Measurement Setups - Post Your Rig!

Thanks to @Biodegraded for doing the work! Geoff explained how and I think I could do it in the future so thanks for the lesson as well.

Biggest discrepency between fpc and EARS is at 5kHz. Ill get some more headphones on fpc for comparison to others.

 

I used the CSD settings you shared here @Biodegraded .
I had to lower the bottom to 40dB so we could see the long decays. What do you think?

^ BP800. The HD800 driver is supposed to have pretty defined ringing at 4.5khz but it doesnt show that well in my graph. Is there an adjustment I could make to make the ringing there more clear without messing up the graph's usefulness as a whole?

Porta Pro measurements later today or tomorrow.

 

Here's a bit of added context, after some PM discussion with Philimon ...

The upper pair of curves are his compensated EARs and flat-plate measurements as in the plot in post #218. Of the lower set of curves, red is the cal for his umik, blue is a B&K -1dB-per-octave after ~200 Hz, purple is the combination of those two - which was used to compensate the purple flat-plate measurement - and tan is @purr1n 's comp curve for his own flat plate (specified here).

Note the spike in the tan curve centred on 6k. It appears that Phil needs a similar spike in his purple curve. Should check with a few more headphones, though.



@Philimon , your CSD floor in the post above is ~50 dB down from the 5-6k peak in your FR. Experienced ears suggest that bad audible effects will generally show up less than 35 dB down (here's the thread).

 

Spoiler: BP800 Distortion

Spoiler: BP800 Distortion (with Noise Floor)

^ Look at that noise floor! Prob the answers to how devastating that might be to csd + distortion measurements can be found in link that you shared above @Biodegraded . Ill read now.

edit: The SPLs used are too low for distortion measurements. Going to read the REW instruction manual next...
edit: 317pages. Not going to read whole thing. Ill read most pertinent sections. Then look for some tips on the forums.

 

For measurements with fpc should I aim for 100db? That's waaay louder than I use for EARS. Even 90db seemed extreme, like going to damage headphone drivers. Maybe that's just me since I'm a low volume listener (headphones and speakers).

I took some measurements just now with Porta Pro and Emu walnut (modded) at 100db. Coming up. External noise wasn't very quiet (ps4 fan and vehicle noise outside) for distortion and csd but ill post them anyway just for practice. Still figuring out axis and settings for graphs.

 

^ fpc using the newest compensation. EARS using SBAF comp


^ Distortion settings need work. I'll look around at others plots... Feel free to advise. Noise floor is much better with volume set for 100db.
FYI I use the Emu with Loki EQ when listening, but measures were without. Reminder these are modded.

Edit: @Hands measurements of Emu-w. @purr1n measurements of Emu-p

 

Other's Porta Pro measurements can be found in this thread and in this thread for comparison. Frequency response and CSD looking pretty good. 100db too much for distortion measurements?

edit: Noise floor was higher than normal here, sorry. PS4 fan blowing and vehicle engine noise outside. I can redo if interest.

 

Apologies again, external noise floor was higher than usual. Can redo if needed.

 

Marv quote, regarding spl for measurements.:

Input gain I have no idea.
Macbook Pro > Schiit Eitr > ModiMBv2 > Vali2+
Mic is Umik-1

 

For your distortion measurements, rather than showing the noise floor, choose 'Mask harmonics below noise floor' in the Controls menu. Different harmonics are significant at different levels below the noise floor, and this latter setting greys them out (while still leaving them visible) at the appropriate places.

For your CSDs, set the floors and scales 35-40 dB below the avg SPL or interesting peaks in the 2k-10k range. Can't quite see through the control panel, but the BP800 immediately above looks better than the PortaPro and Emu ones where the peaks are up around 100 and the floors are down at 55 (45 dB diff so not surprising the ridges come all the way to the front).

Another PortaPro measurement here. Not the best calibration but yep, 5k peak

 

Thanks much @Biodegraded . I'll reformat then post again.

That's why when I see headphones with 5kHz peaks or ringing I dont worry. My ears have been toughened up by over a decade of Porta Pro (and Grado).

 

^ "Plot normalized to fundamental". The noise floor might be making this a bad measurement of these headphones. I'll take more quiet environment measurements next time. Depending I could use some weak isolation chamber (some box).

"Plot normalized to fundamental" I find easier to read... Just know less than 1 is good! But I guess that depends on reference spl. Less than 1 at 100db spl is good!(?)

I see @purr1n and @Vtory demonstrate distortion at various spls. I think I will try and copy this because the differences are stark. A low volume vs high volume listener could have very different experiences. Some examples of their approaches to display distortion.:
https://www.superbestaudiofriends.o...-analysis-and-measurements.11680/#post-366689
https://www.superbestaudiofriends.o...2x-review-and-measurements.11872/#post-371474



^ CSD I see is going to need a little more best judgement. The frequency response of BP800 is throwing off where I should set the limits on the graph. The treble is so depressed, but I don't actually use these headphones without EQ so these whole distortion and csd measurements arent really representative of how I'm really using these anyway... I'll redo later with quieter background. Anyways, I chose for this graph to lower the limits to try and demonstrate the audible treble ringing (at 4.5kHz). Compare to @Bill-P csd measurements. FYI our mods are not the same. I'd say my mods damp the driver more due to mine having a non-porous baffle, while his is porous. But I don't know for sure nor how the mod differences would equate to our CSD looking so different.:

^ Bill-P BP800


^ CSD of BP800 with EARS (definitely not at 100db). This does reveal the 4.5kHz ringing (I could have set limits better here for longer decay trail) as observed as well by Bill-P. So what's causing my FPC CSD not to look like my EARS or Bill-P's fpc CSD? Too high noise floor? Or it could be I need to reduce the 5kHz blip in my Umik-1 compensation because its masking the 4.5kHz ringing? I will redo measures later, need more data. And I will try a set of measures with EQ applied.

edit: forgot to de-click show noise floor in distortion measurements. Next time

 

- Noise floor is real.
- Still cant get 4.5kHz ringing to show. Maybe my fpc compensation curve needs to lower its 5-6kHz peak?
- My regular Loki+ EQ settings applied: 400Hz knob at 7o'cock, 8kHz knob at 9o'clock

 

- I tried axis limits that might help show 4.5kHz but no luck.
- I now see I don't need 90db vs 100db CSD since they dont seem to change like distortion.

 

- Feel my CSD could still use some polishing. Same settings as before except maybe a little axis limit changes.
- Loki+: 8k@9
- Should I use smoothing with distortion measurements?



^ Horizontal ridges indicate reverb (?). Emu is closed cup. Tried quickly reorienting CSD viewpoint to overhead but could not figure out. Need to research.

 

@Biodegraded pmed me some tips. TY
- I don't need to make my CSD that deep. Better to show the 0ms / top of curve. If I need to go that deep to show ringing then its probably not that audible anyway.
- The 4.5kHz ringing is short lived, perhaps thats why its not so obvious in my measurements. Also, mod and rig differences between @Bill-P and I.
- I need more headphones for reference. Ill sign up for some loaner tour headphones that have been measured by purr1n.

 

Okay. How about this? And do I select "Plot normalized to fundamental"? And I like distortion in percentage but others like Bio prefer other.



*Same measurements with Loki+ eq from last post.

 

Here I didnt select "plot normalized to fundamental". This looks nicer.

Hands Emu-w measurements for comparison.

 

BP800 without "plot normalized to fundamental". Nice


^ How is 3rd and 4th harmonic better in 100db measurements? What did I do wrong?

 

Maybe a bit off topic, but does anyone have a miniDSP EARS they'd be willing to loan or sell? I'm having a hard time finding one second hand, and I really could use one to help make a EQ profile for a pair of ethers CXs I bought with some channel imbalance issues.