Four KM-84 Clones - Blind Test

[joulupukki]

Interesting. So with C2 at 2.2/3.9pF the high pass frequency shifts even higher? …and the mic would have an even higher sensitivity in the high end?

Ignore that question. I loaded those same .mdat files in REW and now understand what you mean. I should put in a pad capacitor in C2 in the MP mic and see what ends up happening in real life recordings. Ah, except I don’t wanna touch those tiny screws in that mic body again. I need to find some replacements. When I first started with those mics I wasn’t delicate enough and I ended up nearly ruining the slot on the screw head. :-|

 

[joulupukki]

I don't know if anyone else will find this interesting, but since I can't currently record all 6 microphones at the same time (because I need to get some better XLR connectors), I decided to see how pink noise is recorded by each of them. In the following image, what you'll see is that the KM84+ is used as the "base" EQ. I used FabFilter's Pro-Q3 to do an EQ match with the KM84+ recording pink noise.

If you see the yellow line going UP past zero, that means that in order to sound like the KM84+, the mic EQ has to be raised at that point. Lower than 0 means the EQ at the given frequency has to be lowered.

Something else interesting (confirming what I'm hearing) is that I swapped the MP SDC-84 capsule for a 3U Audio Capsule and did a recording with the MP SDC-84 of pink noise. See the big drop at the high end frequencies on the stock MP SDC-84 capsule?

With the KM84++ I'm seeing that the low end needs to be raised to match the KM84+. Also, like I can hear in the recordings, 5-15kHz has to be raised in order to match the KM84+ ... confirming why I'm not hearing as much definition out of the KM84++ as the KM84+.

This also includes the stock Takstar CM-60 and CM-63 and it's interesting to see that they both have a mid scoop compared with the KM84+. The CM-63 has a lot more high end above 8kHz compared with the KM84+. When I was comparing the CM-60 earlier this year against my Graeme Woller KM84 build, the CM-60 sounded really close except it had significantly more low end.

 

[jp8]

Interesting. So with C2 at 2.2/3.9pF the high pass frequency shifts even higher? …and the mic would have an even higher sensitivity in the high end?

No, maybe my description wasn't clear, but it's actually the other way around. The green line shows low-end roll-off with C2 = 0 pF and roll-off point moves to the left to a lower frequency.

Gain (or sensitivity) increases when C2 decreases and is limited by the open-loop gain of the JFET.

Jan

 

[joulupukki]

I fixed up the commercial XLR cables I had by pulling them apart and re-soldering the broken wires. I should have thought of doing that long ago. They should now last a much longer time.

Logistically, how could I best record 6 microphones at once? I've got 4 of them plugged in to a Zoom H6 and 2 of them into the Apollo Twin X but ...

They are now spread out far enough that I'm not sure that it's gonna be a fair enough comparison because I know even some movement of mic position on a source can make a difference. Thoughts? Here's my crazy makeshift mic clip. In initial recordings the stock Takstar mics definitely have less definition than all of the others and I can hear the mid scoop in them.

 

[kingkorg]

I fixed up the commercial XLR cables I had by pulling them apart and re-soldering the broken wires. I should have thought of doing that long ago. They should now last a much longer time.

Logistically, how could I best record 6 microphones at once? I've got 4 of them plugged in to a Zoom H6 and 2 of them into the Apollo Twin X but ...

They are now spread out far enough that I'm not sure that it's gonna be a fair enough comparison because I know even some movement of mic position on a source can make a difference. Thoughts? Here's my crazy makeshift mic clip. In initial recordings the stock Takstar mics definitely have less definition than all of the others and I can hear the mid scoop in them.

View attachment 140014

They are both spread between each other, and they obstruct each other's vents significantly. It is best to put each one on a stand in same position, same distance from the source, play same exact clip of pink noise, or some other material. Noise is most revealing tho. Even if you make couple millimeters of error in positioning it will be less than this.

 

[jp8]

How about using a stereo mic bar with those long slots for positioning the mics? Add 4 more screws, if you can find them. Then mount the mics as if in a line array, but then in a concave shape. Sound source positioned in the focal point. Perhaps flip the bar and record for a second time. Upper mic, which becomes lower mic in second take, should sound identical in both takes to confirm that sound field is symmetrical around the center of the bar.

Jan

 

[joulupukki]

Yeah, good points. Ideally I'd like to record the exact same thing with all of the mics at once since it'd be the same performance. I used to have a nice stereo bar that came with the SE Electronics sE5 Stereo Pair I had, but sold the whole thing after I realized I liked these KM-84 variants a LOT more.

Maybe to get the general idea of these mics I could just record them two at a time and try the best I can to keep the same mic placement and similar performance volume on each of the instruments. Since it'll be a little longer than a 10-second passage the mic characteristics would better be heard for each one. If I stay about 16" away from the mics when recording I could probably do them two at a time and the distance between two mics would more or less be negligible.

Sorry to say, but the blind test from post #1 was done with 4 of these mics being taped closely together. But, each mic had that "disadvantage" so it's not like one of them had it "better" than the other. Definitely not ideal, but for the experiment it did allow a decent capture of the mic characteristics.

 

[museric]

Thank you @joulupukki for this blind test, and to @jp8 and other participants for all the info - this was great fun! I feel like a kid in a candy store

The reveal was fascinating to me. What I was most surprised by was my second choice being the KM84++. I was half-expecting the MP SDC-84 to be in the top two - and I'm much more excited for the current and future Takstar mod possibilities. I'm going to be picking up a pair of the U3 capsules as well.

I do agree with the assessment to have more air around the physical microphones when recording side by side comparisons, though I'm not sure, if done differently, that this test from first post would have had different results. Cheers everyone!

 

[joulupukki]

Thank you @joulupukki for this blind test, and to @jp8 and other participants for all the info - this was great fun! I feel like a kid in a candy store

The reveal was fascinating to me. What I was most surprised by was my second choice being the KM84++. I was half-expecting the MP SDC-84 to be in the top two - and I'm much more excited for the current and future Takstar mod possibilities. I'm going to be picking up a pair of the U3 capsules as well.

I do agree with the assessment to have more air around the physical microphones when recording side by side comparisons, though I'm not sure, if done differently, that this test from first post would have had different results. Cheers everyone!

Thanks for the comments. With regard to having more space around the mics, the effect of how I did the test in this case seems negligible. The mics, in isolation by themselves seem to still have the same overall characteristics as they do in the blind tests. I’m working on a new set of recordings “in a mix” where the mics aren’t on top of one another and they still seem to exhibit the same frequency response nuances.

 

[museric]

Thanks for the comments. With regard to having more space around the mics, the effect of how I did the test in this case seems negligible. The mics, in isolation by themselves seem to still have the same overall characteristics as they do in the blind tests. I’m working on a new set of recordings “in a mix” where the mics aren’t on top of one another and they still seem to exhibit the same frequency response nuances.

Good to know, and I'm looking forward to any updates, too

 

[soapfoot]

Now I’d be curious to hear this test repeated but with 3 genuine KM84 and one KM184

Would be interesting to hear

1) how similar the vintage 84s sound (in my anecdotal experience this is likely to be “not very”

2) whether the 184 is an extreme outlier, or whether a random trio of vintage 84s are too divergent to be able to reliably spot the 184

 

[joulupukki]

Now I’d be curious to hear this test repeated but with 3 genuine KM84 and one KM184

Would be interesting to hear

1) how similar the vintage 84s sound (in my anecdotal experience this is likely to be “not very”

2) whether the 184 is an extreme outlier, or whether a random trio of vintage 84s are too divergent to be able to reliably spot the 184

I would love to try this too. I've found a local studio that has two ... just need to get their attention and time. They seem to be really busy! I'd love to know how well these clones stack up to them.

 

[Paul W]

I would love to try this too. I've found a local studio that has two ... just need to get their attention and time. They seem to be really busy! I'd love to know how well these clones stack up to them.

You might want to borrow (if you can) or just buy an inexpensive calibrated measurement mic, so that you can do reasonable frequency response measurements and not just relative ones.

Does that studio have one you can use, I wonder? If not, the Dayton Audio EMM-6 is only $55 right now w/free shipping at Amazon.

 

[joulupukki]

You might want to borrow (if you can) or just buy an inexpensive calibrated measurement mic, so that you can do reasonable frequency response measurements and not just relative ones.

Does that studio have one you can use, I wonder? If not, the Dayton Audio EMM-6 is only $55 right now w/free shipping at Amazon.

Thanks for the tip!

 

[joulupukki]

Ok, here's a bonus blind test! Sorry, no video this time and you'll have to pull down the wav files yourself to listen.

In no particular order, the mics in this test are:

• Microphone Parts SDC-84
• @jp8 KM84+
• @jp8 KM84++
• Takstar CM-60
• Takstar CM-63
• My Graeme Woller KM84 build in a Takstar CM-60 body

I no longer have the KM-184 since I had to return it or I would have included it.

This collection of WAV files contain static mixes from each mic. The instruments in this track are acoustic guitar, mandolin, banjo, and upright bass. No fancy playing, just quick and dirty to get something. The only thing I've done is a little panning and some crude volume balance and automation. Oh, and I did add a tiny bit of room/reverb to all instruments except for the upright bass just so it didn't feel completely dry.

The purpose of this test was for me to see how these microphones sound if I were to stack them in a mix from every instrument. I can hear differences between the mics. It's interesting hearing good / bad qualities based on listening to a particular instrument/mic combo. I wouldn't ever record this way for a real album. Two of these mics I definitely wouldn't use.

I recorded the mics in pairs of two so there are three "performances". Mics 1 and 2 were recorded together as were 3 and 4 ... and 5 and 6. Each mic had its own mic stand and clip with about 1 inch of separation. For guitar they were about 6-8 inches away from where the neck meets the body. Mandolin about 12-14" away – same with banjo (with the mics pointing where the neck meets the banjo head). With upright bass the mics were pointed directly in front of the bridge with about 6-8" from the tip of the bridge.

Any guesses as to which ones are which?

 

[Khron]

You might want to borrow (if you can) or just buy an inexpensive calibrated measurement mic, so that you can do reasonable frequency response measurements and not just relative ones.

Thanks for the tip!

Worth keeping in mind that measurement mics are mostly omnidirectional, so using that as a reference can be so-and-so, for cardioids.

 

[jp8]

Yes, but aren't cardioid mics also measured by tjeir manufacturs in setups that are calibrated with omni measurement microphones? If the mic is not too far from the speaker compared to the distance to room walls and you have a not too reverberant room, I'd expect usable results. Or am I wrong?

Jan

 

[Khron]

Yes, but aren't cardioid mics also measured by tjeir manufacturs in setups that are calibrated with omni measurement microphones? If the mic is not too far from the speaker compared to the distance to room walls and you have a not too reverberant room, I'd expect usable results. Or am I wrong?

Jan

... If you're far enough away for proximity effect to not be a thing anymore.

 

[jp8]

Sure, proximity effect is one thing. What I was thinking more about are room reflections that an omni microphone theoretically picks up equally strong in all directions, but a cardioid microphone does not. So you correct your measured cardioid curve with signals from reflections that it picks up strongly attenuated. I remember now that @MicUlli wrote a paper on how to apply a correction. It's in his Mic Measurements Part 2 paper that should be available for download in another thread.

Jan

 

[molke]

Using omni as a reference just gives unnecessary anomalies. Proximity effect has no negative effect, as both DUT and reference mic have it, so they cancel each other in measurement. Omnis as DUT are used in true anechoic settings simply because there's no reflecions, it doesn't make sense to use one if the room isn't truly anechoic.

Then aren’t you just trading one set of anomalies for another? The amount of proximity effect depends on the capsule construction, i.e., dual diaphragm has reduced proximity effect. It also depends on the incidence direction, e.g. there’s no proximity effect at 90°. I thought quasi-anechoic measurements work by windowing the impulse response before the first reflection?