Serious Inconsistency rew measurements data

[ALEXMENDEZ]

Hello to everyone,
I need help to figure out how to solve the problem that I am facing.
When only listening to music, it sounds quite good, no strange noises, good imaging, no drops in frequency, just normal, very nice I would say.
the issue comes with the measurements only

My chain of signal is the following:

windows pc (with umik1 / right usb port)
usb to douk u2 (left usb port) fixed volume 100%
optical to Wiim Ultra (pre-amp, control of level)
usb into Smsl su8 now (temporarily, before topping e50, just failed me with error code d3 today), will use br15 in a few days as it's ordered. (fixed volume, dac mode, 100%)
rca into naim nap 200 amp combined with nac152 pre-amp (knob at 25% to get 75db spl reading when wiim ultra is at 100 volume)
coaxial from sub out on wiim utra to stx 12 inch powered sub
elac unifi 2.0 as front end

i am using rew, with preferences java, with calibration file 90 degrees, all settings copied from some youtube videos.

first problem:
when i pull the spl meter and the generator, speaker cal, at first the spl reads 75db/ 76db and then suddenly starts dropping, but the noise level still the same.
yesterday i had the same settings as today, to get 75/76 db the naim clone was at 30%, today I reach that level when at 25%
i can listen to music for hours with the same level, so the amp is not the problem, I think

second problem:
lets say i take the first reading of 75db
i go to measurement, check levels, all good, output speaker L, with timing reference on speaker L, I hit start, 4 consecutive measurements (empty room, nothing is moving, dead silence)
i get 4 different readings, or 2 very similar, 2 very different, mainly in the low end, sometimes is a few dbs higher, sometimes the highs are 3b hotter, other times there is a huge null in 45hz, etc
i have taken around 60 measurements so far and hardly one is the same than the one before or after.
it does not matter, only Left speaker, only the Right, or both L+R, same problem

what can be the problem?

i have been reading that can be a clock problem between usb port on the laptop, or latency from the wiim ultra, etc

i would really appreciate some help here.
Thank you in advance

 

[ALEXMENDEZ]

sorry, the 3 files before weren't saved properly (i had deleted all the bad measurements and kept only the ones I thought they were more close to reality).
I am starting with rew and is kind of tricky. my apologies!!

the file bellow shows the problem clearly

thank you

 

[Keith_W]

These are the measurements for "DAY 3.mdat". Frequency Response and Step Response. They look about the same. There is a very slight difference in that null at about 60Hz, but that is due to poor signal-noise ratio since you are taking a 256k sweep.

If you ask me, I think you have taken a valid measurement. It could be improved a bit, e.g. try taking a longer sweep (about 1M) to improve the signal-noise ratio.

first problem:
when i pull the spl meter and the generator, speaker cal, at first the spl reads 75db/ 76db and then suddenly starts dropping, but the noise level still the same.
yesterday i had the same settings as today, to get 75/76 db the naim clone was at 30%, today I reach that level when at 25%
i can listen to music for hours with the same level, so the amp is not the problem, I think

I am not sure what you mean when you say this (in bold). Do you mean that subjectively, the pink noise sounds like it is the same SPL, but the SPL meter indicates that the SPL is dropping? Try changing your SPL meter from "FAST" to "SLOW" and see if that makes a difference.

The real problem with your room is that it is very bare. That RT60 is 750ms, and look at all those specular reflections in the ETC. In general you want the RT60 to be less than 400-500ms (depending on the size of your room, your preference, etc), some people say 250-300ms, and some go even lower (150-200ms). And you don't want any spike in the ETC to be more than -15dB to the main impulse. You need some furniture or room treatment.

 

[John Mulcahy]

Latency should not be an issue as you are using the acoustic timing reference. How is your mic supported? Any change in the position of the mic or anything near it will alter the measurement. Ideally it should be on a mic stand at a fixed position, you can't really get consistent results if it is handheld.

 

[sm52]

On the first day you measured channel L with reference L. Channel R with reference R. You can’t do that. There should always be one reference. In the second group (No Toe in New) of measurements the reference is L. In the third group (Toe in 30 cm) the reference is R. Besides this problem, I don’t see any others. If after some time the signal level drops somewhere, there is a bad contact somewhere.