JBL 708i Monitor Review (DSP: Part 2)

[Naturlyd]

Because the tweeter level was too hot based on my acoustic measurements -- something no tech will be able to do, not easily anyway. What you saw was post compensation of the levels using the trim controls by me.

I am afraid the elevated level is an artifact of the DSP not beeing up to par and might be missing a EQ point or the correct level at some point. Its really frustrating. Here you go with a more or less perfect baseline/construction. With the prospect of over the air upgrades on EQ and other ways of optimization. To end up with a speaker system beeing worse several years after release..

 

[jhaider]

I think it makes you more critical of speakers you see in the future because you've had more exposure to good speakers that have been reviewed with perfect or near perfect spinoramas (at various price points). I looked back at my comments in the 708p ASR thread and it looks like I quite liked the speaker, but my opinion has changed since then, probably for the reasons I mentioned.

I think it’s noteworthy that a speaker made by an engineering team with the direct feedback loop to the research that went into designing the Spinorama model and access to blind testing facilities made the choices they did.

That is a separate discussion from issues of production trustworthiness. Two of three 7-series samples measured by @amirm - 705P and this 708i - showed clear defects IMO. Only 708P measured as expected.

4) Why did you have to adjust the input trims? In the review one get the impression that level matching between the low and high channels are done with the trims. The level matching is done by the config file from Harman. Trim is for adjusting to the sources output level. As one can adjust the amplifiers gain modes 26/34/37dBu (in Audio Architect..) my gain structure is usually fine with the trim fully open.

I hadn’t considered it before, as I’ve never biamped mine - but if the analog pots are not internally ganged between two channels when the "biwire" tuning is used, that seems like a design flaw to me.

I agree with this poster. This speaker is part of s system but it was measured in both occasions as a module of that system.

I will repeat again: these are not standalone speakers. They are elements of a sound system. Find someone to set them up or use the P versions. Simples!

I don’t understand why you think that. Here what was measured was the complete DSP-amplification-transducer chain. That is the system. Just add source, and done.

At any rate, I'm generally fairly dumb, have zero formal education in electrical engineering or acoustics, and no experience setting up sound systems outside of the home audio context - yet I managed to set up a trio of 708i's with a Crown DCIn amp just fine even though Audio Architect does not run on Macs and I generally resent having to use windows for anything personal.

 

[sarumbear]

I don’t understand why you think that. Here what was measured was the complete DSP-amplification-transducer chain. That is the system. Just add source, and done.

There was no crossover, was there in the part 2 test? The JBL documentation (and I agree with everyone that they are bad) shows that the BSS unit is doing the LP/LF filtering. Your posts show no such filtering either, just EQ.

 

[jhaider]

There was no crossover, was there in the part 2 test? The JBL documentation (and I agree with everyone that they are bad) shows that the BSS unit is doing the LP/LF filtering. Your posts show no such filtering either, just EQ.

The Crown DCIn amp used provided all of the intended processing.

 

[sarumbear]

The Crown DCIn amp used provided all of the intended processing.

Do you see the effect of a high order LP filter on the woofer on this chart?

Besides, on his post @MAB showed us that the DSP is not acting as a crossover. Do you disagree with that?

In summary of what the speaker actually is: It's designed to use one or two amplifiers per customer's choice. It has passive crossovers only. It uses DSP for shelving and notch filtering only.

So that you don't misinterpret my words, here is a picture of the DSP filters directly from London Architect EQ files for HF and LF:

 

[jhaider]

Do you see the effect of a high order LP filter on the woofer on this chart?… the DSP is not acting as a crossover. Do you disagree with that?

Irrelevant. All of the intended processing to complete these speakers is in use as tested by @amirm.

We would see the same result from the speakers tested with BSS macros. Refer to my earlier post (specifically, the magnitude of the ~7kHz lump compared to what @amirm measured) to see why those filters don’t work as they should in this case.

 

[sarumbear]

Irrelevant. All of the intended processing to complete these speakers is in use as tested by @amirm.

We would see the same result from the speakers tested with BSS macros. Refer to my earlier post (specifically, the magnitude of the ~7kHz lump compared to what @amirm measured) to see why those filters don’t work as they should in this case.

There is no effect of a HP filter nor on the filters posted but that is irrelevant?

I read your linked post you said:

It's worth noting that, as I pointed out in Part I, the engineers behind 7-Series were not bullish on the biamp option. They said it added little, and explicitly expressed a preference to me and many others around that time for a stronger amp channel in single amp mode over two channels in biamp mode.

which is pretty odd as the P version is a bi-amp active speaker. If bi-amp adds little why would they used it? Unnecessarily increase the price?

 

[jhaider]

There is no effect of a HP filter nor on the filters posted but that is irrelevant?

Yes. It’s working as designed. There are clearly highpass and lowpass filters in place in biamp mode.

which is pretty odd as the P version is a bi-amp active speaker. If bi-amp adds little why would they used it? Unnecessarily increase the price?

That’s what the market expects in powered studio monitors. It’s pretty clear they don’t think there’s a performance difference - they market i and P with the same spins, and the P sample measured by Amir and Erin validated that specification.

Also those chip amps are a lot less expensive. Maybe a single 600W one would be more expensive.

 

[crappypanther]

Glad I didn't buy this. Although there was once a desire due to the good reputation of the JBL LSR6328P

 

[GXAlan]

The DCi 4|300N is an officially supported configuration.

They do mention the speaker out polarity being - + + - which is different than expectations and correlates with a comment that one of the drivers may have been out of phase when tested. @amirm

https://jblpro.com/en/site_elements/7-series-set-up-guide

 

[amirm]

The DCi 4|300N is an officially supported configuration.

They do mention the speaker out polarity being - + + - which is different than expectations and correlates with a comment that one of the drivers may have been out of phase when tested. @amirm

https://jblpro.com/en/site_elements/7-series-set-up-guide

I noted the odd configuration and wired it as such. Checked again when I took it apart to ship.

 

[GXAlan]

I noted the odd configuration and wired it as such. Checked again when I took it apart to ship.

Thanks for confirming!

 

[sarumbear]

Yes. It’s working as designed. There are clearly highpass and lowpass filters in place in biamp mode.

Please show us in @amirm’s test results. I can see no effect of any crossover. It’s only fair that you explain that clarity you say in results.

 

[jhaider]

Please show us in @amirm’s test results. I can see no effect of any crossover. It’s only fair that you explain that clarity you say in results.

I don’t know what you want, but based on the measured response the filtering is working as intended, except for some differences resulting from variance in the raw tweeter response, as clearly shown by comparative measurements.

 

[Robbo99999]

Please show us in @amirm’s test results. I can see no effect of any crossover. It’s only fair that you explain that clarity you say in results.

I don’t know what you want, but based on the measured response the filtering is working as intended, except for some differences resulting from variance in the raw tweeter response, as clearly shown by comparative measurements.

Well in support of what @sarumbear is saying whenever I've worked with crossovers in miniDSP for my sub & speakers the roll off they create is a lot steeper than what we see with the woofer in the following (unless the measurement of the woofer is slightly contaminated by the tweeter and can't be fully trusted):

There's no "exponential" roll off happening there with the woofer, it's just at gradual same slope decline almost. The crossovers I use in my miniDSP for sub & speakers is a 24dB something or other (the standard default crossover used in miniDSP, is it Linkwitz-Reilly) - yeah and the roll off is really sharp (really sharp) on those in contrast to what's happening with the woofer above. Mind you the tweeter roll off looks more analogous to a crossover though, so maybe the woofer measurement is contaminated but the tweeter measurement is maybe more accurate, I don't know (unless the tweeter just falls off a cliff naturally at lower frequencies without any crossover intervention).

EDIT: but Amir did say that there was a passive crossover active within the speakers at all times regardless of whether bi-amping or not. So there's obviously somekind of crossover already being applied, which is why it isn't included within the EQ filters for the speakers:

So I think the woofer measurement is contaminated, unless the passive crossover already included in the speakers chooses to roll off the tweeter more severely than the woofer.

 

[beatelund]

Well in support of what @sarumbear is saying whenever I've worked with crossovers in miniDSP for my sub & speakers the roll off they create is a lot steeper than what we see with the woofer in the following (unless the measurement of the woofer is slightly contaminated by the tweeter and can't be fully trusted):

There's no "exponential" roll off happening there with the woofer, it's just at gradual same slope decline almost. The crossovers I use in my miniDSP for sub & speakers is a 24dB something or other (the standard default crossover used in miniDSP, is it Linkwitz-Reilly) - yeah and the roll off is really sharp (really sharp) on those in contrast to what's happening with the woofer above. Mind you the tweeter roll off looks more analogous to a crossover though, so maybe the woofer measurement is contaminated but the tweeter measurement is maybe more accurate, I don't know (unless the tweeter just falls off a cliff naturally at lower frequencies without any crossover intervention).

EDIT: but Amir did say that there was a passive crossover active within the speakers at all times regardless of whether bi-amping or not. So there's obviously somekind of crossover already being applied, which is why it isn't included within the EQ filters for the speakers:

So I think the woofer measurement is contaminated, unless the passive crossover already included in the speakers chooses to roll off the tweeter more severely than the woofer.

For the tweeter in 708i the 2409H 8ohm they are using a condensator with 1uF. It is always active regardless of single wire or bi amp mode.

In bi amp mode the passive crossover for the woofer is bypassed
and it is using a 1,5khz LR24db crossover both in the BSS presets and Crown DCi N

 

[beefkabob]

Not having to fuck with all the DSP work is exactly why I wanted the 708p over a DIY or DSPIY solution like this.
I think anyone seriously using the 708i is going to hire a professional to do major room correction on a many-channel, professional system, like a 10+ speaker theater, and these will be the surround speakers more likely than not.

 

[Robbo99999]

For the tweeter in 708i the 2409H 8ohm they are using a condensator with 1uF. It is always active regardless of single wire or bi amp mode.

In bi amp mode the passive crossover for the woofer is bypassed
and it is using a 1,5khz LR24db crossover both in the BSS presets and Crown DCi N

Interesting, thanks. So, for the tweeter, what kind of digital crossover would that be analogous to? So Linkwitz-Reilly 24dB for instance or something else? Regarding the woofer, I understand what you mean, but it's not being shown in this pic, would the crossover not be shown in:

Also from Amir's measurement of the woofer it certainly doesn't look like a 1.5kHz LR24dB, but perhaps his measurement is being contaminated by the tweeter.

 

[beatelund]

Interesting, thanks. So, for the tweeter, what kind of digital crossover would that be analogous to? So Linkwitz-Reilly 24dB for instance or something else? Regarding the woofer, I understand what you mean, but it's not being shown in this pic, would the crossover not be shown in:

Also from Amir's measurement of the woofer it certainly doesn't look like a 1.5kHz LR24dB, but perhaps his measurement is being contaminated by the tweeter.

Tweeter filter is 20khz HF Butter 6db.
It also has a 11ohm resistor in parallell so i dont know how that converts. But i suppose it Will be lower Maybe 12-15khz But 6db.

 

[jhaider]

I think anyone seriously using the 708i is going to hire a professional…

It’s really not hard, especially if you just run it in single amp mode. Which is what the engineers responsible for the speakers recommended anyway.

After all this I’d be taking measurements of any B-stock 7-series purchases to make sure the tweeters aren’t out of spec. Unfortunate but true.