Predicted v Measured in room response

[dweeeeb2]

Hi, Ive been looking at spinorama data / speaker reviews and Ive been putting a bit of faith in the "predicted in room response" curve. I understand that its just an average of possible rooms etc and its a best guess, but I was watching a vid on the genelec 8351B (above my price range unfortunately) and I saw that their measured in room response was quite different to the predicted response by Amir. I was just wondering of this difference is normal and expected or is it an outlier?
(sorry about the formatting, im on my phone)
EDIT: speaker vid (sales vid):

 

[abdo123]

That’s a pretty good in-room response. It’s shocking that not a single peak reaches +10dB.

 

[Scgorg]

The estimated in-room response assumes both a certain level of reflectivity, and a relatively far listening distance. Additionally it is usually only accurate-ish above, say, 300-1000hz depending on your room size, this is because it does not account for room modes or other forms of strong destructive/constructive interference (such as reflections from a mixing desk). If your room is not very reflective, or if your listening distance is relatively close (say, 2 meters or closer) it becomes less accurate, and the actual in-room response will look closer to the anechoic frequency response of the loudspeaker.

 

[Deleted member 48726]

Well if all rooms were identical..
I've wondered about what to use that exact info for in the reviews.

If one wants to investigate room impact REW has a quite nice room simulator where reflections can be factored among other things.

 

[Eetu]

The Predicted In-Room Response is for far-field and that's why it includes 44% Sound Power, 44% Early Reflections and only 12% Listening Window.

But when you are listening near-field you're mostly hearing direct sound so Listening Window is way more important.

Add some below-the-Schroeder messiness, comb-filtering and the measurements in the video start to make total sense.

 

[staticV3]

I saw that their measured in room response was quite different to the predicted response by Amir. I was just wondering of this difference is normal and expected or is it an outlier?

This is a Far Field prediction:


This is a Near Field measurement:

see here:

that's why it does not match the prediction.

However, later in the video he does a Far Field measurement as well:
-->

And if you overlay this Far Field measurement with Amir's Far Field prediction, you get:

a perfect match
(sans room modes)

 

[dweeeeb2]

Thanks guys, very helpful as Im looking at playing with near v far distance listening in my current setup, thanks for clarifying.