Sony SS-CS5 bookshelf speaker (review by Erin)

[sweetchaos]

Thanks to @hardisj for his review:

Sony SS-CS5 Bookshelf Speaker Review

Sony SS-CS5 Review

www.erinsaudiocorner.com

Erin's conclusion:

Let’s be real. This is a budget darling. I paid $88 for the pair. Most people are paying $130 or less. The performance is OK. It’s not going to knock your socks off with fidelity (nor bass) but it’s a good option for someone starting out.

Discuss!

 

[Doodski]

Erin's conclusion:

Let’s be real. This is a budget darling. I paid $88 for the pair. Most people are paying $130 or less. The performance is OK. It’s not going to knock your socks off with fidelity (nor bass) but it’s a good option for someone starting out.

I imagine shipping a empty shoebox around the earth a couple of times before sending it to the end user would cost more than these speakers are worth.

 

[sweetchaos]

5 reviews in 6 days?

Erin is on a roll...

Don't you sleep, man?!

 

[sweetchaos]

Erin's comparison with Neumi BS5:

 

[Dennis Murphy]

Interesting comparison. I've worked with the Sony quite a bit and have done some cabinet and crossover mods, and I would take issue with Erin's explanation for the peak around 1 kHz. Erin attributes this to internal cabinet resonances and/or a port resonance, but it's really just inadequate baffle step compensation. There is a port resonance, but it's higher up--around 1600 Hz and doesn't show up in the frequency response measurement. Strengthening the cabinet, or adding more sound treatment won't help with the 1 kHz peak, which to my ears is the main issue with the speaker. You either have to build in more baffle step compensation in the low pass circuit, or add "wings" to either side of the cabinet to attenuate the baffle step somewhat. The latter approach is clumsy. The former is relatively easy, but you still won't have a speaker that goes very low or can play very loud. Within its limits, though, a modded version is capable of very decent sound.

 

[hardisj]

Interesting comparison. I've worked with the Sony quite a bit and have done some cabinet and crossover mods, and I would take issue with Erin's explanation for the peak around 1 kHz. Erin attributes this to internal cabinet resonances and/or a port resonance, but it's really just inadequate baffle step compensation.

The cabinet has clear resonance. This is seen easily in the accelerometer tests. Especially at 1kHz. Baffle step can fix the delta, sure. But the cabinet definitely is ringing at 1kHz.

 

[Dennis Murphy]

The cabinet has clear resonance. This is seen easily in the accelerator tests. Especially at 1kHz. Baffle step can fix the delta, sure. But the cabinet definitely is ringing at 1kHz.

View attachment 172571

But that's not the cause of the peak. Otherwise simply adding width to the cabinet wouldn't clear up the issue. I also don't think that ringing would be excited as much at low playback levels, but the peak is always there. Most speakers in this price class don't build in enough baffle step compensation, not only because of the expense but because it lowers rated sensitivity. Another member of the forum--an electrical engineer specializing in radar theory--performed a more elaborate modification to the width of the cabinet and achieved a similar reduction in diffraction that smoothed out the peak.

 

[hardisj]

Yea, I think he contacted me a minute ago. I’m gonna try his mod out and re-test.

I believe the resonances are contributing to the peak and that the baffle step you talk about is impacted by this.

The reason I say this is because the peak jumps so quickly. Baffle step wouldn’t have such an immediate impact. Or I wouldn’t expect it to. of course, increasing the baffle extends this step and “covers up” the resonance.

What I’m saying is that I don’t think that you’re wrong. I just think the peak from resonance is a contributing factor.

 

[sarumbear]

Interesting comparison. I've worked with the Sony quite a bit and have done some cabinet and crossover mods, and I would take issue with Erin's explanation for the peak around 1 kHz. Erin attributes this to internal cabinet resonances and/or a port resonance, but it's really just inadequate baffle step compensation. There is a port resonance, but it's higher up--around 1600 Hz and doesn't show up in the frequency response measurement. Strengthening the cabinet, or adding more sound treatment won't help with the 1 kHz peak, which to my ears is the main issue with the speaker. You either have to build in more baffle step compensation in the low pass circuit, or add "wings" to either side of the cabinet to attenuate the baffle step somewhat. The latter approach is clumsy. The former is relatively easy, but you still won't have a speaker that goes very low or can play very loud. Within its limits, though, a modded version is capable of very decent sound.

If the issue is due to the 2Pi to 4Pi conversion step response, why it shows as an almost symmetrical low-Q peak?

Also, the waterfall clearly shows that there is a major resonance at 1kHz, which corresponds to the FR peak within a few tens of a Hertz. Coincidence?

 

[Dennis Murphy]

Yea, I think he contacted me a minute ago. I’m gonna try his mod out and re-test.

I believe the resonances are contributing to the peak and that the baffle step you talk about is impacted by this.

The reason I say this is because the peak jumps so quickly. Baffle step wouldn’t have such an immediate impact. Or I wouldn’t expect it to. of course, increasing the baffle extends this step and “covers up” the resonance.

What I’m saying is that I don’t think that you’re wrong. I just think the peak from resonance is a contributing factor.

That seems reasonable. It's certainly ugly. You get better resolution around 1 kHz, so it's certainly possible that the cabinet is contributing. I still have my Sony's, although I know one has a blown main tweeter (it's cheap) and the other may have my Xover mod, but I am curious to revisit this in light of your waterfall.

 

[hardisj]

I actually thought about seeing if you might be interested in taking a look at the Neumi speakers if I don't try to dig into them a bit more. Only if you a) have the time and b) think there would be a benefit to "fixing" it.

On my end I was thinking about disconnecting each component from the crossover and measuring it full-range (except for the tweeter) while still in the enclosure. Then loading the response curves into VCAD to see if I could come up with a more refined crossover. Either that, or just making adjustments on the fly with my DSP. In the "ports sealed" mode it looks reasonable but that one does seem to indicate baffle step issues as well. I just don't know if there is enough sensitivity there to make up for it because I'd hate to have to pad the tweeter down more.

If that's something you'd be interested in - or for any of the budget speakers I measure - let me know and I'd be happy to work together with you on improving what can be improved. Ultimately, what I'd like to provide is a resource for those on a budget to get some killer performance for as little money as possible.

 

[Dennis Murphy]

I actually thought about seeing if you might be interested in taking a look at the Neumi speakers if I don't try to dig into them a bit more. Only if you a) have the time and b) think there would be a benefit to "fixing" it.

On my end I was thinking about disconnecting each component from the crossover and measuring it full-range (except for the tweeter) while still in the enclosure. Then loading the response curves into VCAD to see if I could come up with a more refined crossover. Either that, or just making adjustments on the fly with my DSP. In the "ports sealed" mode it looks reasonable but that one does seem to indicate baffle step issues as well. I just don't know if there is enough sensitivity there to make up for it because I'd hate to have to pad the tweeter down more.

If that's something you'd be interested in - or for any of the budget speakers I measure - let me know and I'd be happy to work together with you on improving what can be improved. Ultimately, what I'd like to provide is a resource for those on a budget to get some killer performance for as little money as possible.

I certainly wouldn't mind taking a whack at the Neumi. I'm sure a smart guy like you could improve its crossover, but I think your time is a lot more valuable running the Klippel machine and writing great reviews. My only hesitation is the mess created by the front port. I'm not sure there's a lot I can do about that other than bring the overall level down a little.

 

[amirm]

The impact of the 1 kHz is small relative to overall boost in that region. You can see it in my measurents:

You can see just a 1 dB or so peaking which is not going to very audible. Erin's measurements of the same are too smoothed to show it.

EQ I developed fixes the broader deviation:

 

[sarumbear]

@amirm Why do you think there is a wide difference with the horisontal polar diagram test?

 

[dfuller]

@amirm Why do you think there is a wide difference with the horisontal polar diagram test?

It doesn't look substantially different to me. Amir's using 1dB increments where Erin is using 3dB increments.

 

[amirm]

@amirm Why do you think there is a wide difference with the horisontal polar diagram test?

As noted, these contour maps are tricky to create and interpret. You can easily change their looks with scaling, palate, etc. This is why I show the beam width with its standardized steps.

 

[sarumbear]

It doesn't look substantially different to me. Amir's using 1dB increments where Erin is using 3dB increments.

Yellow is at -12dB on both charts. There is certainly more resolution in @amirm's chart but the actual values assigned to colours are the same. Maybe the eye getting fooled.

 

[sarumbear]

As noted, these contour maps are tricky to create and interpret. You can easily change their looks with scaling, palate, etc. This is why I show the beam width with its standardized steps.

Do you think normalising will help?

 

[sweetchaos]

Here's both normalized....

The single image is Amir's.
The overlaid image is Erin's.
I don't see a difference.

 

[sarumbear]

Here's both normalized....
View attachment 172703
The single image is Amir's.
The overlaid image is Erin's.
I don't see a difference.

Thank you for doing this. Normalisation helps interpreting the data.