Assistance setting up REW

[sm52]

Yes, it's weird. Upload 2 passes of both together here. .mdat file. We'll see tomorrow.

 

[HendrixStyle]

Att. TY

 

[sm52]

If we consider separately the right side and separately the left, then everything is perfect. You have very good measurements. The signal-to-noise ratio is good. Impuls is good. The waterfall is good. The spectrogram is good. 95% don't have that. But the joint work of both parties does not work out. After 2 kHz, they interfere with each other. This is according to the charts. Is it the same by ear?

 

[HendrixStyle]

Hmmm - good question. Hard to say; it sounds good in here right now, and I am not hearing anything funky nor out of balance. After reading your post, I am now trying to convince myself that the two sides interfere with each other, but I am not hearing it.

 

[sm52]

Then there is a suspicion that the measurement of the two sides together is incorrect. Check that the recording and playback devices in Windows on the 'advanced' tab are set to the same frequency as in REW.

 

[HendrixStyle]

When you say frequency, do you mean the sample rate? I have 48k all around.

 

[sm52]

Yes, sample rate. What we see on HF happens when the left side is in phase and the right side is out of phase. Or vice versa. This can be checked by ear. You need to close REW. Play tones from 1 kHz and above on the player. If the sound is always centered, the speakers are fine. Then you need to check the RME settings. Or take another sound card for testing.

I just looked at the information about the test signal. You have a Length of 128k. Put 512k. And check the ASIO buffer size in the sound card settings. Put 1024.

 

[John Mulcahy]

To measure both 8c together select that option on the Measure dialog, is that what you are doing? There look to be small differences in the relative timing of left and right when measuring both, they should keep the same relative timing as long as the mic doesn't move.

 

[HendrixStyle]

Thx again

To measure both 8c together select that option on the Measure dialog, is that what you are doing?

Yes. And I am not touching the mic; it is on a stand.

Played 1k and up in CUbase using the test gen, and the sound remeianed centered.

Nothing else to check in the RME control panel; buffer is already at 1024; I can try @ 2048. I do not have another scard. When I select '8C', I cant change the length; it's greyed and 128 is it. When I choose 'From REW', I cant choose r/l or r&l, I then wind up with one line on the graph, and the measurement says:

using estimated IR delay relative to Acoustic reference on L with no timing offset

 

[sm52]

buffer is already at 1024

no more. 1024 with a margin. Below 512 is also not necessary.

take another sound card for testing.

 

[John Mulcahy]

To measure both 8c together select that option on the Measure dialog, is that what you are doing?

Yes. And I am not touching the mic; it is on a stand.

Probably something for D&D to look at then. The speakers handle synchronisation between themselves for sweep replay, though when I did the same test on my pair the results were identical every time.