JBL 308P MKII Studio Monitor Review

[Pearljam5000]

@richard12511
Thanks.
It's good to know they're worth the price difference

 

[amirm]

Actually, it is not a big deal to test the noise of the monitor's tweeter from 1" distance, at least to compare with others.

I wanted to go this way but where is the exact end of a tweeter? Some have guards, others do not. Some have waveguides, others do not.

And with 1 inch distance, accuracy errors can be large with small variations. Longer distance helps with this but then signal to noise ratio becomes a problem.

I think we need to do this but need some protocol that is more defensible.

 

[richard12511]

Need to test a Phantom! Other companies are years out from something that compares as a package when it comes to dispersion, bass extension, and even response. If they made a non-app version with regular inputs and let users adjust the bass tilt, I'd put it up there as a Genelec 8260 alternative.

Send yours in

 

[Robbo99999]

Within their normal range, I haven't seen speakers change their response curve

Measuring in-room, at level too low, the ambient noise will intrude on the low frequency measurements, so, don't go too low.

Here is a set of measurements fom about 45 to 93dB, with 2db intervals in SPL, in-room.

You can see the confusion at low SPL (bottom left) where ambinet noise intrudes on the low frequency measures. From 70dB on up, the traces match despite increasing SPL

View attachment 92105

Yeah, cool, I'm familiar with the drawbacks of measuring at different SPL's for RoomEQ purposes, but I'm not talking about that. I'm talking about using the Anechoic Measurement Data that Amir has provided in order to create an "AnechoicEQ" to make the speaker flat. I'm ultimately trying to determine if I can use the data in his attached text file, without any kind of "normalisation process" to a different SPL, because the data is around 110dB whereas his graphs are around 85dB. I believe I can use the anechoic data "as is" without applying any "normalisation", as I have done by creating an "AnechoicEQ of the Listening Window" here: https://www.audiosciencereview.com/...-mkii-studio-monitor-review.17338/post-561866

(I'm finding it kinda surprising that a couple of experienced members haven't been grasping what I'm getting at, unless I'm missing something, but I don't thing I am, I merely trying to get confirmation that my EQ is valid based on using the data Amir has included "as is").

 

[amirm]

@amirm , or @MZKM or others - for EQ purposes is it OK for me to use the data in the CEA2034 file, because I notice it is at the higher 110dB level rather than the 85-90dB range seen in the graphs in the review. Do I need to apply a correction to the data, or can I just base my EQ on the data in the CEA2034 file which is at the higher 110dB level? (I'm thinking I can just use it at the current 110dB level, but wanted to be sure).

As explained, Klippel does not know about the amplifier gain in powered speakers. It thinks they are all passive so computes far higher output level. I have a manual process to compensate but it is trial and error and super slow process (10 minutes to try one alternative). To avoid questions about this speaker having been measured at too high a level, I did some manual compensation to bring it to what it more or less was (86 dB SPL). But forgot that I had created the spins at the original SPL levels. I can regenerate them if needed.

But yes, it doesn't matter if you use these artificially elevated levels or not for eq purposes.

 

[amirm]

On whether to EQ relative to on-axis or predicted in-room response (PIR), I have tried both across many speakers. To the extent your situation approximates that of the model behind PIR, then you are better off using that. For near-field that doesn't hold so I would use on-axis. In general I try on-axis first but if I fail to get what I expect, then I use PIR. If that also fails, then I give up.

As always you must listen and use your ears to confirm any EQ changes.

 

[Robbo99999]

As explained, Klippel does not know about the amplifier gain in powered speakers. It thinks they are all passive so computes far higher output level. I have a manual process to compensate but it is trial and error and super slow process (10 minutes to try one alternative). To avoid questions about this speaker having been measured at too high a level, I did some manual compensation to bring it to what it more or less was (86 dB SPL). But forgot that I had created the spins at the original SPL levels. I can regenerate them if needed.

But yes, it doesn't matter if you use these artificially elevated levels or not for eq purposes.

Exactly thanks, that is what I have been asking. This must mean that the deviations in dB around a central point (if you draw an average straight line of the response between say 50Hz and 20Hz), would be the same regardless of whether you have your data displayed at 110dB or the around 85dB that you normally show your graphs. I guess I'm right in this, unless I hear back.

 

[amirm]

You are correct. Just make sure when you plot these yourself in REW, etc. that you use the same 50 dB scale. Otherwise the higher max SPL will compress the graph and hide issues.

 

[Robbo99999]

You are correct. Just make sure when you plot these yourself in REW, etc. that you use the same 50 dB scale. Otherwise the higher max SPL will compress the graph and hide issues.

Excellent, thanks, that's what I did, the same 50dB window.

 

[amirm]

On JBL305P measurements, I took my existing data and applied my current numerical optimizations to it. It basically made no difference. The 305P was the second speaker I ever measured with the system and has not benefited from some other optimizations I use now. To wit, it has only 500 measurement points compared to 800+ I use now. And as posted, it has 10 points/octave rather than 20. And of course we don't have full suite of tests. Time permitting, I will measure it again.

For now, it is not in the same class as the JBL 308P. It can't play as loud. Doesn't have nearly as much bass. And I don't think it sounds as linearly as 308 can.

 

[RayDunzl]

Yeah, cool, I'm familiar with the drawbacks of measuring at different SPL's for RoomEQ purposes, but I'm not talking about that. I'm talking about using the Anechoic Measurement Data that Amir has provided in order to create an "AnechoicEQ" to make the speaker flat. I'm ultimately trying to determine if I can use the data in his attached text file, without any kind of "normalisation process" to a different SPL, because the data is around 110dB whereas his graphs are around 85dB.

I thought I demonstrated above that different SPL doesn't change frequency response.

Oh well.

 

[amirm]

I thought I demonstrated above that different SPL doesn't change frequency response.

Oh well.

He wanted a cooler cat explaining it!

 

[Robbo99999]

I thought I demonstrated above that different SPL doesn't change frequency response.

Oh well.

He wanted a cooler cat explaining it!

In fairness it's not quite the same mechanism involved, even though the end result happens to be similar....so wanted definitive confirmation. Ha, so wasn't "Cool Cat Factor", you're both cool, no denying baby!

 

[BostonJack]

It is on purpose, trebles wouldn't glide so smooth..

No white finish option and no winter sales in my region are bigger crimes imho

That clip makes dumping a bike look like a superpower! Talk about athletic kinesthetic awareness!

 

[stevenswall]

Send yours in

Don't have it anymore. "Just" my Genelec 8260 monitors, an IN8 and Hivi that have already been tested, and a Google Home Max.

I'd chip in $50 though if a couple dozen people wanted to get one. After testing have whomever makes the largest donation get it.

 

[DuncanTodd]

Personally I find this speaker to be quite beautiful. In particular, love the shape of the waveguide. I also own the JBL 530, and I agree with you there. That is an ugly speaker, especially if you take the grill off .

Interesting. The waveguide is what I don't like about it, it's even worse when it looks like cheap shiny plastic like on this one. On the 5XX series they don't have grilles over the waveguide which is what I find most offputting. They look like bholes... I use some Arenas for surrounds and I think they have a similar design but luckily the grilles take care of that issue

 

[tuga]

Adam once made a floor stander (Tensor Beta), don't know the price (probably high). It's no longer available.

 

[MZKM]

I wanted to go this way but where is the exact end of a tweeter? Some have guards, others do not. Some have waveguides, others do not.

And with 1 inch distance, accuracy errors can be large with small variations. Longer distance helps with this but then signal to noise ratio becomes a problem.

I think we need to do this but need some protocol that is more defensible.

I don’t think those instances really matter, simply measure distance from the baffle.
Its not like you are going to sit closer to a speaker where the tweeter is sunken in a waveguide.

I’d say 2ft is a close but not unrealistic listening distance for computer speakers. Or measure at 1/2m (~20in) and then we extrapolate to the standard 1m. Or, just measure at 1m. You only need to be reflection free down to like 1500Hz.

 

[LTig]

Ooops - very expensive.

 

[CerealKiller]

308P's stand-alone or 305P's combined with a subwoofer. Anyone have any opinions/experiences they care to share?

I own 305p + PreSonus temblor t10. Used it that way for a year. Its good at nearfield on the desk. Lately took them to the living room,
tried the combo for some months and finally removed the sub, it Made some trouble in the form of resonancia, front back walls aré not suficiently apart i supouse, regardles, i can make the 305 distort if driven too loud but thats simply too loud for my listening gaste. Temblor T10 Is overpowered for 305p