Kali Audio IN-8 Studio Monitor Review

[amirm]

Very much look forward to Kali Lp6 review if it happens.

Same here. Just had an offer to test the LP-6 so hopefully we get to do that soon.

 

[beefkabob]

Reviewers get sent the equipment. If you have a big audience, you may get to keep it. Then you don't get sent anything if you have give negative reviews or if you talk about objective things. It's almost all a huge scam, plain and simple. 99% chance of getting fd.

 

[3ll3d00d]

It lets you change their colors or if you want them displayed. Nothing much past that. I need to export the data and plot it elsewhere for something different. If anyone knows of a graphics app or website that creates 3-D graphs from these, let me know. Short of that, I may have to write something in matlab.

I wrote the qt/python app that was in the diyaudio thread on this subject (parked in https://github.com/3ll3d00d/pypolarmap). It produces such views from imported IR data with selectable colour maps (someone made the point in the other thread about how perceptually uneven jet is) and normalisation on a selected reference axis. If you want something interactive them probably easy to resurrect that though not hard to write some simple script to do this either. In what format does it export data?

 

[folzag]

Reviewers get sent the equipment. If you have a big audience, you may get to keep it. Then you don't get sent anything if you have give negative reviews or if you talk about objective things. It's almost all a huge scam, plain and simple. 99% chance of getting fd.

Yep. Get the equipment and a nominal honorarium/writing fee. Post a glowing review. Sell the equipment on Craiglist or eBay. Rinse and repeat.

I bought two pairs of Sony WH-1000MX3's just last week in exactly this scenario. The dude had 5 pairs! I was suspicious they were stolen, but after meeting him in person and chatting a while, nope. Just a part-time online reviewer.

 

[amirm]

I wrote the qt/python app that was in the diyaudio thread on this subject (parked in https://github.com/3ll3d00d/pypolarmap). It produces such views from imported IR data with selectable colour maps (someone made the point in the other thread about how perceptually uneven jet is) and normalisation on a selected reference axis. If you want something interactive them probably easy to resurrect that though not hard to write some simple script to do this either. In what format does it export data?

Thanks. It supports ASCII, VACS and binary Matlab from what I have seen (these are measurement graph exports).

 

[folzag]

@amirm, are you doing the listening tests with them on a desk top? Shouldn't there be two listening tests, near-field and far-field?

There's the assertion about the predicted in-room response, which I take to mean far-field, it would be interesting to validate that subjectively as well.

 

[confucius_zero]

The horror of the hype train...

 

[napilopez]

Thanks once again amir! I appreciate you breaking down the early reflections curve into its component curves - very useful Imo.

I have to say though, I don't read this performance quite as poorly as many of you - thoug less then expected from Kali given it's made of ex-JBL engineers. Still, the JBL Control 1 is definitely worse in my book, and I think the NHT too.

My impression is that the speaker has some issues, but they seem easy to fix. Some potential considerations so maybe @Sancus doesn't have to give it away to someone he hates:

First, the DI curve is super smooth, and most of the bumps and dips are present on the big three curves(LW, Early reflections, sound power). That means EQ-ing this speaker to an almost ideal response should be pretty trivial, since none of its problems appear to be major directivity issues.

Normally we don't want to EQ above the schroeder frequency, but since we have anechoic data (again, thanks amir!), just bring down the bumps that are consistent at every angle, or bring up that low-mid/upper bass dip.

Looking at the SPL graphs, the treble is also better-controlled than I expected, considering coaxial drivers from most companies many companies that aren't called KEF often lead to some messiness.

Does this include SIBR effects which is dependent on distance from front wall? Some speaker designs include wall support and need to be very close to the wall.

It is a computed curve which shows a theoretical average of SBIR effects.

@Thomas_A makes a good point. The predicted-in-room PIR is not particularly reliable below the schroeder frequency. Seems to me like proper placement in a real room help fill in that bass.

Notably, see how from 400-20KHz, the speaker is very linear and has the almost perfect amount of tilt.
That said, I did a quick check and Kali's manual says ideal position is "at least 0.5 meters (about 20 inches) away from any walls)," so close wall placement doesn't seem to be the intention. Even at that distance though, you should get some reinforcement.

Does the software allow normalization for contour plots? Would make the directivity more readable..

Lastly, I agree with @gr-e that the contour plots should be normalized to the listening window. Contour plots are meant to show directivity, after all. If not normalized to the on-axis, the directivity info becomes much harder to puzzle out if the speaker doesn't have a perfectly linear response. The errors we see in the contour plot seem to be mainly the same ones we already see in the on-axis.

For me, good directivity is almost always the priority. The rest can be largely fixed, especially speakers meant for enthusiasts with easy access to EQ.

EDIT: To be clear though: speakers should be good out of the box and following manufacturer directions. You shouldn't have to fiddle with them far beyond what the manufacturer tells you to in order to get them to sound good. That's as much a design error as any other.

But speaking from the perspective of an individual speaker enthusiast, I'd rather have a speaker with well controlled directivity than one that's perfectly flat.

 

[amirm]

@amirm, are you doing the listening tests with them on a desk top? Shouldn't there be two listening tests, near-field and far-field?

For these monitors, yes. They are typically used this way. For stand-alone speakers I will use a different setup.

 

[urfaust]

Disapointing results, the smaller 6" versions also got quite some hype back then. I wonder how those would compare since they have inherently a different design without coaxial. This is going to get fascinating once there s a good sample of classic studio monitors.

 

[amirm]

First, the DI curves super smooth, and most of the bumps and dips are present on the big three curves(LW, Early reflections, sound power).

The DI curve for early reflections is broken. Klippel has confirmed and testing a fix. It is not yet working in the main graph but is as a separate graph:

As you see, it is not smooth at all.

 

[aarons915]

First, the DI curves super smooth, and most of the bumps and dips are present on the big three curves(LW, Early reflections, sound power). That means EQ-ing this speaker to an almost ideal response should be pretty trivial, since none of its problems appear to be major directivity issues.

Normally we don't want to EQ above the schroeder frequency, but since we have anechoic data (again, thanks amir!), just bring down the bumps that are consistent at every angle, or bring up that low-mid/upper bass dip.

This was what I was getting at in another thread, measuring the speakers is a great service to consumers but I was curious if the Klippel software is able to provide PEQ filters to match a target (ie flat) similar to REW. If not can the data be exported into REW so we can mess around with the EQ function ourselves? All that would really be required are the Spinorama curves, not every measurement. You can eyeball the graph, of course, but it would be a bit more precise with some kind of software giving us the PEQ values.

Since the IN-8 does have such well behaved directivity indices, simply EQ'ing the listening window to be neutral should make the whole speaker very good.

 

[SIY]

Same here. Just had an offer to test the LP-6 so hopefully we get to do that soon.

You'll like it. Remarkably good vertical dispersion besides very flat on-axis response.

 

[mhardy6647]

... I mean this speaker is very cheap for an active 3-way with 8" woofer...

I think this speaks volumes (no pun intended), don't you?

* active
* three-way
* 8" woofer
[* coaxial MR/tweeter]
* cheap

I mean, what's wrong with this picture?

 

[amirm]

This was what I was getting at in another thread, measuring the speakers is a great service to consumers but I was curious if the Klippel software is able to provide PEQ filters to match a target (ie flat) similar to REW. If not can the data be exported into REW so we can mess around with the EQ function ourselves? All that would really be required are the Spinorama curves, not every measurement. You can eyeball the graph, of course, but it would be a bit more precise with some kind of software giving us the PEQ values.

Since the IN-8 does have such well behaved directivity indices, simply EQ'ing the listening window to be neutral should make the whole speaker very good.

Klippel NFS can generate the compensation curve that once applied to in-room measurements, produces anechoic results. It is relative to measured in-room results however. So not the same as what REW does.

 

[Sancus]

Some potential considerations so maybe @Sancus doesn't have to give it away to someone he hates:

No worries, it's already on its way back to the vendor(Not to make this an ad -- but I bought it from Sweetwater and their customer service has been fantastic). I'm not keeping it.

Given I had planned to use it horizontally placed as a center, dealing with that mess of a vertical response doesn't seem worth it. I mean look at those asymmetric hotspots. Yuck.

 

[napilopez]

The DI curve for early reflections is broken. Klippel has confirmed and testing a fix. It is not yet working in the main graph but is as a separate graph:
View attachment 45867

As you see, it is not smooth at all.

Yeah, I remember there was an issue in the main graph. I was referring to the SPDI and inferring the ERDI from the LW/ER graphs.

I guess we're reading it differently but your 2dB scaling above does exacerbate the bit of treble irregularity in the ERDI. If I scale it back to match the aspect ratio of the CTA-2034A it looks like this:

Which doesn't look all that bad to me.

This was what I was getting at in another thread, measuring the speakers is a great service to consumers but I was curious if the Klippel software is able to provide PEQ filters to match a target (ie flat) similar to REW. If not can the data be exported into REW so we can mess around with the EQ function ourselves? All that would really be required are the Spinorama curves, not every measurement. You can eyeball the graph, of course, but it would be a bit more precise with some kind of software giving us the PEQ values.

Since the IN-8 does have such well behaved directivity indices, simply EQ'ing the listening window to be neutral should make the whole speaker very good.

No need to eyeball the graphs or make Amir do any more work =].

Vituixcad, a free speaker design app, has a very handy "SPL Trace" function. Open Vituixcad, go to Tools, select SPL trace. Import a screengrab of the klippel measurements. Set your graph boundaries and SPL levels, then tap on the curve you want to copy, which in most cases would be the listening window for PEQ purposes. It looks like this:

Might take a bit of adjustment to get it just right, but it's usually pretty easy. Then just tap export, and you have a response you can import into your app of choice and create filters for.

 

[aarons915]

No need to eyeball the graphs or make Amir do any more work =].

Vituixcad, a free speaker design app, has a very handy "SPL Trace" function. Open Vituixcad, go to Tools, select SPL trace. Import a screengrab of the klippel measurements. Set your graph boundaries and SPL levels, then tap on the curve you want to copy, which in most cases would be the listening window for PEQ purposes. It looks like this:


Might take a bit of adjustment to get it just right, but it's usually pretty easy. Then just tap export, and you have a response you can import into your app of choice and create filters for.

Thanks a lot, that is awesome and would be really useful. I also agree the ERDI isn't that bad but it would take quite a bit of EQ to fix all of it's issues, that's a speaker I would pass on personally. Many speakers have a few resonances that could be fixed and made to be much better, a good example is the LS50 that I currently use, here's a "Spin" I found in a study that doesn't name the speaker but calls it a "2-way Coaxial with an MSRP of $1499", so I'm pretty sure it's the LS50.

You can see that fixing the listening window with a few filters at 750Hz and 2k would make this a much smoother speaker, which I have done and can confirm. This is a good example of a speaker with a few problems that can be made even better with just a few filters

 

[napilopez]

Thanks a lot, that is awesome and would be really useful. I also agree the ERDI isn't that bad but it would take quite a bit of EQ to fix all of it's issues, that's a speaker I would pass on personally. Many speakers have a few resonances that could be fixed and made to be much better, a good example is the LS50 that I currently use, here's a "Spin" I found in a study that doesn't name the speaker but calls it a "2-way Coaxial with an MSRP of $1499", so I'm pretty sure it's the LS50.

View attachment 45871

You can see that fixing the listening window with a few filters at 750Hz and 2k would make this a much smoother speaker, which I have done and can confirm. This is a good example of a speaker with a few problems that can be made even better with just a few filters

Yeah, the LS50W measures a fair bit flatter than the LS50, and they basically did the same right in the DSP, its seem. (still has the 1.5K dip but disappears off axis, so not hugely audible). Where did you find that study, btw? First time i've seen a spin of the LS50!

There are major issues with Kali's response but the biggest problem is definitely that 100-400-ish scoop imo. You're 5dB down from 500Hz to 300Hz, ouch. That would be the first thing I'd fix, and then the in-room response wouldn't be so bad.

One of the things I'm really grateful about the Klippel. Since almost everyone else (including myself) is using gated measurements for their measurements, this region tend to have really low resolution, so such issues might not be readily apparent.

But yeah, there definitely seem to be better options in the price range that require less effort, unless you really need that vertical directivity because you do your mixing while jumping on a trampoline or something.

 

[HammerSandwich]

The DI curve for early reflections is broken. Klippel has confirmed and testing a fix.

Stunning. How many have they sold? Starting when? And this bug is found by a crazy hobbyist's website?!??

Great work, everybody! Especially @amirm. Thank you!