JBL LSR305 (1st gen) quasi-anechoic measurements

[dominikz]

Today I finally measured one of my 1st generation JBL LSR305s, using the quasi-anechoic method, and used it to create a spinorama-style report. It is far from the precision of an NFS, but I hope it might still provide some qualitative insight on these pretty common loudspeakers.

Some information on measurement procedure first:

• Made with the quasi-anechoic method, in principle equivalent to the one from the amazing guide by @napilopez
• Measurements were done in an apartment using a makeshift turntable and stand
• EDIT 2021-06-28: Measurements were done with Cross-Spectrum labs calibrated Dayton Audio EMM-6 microphone, pointed horizontally (towards the loudspeaker) and 0° calibration file was applied
• Nearest reflective surface was the ceiling, ~1,2m away from center of the woofer
• Measured at 0,5m distance so I could use ~5,2ms gate for the HF part of the measurement (IMHO this should satisfy far-field criteria as at it is ~4,4 times the radiating diameter of the woofer)
• SPL in REW was calibrated by a cheap SPL meter - measurement level was ~83-85dB SPL at 0,5m
• For the horizontals measurement axis was the tweeter
• For the verticals I had to move the measurement axis to the LED below the tweeter, as that was the only way to safely balance the loudspeaker on the stand - however the 0°/on-axis measurement for both the tweeter and LED axis seems very close, so I hope the total error is not very large
• Used 10° increments for the measurements, and covered one half-plane with the horizontals (0-180°, should be OK as the loudspeaker is symmetrical), and the full 360° for the verticals
• The nearfield LF measurements were spliced with the gated HF measurements at ~250Hz. Note that due to splicing there is some uncertainty in how the low frequencies are depicted (should be close, though - see comparisons with other measurements below)

Spinorama:

I tried comparing it with a spin made by the vendor (JBL), by digitizing the very low res image available online:

It appears there some qualitative overlap, especially in the DIs - most of the differences seem to be in the 1-2k range, and these could be due to unit-to-unit variation, imprecise digitization of data, low resolution of the original image, or of course - due to errors in my measurement rig or process.

Horizontal directivity:

Vertical directivity:

We can see above I was slightly below the tweeter axis - as mentioned in the introduction.

I was also interested how the 1st gen compares with the newer 305p MKII, here's a listening window comparison from spins by @amirm and @hardisj:

There is some similarity, though it looks like the newer model might be brighter (but it also has a less pronounced 12k dip and slightly better DI).

Lastly I compared just the on-axis measurement of the two units I own to see unit-to-unit variation. The differences mainly show above ~1kHz:

There is a slightly bigger difference around the crossover, but above 2,5k both units seem within ~1,5dB of each other.
NOTE: Originally one of the units I own had a ~2dB tweeter level difference vs the other, and I had to replace the tweeter on one of them to get the slightly closer match that you see above.

In case @MZKM and/or @pierre (or others) might be interested to do their magic, CTA-2034-A export from VCAD is attached.

EDIT [2021-07-01]: Added distortion and individual driver nearfield measurements and a few misc. tests in post #15.

 

[dominikz]

Maybe also interesting - here's how the spin compares to the nearfield (~0,8m) MMM in-room measurement:

As we can see, due to measuring in the nearfield direct sound dominates and MMM response is very close to LW above ~700Hz.
You can also see how my acoustically terrible room and poor placement choice performs in the LF Thank goodness for EQ.

 

[sweetchaos]

Nicely done. Thanks for the effort and time.
Any other speakers you can measure?

I feel like these posts should be in their own sub-forum…maybe. Perhaps separate the measurement posts from discussion posts. Just a thought.

 

[ernestcarl]

Maybe also interesting - here's how the spin compares to the nearfield (~0,8m) MMM in-room measurement:
View attachment 137572
As we can see, due to measuring in the nearfield direct sound dominates and MMM response is very close to LW above ~700Hz.
You can also see how my acoustically terrible room and poor placement choice performs in the LF Thank goodness for EQ.

This doesn't look all that terrible to me. That ~8 dB dip at 68 Hz could be significantly improved with the help of a sub(s).

 

[dominikz]

Any other speakers you can measure?

Nothing new or exciting at the moment, I'm afraid My main pair of speakers are Revel M16, which were already measured here; and we also have a pair of long-discontinued Genius SP-HF3000A multimedia loudspeakers.
But who knows, I do hope to do more of these measurements
Thanks for the kind words!

This doesn't look all that terrible to me. That ~8 dB dip at 68 Hz could be significantly improved with the help of a sub(s).

Indeed - though at the moment I'm not looking to add a sub to this system so I'm just using EQ to mitigate

 

[changer]

Hi Dominik,
good effort! What effective radiating diameter did you assume for the woofer? I am asking because in Arta AN4 the far field threshold was said to be at 6 times the radius of the woofer as a rule of thumb or there is also a formula https://www.artalabs.hr/AppNotes/AN4-FreeField-Rev03eng.pdf
Don't know what's the magnitude of the error, but 50 cm could be a little short?

 

[napilopez]

Great stuff! What mic did you use?

 

[dominikz]

Hi Dominik,
good effort! What effective radiating diameter did you assume for the woofer? I am asking because in Arta AN4 the far field threshold was said to be at 6 times the radius of the woofer as a rule of thumb or there is also a formula https://www.artalabs.hr/AppNotes/AN4-FreeField-Rev03eng.pdf
Don't know what's the magnitude of the error, but 50 cm could be a little short?

Thanks a lot!

Please note that diameter of a circle is 2x its radius - so >6 times the radius of the woofer is actually >3 times its diameter. This condition was satisfied in my measurement.

Estimated effective radiating diameter of the LSR305 woofer is ~11,3cm (radius ~5,65cm) which means that with 50cm measuring distance the factor of distance / effective radiating diameter is ~4,4 (~8,8 compared to the radius) therefore well over 3 (well over 6 compared to the radius).
A very useful whitepaper by the late Jeff Bagby also suggests that measuring at a distance 3-5 times the effective radiating diameter of a loudspeaker is enough to satisfy far field requirements (it is therefore in line with the ARTA paper you posted as well).

In addition, before starting the spin measurement, I measured the gated on-axis response at 30cm, 50cm and 80cm distance to make sure that 50cm was OK - in practice the gated responses at 80cm and 50cm were almost identical (except that 50cm allowed longer windows so had better resolution in the lower frequencies).

My choice of 50cm as the measuring distance was mainly motivated by easier extrapolation of measured SPL to 1 and 2m distance.

Hope this clarifies a bit!

Great stuff! What mic did you use?

Thanks! I use a Cross-Spectrum labs calibrated Dayton Audio EMM-6, it was pointed horizontally (towards the loudspeaker) and 0° calibration file was applied - forgot to mention it in the OP! (Edit: added now!)

 

[napilopez]

Thanks a lot!

Please note that diameter of a circle is 2x its radius - so >6 times the radius of the woofer is actually >3 times its diameter. This condition was satisfied in my measurement.

Estimated effective radiating diameter of the LSR305 woofer is ~11,3cm (radius ~5,65cm) which means that with 50cm measuring distance the factor of distance / effective radiating diameter is ~4,4 (~8,8 compared to the radius) therefore well over 3 (well over 6 compared to the radius).
A very useful whitepaper by the late Jeff Bagby also suggests that measuring at a distance 3-5 times the effective radiating diameter of a loudspeaker is enough to satisfy far field requirements (it is therefore in line with the ARTA paper you posted as well).

In addition, before starting the spin measurement, I measured the gated on-axis response at 30cm, 50cm and 80cm distance to make sure that 50cm was OK - in practice the gated responses at 80cm and 50cm were almost identical (except that 50cm allowed longer windows so had better resolution in the lower frequencies).

My choice of 50cm as the measuring distance was mainly motivated by easier extrapolation of measured SPL to 1 and 2m distance.

Hope this clarifies a bit!


Thanks! I use a Cross-Spectrum labs calibrated Dayton Audio EMM-6, it was pointed horizontally (towards the loudspeaker) and 0° calibration file was applied - forgot to mention it in the OP! (Edit: added now!)

Thanks for clarifying! The drooping treble is a little unusual so I was wondering if it had anything to do with the mic calibration. And I'm assuming the speaker was on its standard EQ settings?

Sound and recording's measurements of the LSR305 don't have that mid forward, downward trend either. Perhaps JBL was just inconsistent with this speaker! Or it changed the tuning along the way. I don't think this would be caused by the measurement distance, espcially as the 0.8m MMM measurement shows the droop as well.

 

[Koeitje]

Nothing new or exciting at the moment, I'm afraid My main pair of speakers are Revel M16, which were already measured here; and we also have a pair of long-discontinued Genius SP-HF3000A multimedia loudspeakers.

I'm wondering if you have the same experience as me. I didn't do an AB comparison of the the JBL's with my M106 directly, but when I tried the JBL's I felt I was hearing a speaker with a sound signature that was pretty close. The M106 is clearly the better speaker, don't get the me wrong, but compared to other loudspeakers the difference in tonality wasn't that big.

Now they are the rears in my surround setup with the M106 as fronts, but even if they were my primary speakers I would be happy with them.

 

[daftcombo]

Very nice measurements. The original 305 measures better imo.

Can you measure THD also?

 

[dominikz]

Thanks for clarifying! The drooping treble is a little unusual so I was wondering if it had anything to do with the mic calibration. And I'm assuming the speaker was on its standard EQ settings?

Yes - both LF and HF switches at the 0dB setting.

Sound and recording's measurements of the LSR305 don't have that mid forward, downward trend either.

That's interesting - the on-axis response from S&R I found is pretty low resolution; but this is how it compares to my measurement and the one from the vendor (JBL):

My previous measurements (MMM or gated) of both of my JBL units show the same approximate shape of the response in HF. Note that on one of them I had to replace the tweeter so below I have a comparison to measurements before replacement - therefore in total 3 different tweeters are compared here:

Green plot shows the old, faulty (2dB less sensitive tweeter). Tweeter unit variations definitely change the shape of the 1-2kHz part of the response significantly (the crossover region), which could affect quite a lot how this area looks in each individual unit's response.

My previous measurements of Revel M16 on the other hand seem to match JBL's and Amir's so I doubt I have a systematic error with the microphone:

Perhaps JBL was just inconsistent with this speaker! Or it changed the tuning along the way. I don't think this would be caused by the measurement distance, espcially as the 0.8m MMM measurement shows the droop as well.

Yeah, definitely strange! From my experience with these JBL monitors there's definitely more production variance vs my Revels. The response of the two M16s matches very closely up until almost 18kHz while with the JBLs each of the three tweeters had a different response. In addition one of the JBL units had a too narrow slot for the backplate and exhibited pretty strong resonance ~250Hz. The other one didn't.

I'm wondering if you have the same experience as me. I didn't do an AB comparison of the the JBL's with my M106 directly, but when I tried the JBL's I felt I was hearing a speaker with a sound signature that was pretty close. The M106 is clearly the better speaker, don't get the me wrong, but compared to other loudspeakers the difference in tonality wasn't that big.

Now they are the rears in my surround setup with the M106 as fronts, but even if they were my primary speakers I would be happy with them.

I fully agree that there is a lot of similarity between the M16 and LSR305 when A/B-ing them - both sound pretty good to me! However, one time as an experiment I set-up one of the LSR305s as the center channel with my M16 being L and R - and in this case IMHO the difference in tonality of the center channel was quite audible and even a bit distracting. I assume this would be less of a problem with surround channels, though!

Very nice measurements. The original 305 measures better imo.

I'd agree with that - though I feel there's a lot of unit-to-unit variance at play here.

Can you measure THD also?

I'll see what I can do

 

[napilopez]

Yes - both LF and HF switches at the 0dB setting.

That's interesting - the on-axis response from S&R I found is pretty low resolution; but this is how it compares to my measurement and the one from the vendor (JBL):
View attachment 137916

My previous measurements (MMM or gated) of both of my JBL units show the same approximate shape of the response in HF. Note that on one of them I had to replace the tweeter so below I have a comparison to measurements before replacement - therefore in total 3 different tweeters are compared here:
View attachment 137917
Green plot shows the old, faulty (2dB less sensitive tweeter). Tweeter unit variations definitely change the shape of the 1-2kHz part of the response significantly (the crossover region), which could affect quite a lot how this area looks in each individual unit's response.

My previous measurements of Revel M16 on the other hand seem to match JBL's and Amir's so I doubt I have a systematic error with the microphone:
View attachment 137913


Yeah, definitely strange! From my experience with these JBL monitors there's definitely more production variance vs my Revels. The response of the two M16s matches very closely up until almost 18kHz while with the JBLs each of the three tweeters had a different response. In addition one of the JBL units had a too narrow slot for the backplate and exhibited pretty strong resonance ~250Hz. The other one didn't.


I fully agree that there is a lot of similarity between the M16 and LSR305 when A/B-ing them - both sound pretty good to me! However, one time as an experiment I set-up one of the LSR305s as the center channel with my M16 being L and R - and in this case IMHO the difference in tonality of the center channel was quite audible and even a bit distracting. I assume this would be less of a problem with surround channels, though!


I'd agree with that - though I feel there's a lot of unit-to-unit variance at play here.

I'll see what I can do

Thank you for the detailed response! Yeah if I had to guess at this point, it'd be some unit-to-unit variation, whether because of production changes or QC; hopefully the former.

Those were the same S&R measurements I was looking at and I'd expect a bit closer of a correlation; in my experience doing quasi-anechoic measurements the region from 2kHz to 10kHz should basically be identical to anechoic results (usually within +/- 1dB aside from the odd dip or peak), like your M16 measurements. So it definitely points to something with the speaker. Goes to show the importance of being able to measure your own speakers too though; if you EQ'd based on S&R's results you might not get a tuning you'd like!

Thanks for sharing!

 

[richard12511]

I'm wondering if you have the same experience as me. I didn't do an AB comparison of the the JBL's with my M106 directly, but when I tried the JBL's I felt I was hearing a speaker with a sound signature that was pretty close. The M106 is clearly the better speaker, don't get the me wrong, but compared to other loudspeakers the difference in tonality wasn't that big.

Now they are the rears in my surround setup with the M106 as fronts, but even if they were my primary speakers I would be happy with them.

I did a blind comparison with the M105 and 308p last year(8030c too). The M105 beat the 308p, but not by a lot.

 

[dominikz]

Can you measure THD also?

OK, managed to do more measurements today.
These THD measurements were done in room at 50cm distance from tweeter, while SPL on the graphs is estimated for 1m distance (i.e. reduced by 6dB due to distance doubling).
I'm only showing 2nd, 3rd and 5th harmonic to make the graphs less busy (as these dominate the THD anyway) and total THD (2-9th harmonic).

I didn't go over 90dB as I didn't want to risk upsetting the neighbours
The distortion peak that becomes prominent after 86dB is visible in other harmonics too, and it is also audible when doing sweeps above ~85-90dB SPL at 1m. It seems to be related to backplate resonance and it is more prominent on one of my speakers vs the other one.

What I found interesting is the rising HF distortion, as it is much more prominent on one speaker than the other. To see what causes it I tried replacing the tweeters between the two units and this is what resulted (playback level: 80 dB SPL at 1kHz at 1m):

As you can see, one unit has more HF distortion regardless which tweeter is in, implying that this added distortion is caused by the amplifier unit rather that the tweeter driver.

Interesting thing is that I get a better match between units with the tweeters switched:

Lastly, let me show individual driver components measured in nearfield (vs on-axis):

Hope some find this interesting!

 

[daftcombo]

OK, managed to do more measurements today.
These THD measurements were done in room at 50cm distance from tweeter, while SPL on the graphs is estimated for 1m distance (i.e. reduced by 6dB due to distance doubling).
I'm only showing 2nd, 3rd and 5th harmonic to make the graphs less busy (as these dominate the THD anyway) and total THD (2-9th harmonic).
View attachment 138499
View attachment 138500
View attachment 138501View attachment 138502View attachment 138504
I didn't go over 90dB as I didn't want to risk upsetting the neighbours
The distortion peak that becomes prominent after 86dB is visible in other harmonics too, and it is also audible when doing sweeps above ~85-90dB SPL at 1m. It seems to be related to backplate resonance and it is more prominent on one of my speakers vs the other one.

What I found interesting is the rising HF distortion, as it is much more prominent on one speaker than the other. To see what causes it I tried replacing the tweeters between the two units and this is what resulted (playback level: 80 dB SPL at 1kHz at 1m):
View attachment 138519
As you can see, one unit has more HF distortion regardless which tweeter is in, implying that this added distortion is caused by the amplifier unit rather that the tweeter driver.

Interesting thing is that I get a better match between units with the tweeters switched:
View attachment 138527

Lastly, let me show individual driver components measured in nearfield (vs on-axis):
View attachment 138530

Hope some find this interesting!

Excellent work, thank you very much!

Really weird one unit has more distortion than the other, even with the tweeters switched.

 

[dominikz]

Excellent work, thank you very much!

Thanks!

Really weird one unit has more distortion than the other, even with the tweeters switched.

I guess unit-to-unit variance would be one of the things where budget shows - given the level of differences I saw with my two units (FR deviations, faulty tweeter, different levels of backplate resonance, different levels of HF amplifier distortion) I wouldn't really dare to do loudspeaker correction on these based on third-party measurements - regardless how precise

However, in practice these sound pretty good IMO - so the above 'issues' are mostly of academic interest to me

 

[changer]

OK, managed to do more measurements today.
These THD measurements were done in room at 50cm distance from tweeter, while SPL on the graphs is estimated for 1m distance (i.e. reduced by 6dB due to distance doubling).

Are these stepped sine measurements, from within RTA window? These differ quite a bit compared to sweeps. And they take more time, if go for a higher steps per octave ratio.

 

[dominikz]

Are these stepped sine measurements, from within RTA window? These differ quite a bit compared to sweeps. And they take more time, if go for a higher steps per octave ratio.

No, these are just sweeps - sorry

 

[DavidMcRoy]

Today I finally measured one of my 1st generation JBL LSR305s, using the quasi-anechoic method, and used it to create a spinorama-style report. It is far from the precision of an NFS, but I hope it might still provide some qualitative insight on these pretty common loudspeakers.

Some information on measurement procedure first:

• Made with the quasi-anechoic method, in principle equivalent to the one from the amazing guide by @napilopez
• Measurements were done in an apartment using a makeshift turntable and stand
• EDIT 2021-06-28: Measurements were done with Cross-Spectrum labs calibrated Dayton Audio EMM-6 microphone, pointed horizontally (towards the loudspeaker) and 0° calibration file was applied
• Nearest reflective surface was the ceiling, ~1,2m away from center of the woofer
• Measured at 0,5m distance so I could use ~5,2ms gate for the HF part of the measurement (IMHO this should satisfy far-field criteria as at it is ~4,4 times the radiating diameter of the woofer)
• SPL in REW was calibrated by a cheap SPL meter - measurement level was ~83-85dB SPL at 0,5m
• For the horizontals measurement axis was the tweeter
• For the verticals I had to move the measurement axis to the LED below the tweeter, as that was the only way to safely balance the loudspeaker on the stand - however the 0°/on-axis measurement for both the tweeter and LED axis seems very close, so I hope the total error is not very large
• Used 10° increments for the measurements, and covered one half-plane with the horizontals (0-180°, should be OK as the loudspeaker is symmetrical), and the full 360° for the verticals
• The nearfield LF measurements were spliced with the gated HF measurements at ~250Hz. Note that due to splicing there is some uncertainty in how the low frequencies are depicted (should be close, though - see comparisons with other measurements below)

Spinorama:
View attachment 137533

I tried comparing it with a spin made by the vendor (JBL), by digitizing the very low res image available online:
View attachment 137557
It appears there some qualitative overlap, especially in the DIs - most of the differences seem to be in the 1-2k range, and these could be due to unit-to-unit variation, imprecise digitization of data, low resolution of the original image, or of course - due to errors in my measurement rig or process.

Horizontal directivity:
View attachment 137534
View attachment 137535

Vertical directivity:
View attachment 137536
We can see above I was slightly below the tweeter axis - as mentioned in the introduction.
View attachment 137556
View attachment 137537

I was also interested how the 1st gen compares with the newer 305p MKII, here's a listening window comparison from spins by @amirm and @hardisj:
View attachment 137555
There is some similarity, though it looks like the newer model might be brighter (but it also has a less pronounced 12k dip and slightly better DI).

Lastly I compared just the on-axis measurement of the two units I own to see unit-to-unit variation. The differences mainly show above ~1kHz:
View attachment 137558
There is a slightly bigger difference around the crossover, but above 2,5k both units seem within ~1,5dB of each other.
NOTE: Originally one of the units I own had a ~2dB tweeter level difference vs the other, and I had to replace the tweeter on one of them to get the slightly closer match that you see above.

In case @MZKM and/or @pierre (or others) might be interested to do their magic, CTA-2034-A export from VCAD is attached.

EDIT [2021-07-01]: Added distortion and individual driver nearfield measurements and a few misc. tests in post #15.

Wow, thanks for this! I’m using 8 of these for surround and height channels in a Dolby Atmos system. Very helpful!