Article: KEF LS50 (David) Versus JBL 4722 Cinema (Goliath) Speaker Comparison with Binaural Recordings

[Kvalsvoll]

A well done experiment where you have actually done some effort to make it possible to listen for differences in the speakers alone, and you provide recorded samples so that we all can hear at least some of it.

I am not that much into brands, especially other people's brands, so I will use the names big and small for the speakers.

Listening to the samples in Room2 I find:

- Big is dry with quite flat soundstage, small is livelier with more 3D and depth.
- Big is much better in the low-mid and upper-bass, where small sounds like something is missing, transients and attack lacks power, big simply has more realism and power.
- Tonal balance is different.
- Minor differences in high freqs, where small sounds a little more colored.
- The original sounds much better - clarity, bigger more precise soundstage with real 3d, better highs. But this is expected, and explained in your article. The big speaker comes quite close to preserve balance and realism on bass and lower mid.
- Big speakers sound more like the original, but that does not mean is sounds better, becaus this is a recording that now includes the room. When I play back this on my system, all the room and coloration from my system adds to the sound.

So why do they sound so different, when they measure equal. Thing is, they do not measure equal.

If you look at the recorded samples, there is visual difference in the waveforms, wich is caused by differences in phase shift. Frequency spectrum reveals that there are quite large differences below 500hz. At 55hz the small has a huge dip. Differences are in the order 10dB, which is expected to be clearly audible.

So even though the speakers where calibrated with room correction to the same target and measured similar, they actually are different, and this can be measured.

Differences can be caused by radiation charactertistics, which alter not only decay and room contribution, but also frequency response through interaction with boundaries. The DRC did not fix all of that.

The small speaker is most likely overdriven and compress in the lower mid - upper bass. If your calibrated playback is 83dB for -20dBFS each speaker, this is about the same as 0dB on my systems, and that is quite loud, louder than most trad hifi-systems can do. Lowering the master by 10dB will give a sound level closer to what many people listen to, and less chance of overload. I suspect this will still be bordeline for this small speaker, at least you should cross over at 120hz.

To try to play at 0dB with this small speaker is not fair, they do not have the capacity needed, but they may still be good speakers and for someone who never play that loud, it would be interesting to know if there are other aspects of the sound that is compromised. Or can the small speaker sound as good as a big. This is what makes your experiment interesting.

Reducing the master volume by 10dB is one thing to try. Then both system should be properly calibrated to the bass system, crossed over perhaps around 120hz, with proper eq on the bass system and delay settings for the main speakers. That will isolate the differences more to the speakers alone - big vs small.


Regards,
Øyvind Kvålsvoll

[Kvalsvoll]

@mitchco, I should have specified exactly which of your recorded sample files I used for comparing. I am sure I used the speaker+sub files for both the large and the small speakers:

JBL 4722 plus subs

Kef LS50 plus subs

 

Those recordings are not from 2 systems that measure equal, the differences in those samples show that there are significant differences both in frequency domain and time domain. I attached 2 pictures of the frequency spectrum, one smoothed and one raw. To see for yourself you can open the samples in audacity, look at the waveforms (phase differences), analyse frequency spectrum (just make sure you select exactly the same time interval from each sample).

 

If the 2 systems measure equal, they would not present differences in frequency spectrum in such magnitudes. As to why they actually are different, when you measured them to be similar, that is another question, several possibilities exist. And finding what causes the differences in response has significance for the validity of this test, because it is not necessarily caused by the speakers alone.

[Kvalsvoll]

3 hours ago, mitchco said:

...

The point of the article is to show that two speakers eq'd "similarly not exact" sound remarkably close, yet the big difference being how much room sound is let into the recording by the wide directivity differences between these two specific speakers that represent the near opposite ends of the directivity index scale. The intent is to have folks listen to binaural recordings to hear the audible difference and determine what one's preference is with respect to how much room sound one likes mixed in with the direct sound. It is nothing more than that.

 

Enjoy the music!

 

And I believe you did a good job at that, the recordings show off those differences very well.

 

My findings led me to suspect that some kind of minor configuration error during the recording session could cause the responses to be more different than your measurements indicated. Looking at the plots once more, and from reading your comments, I conclude the differences in music sample spectrums are within what can be expected.

 

Would they sound more equal if setup was more accurate? Perhaps, but that would require more effort, which may not be worth it, considering the experiment actually show what you wanted as it is.