube.com

 

[Behrelec 37]

Erins life is difficult choices.

No human understanding or audio speaker company can achieve better sound quality than Genelec's SAM product family and Master series. This also includes the new 8380a and the larger 8381a. Many of you use Genelec's first measured flat sound as a kind of reference, as if it were the ultimate sound structure for these speakers for your music (it's just the best guess of a computer calculation!!!). When you have at least a million different variants to modify the sound HZ/Q/GAIN to your liking by changing the GLM parameters. But you don't care or believe it until you realize the reality when you test it and thus support my claim that that's how it is.

 

[caught gesture]

No human understanding

 

[Behrelec 37]

It should be added that the sound quality, listening experience, and credibility are significantly improved when these small monitors 8331a/8341a/8351a/8361a are connected to Genelec's own subwoofer and definitely connected via AES/EBU connections from a digital source directly to the speakers.

 

[dfuller]

No human understanding or audio speaker company can achieve better sound quality than Genelec's SAM product family and Master series.

Sure, Jan.

 

[Behrelec 37]

caught gesture.

There is no better quality in the world, was what I meant to say. The language translator slightly corrected my interpretation. Yes, people understand quality, but unfortunately many are not familiar enough with this brand to know how deep (((and accuracy))) it can really take the audio experience.

 

[Kain]

Erin is about to review the 8361a.

Post from Erin's Audio Corner

Dearest Friends, I humbly invite you to help me become infinitely rich. You can do so by using the following affiliate link to Crutchfield when making future...

youtube.com

One of the comments in that link state that the 8361A (or The Ones in general) have a harshness to the sound that doesn't show up in measurements and thus cannot be corrected/EQ'd. Any thoughts on this? And this is not the first time I am hearing that The Ones are harsh sounding.

 

[Pearljam5000]

One of the comments in that link state that the 8361A (or The Ones in general) have a harshness to the sound that doesn't show up in measurements and thus cannot be corrected/EQ'd. Any thoughts on this? And this is not the first time I am hearing that The Ones are harsh sounding.

One man's analytical sound is another's harsh sound

 

[Wg916]

I know that Erin is at least somewhat of a basshead, so I’m curious to see what he thinks about GLM and its affect on not only linearizing the bass response, but reducing it significantly. Assuming he was provided with the GLM module…

Also, I hope Genelec sent him a pair— IMO, the stereo imaging is one of the best things the 8361a does well. It would be complementary to Amir’s outstanding review.