Search the Community

The above clip includes a pink noise listening test. This test aims to demonstrate a phenomenon caused by the path difference from each speaker to your left and right ears. This causes frequencies at about 2kHz partially cancel, thus causing a dip in the frequency response of about 6dB, with the exact dip being dependent on your room. Basically, the pink noise sounds slightly duller when the listening position is at the exact centre between the speaker. I tried the test in the YouTube clip above on two completely different systems, I could clearly hear the effect on both systems, which I certainly found interesting. What seems to be of concern here, is what might be happening to the final mix in the recording studio. If the recording engineer working on the final mix happened to be using nearfield monitors, with vocals and other key musical information panned dead centre, then the recording engineer may end up boosting frequencies at around 2kHz to compensate. Alternatively, If the recording engineer working on the final mix happened to be using headphones or was further back from the monitors such that direct sound is blended with reflections from the room, then no such boost would be applied. I also believe that the above is part of the theory behind the “BBC Dip”. Maybe not so much a case of a small midrange dip making things a little more subjectively pleasing, but maybe compensating for the boost applied to recordings. Two points strike me: Firstly, if you are listening to a well-balanced system in a nicely treated room, or indeed simply listening on headphones, you may not be hearing what the recording engineer heard in the studio. If the recording engineer was listening on nearfield monitors producing a 6dB dip in the sound, then the perfect mix to his ears would need 6dB boost at 2kHz to sound balanced. Then someone listening to the recording on headphones would then simply hear the unnecessary 6dB boost. Hear we dip into the subjective. I can think of tracks that that to my subjective ears sound too prominent in the mid-range, and to my subjective ears sound more balanced with a DSP cut at 2kHz. It is of course easy to make the assumption that the original mix has an inadvertent 2kHz “boost”, and by my subjective 2kHz cut, I am actually making what I hear accurate to what was “heard in the studio”. Therefore, my 2kHz DSP cut is making the sound more accurate to the recording. But this is of course just a guess, maybe the track was mixed on headphones or the recording engineer was further back from the monitors such that direct sound is blended with reflections from the room, and my 2kHz cut is noting more than a subjective preference. Without full details of exactly what happened during the final mix, we will never know. As an example of the above, earlier in the week I was listening to Radiohead’s Airbag on headphones. To my ears this definitely sounds more balanced with a modest cut at 2kHz. (although something like -3dB does it for me, - 6dB drifts into dulling the track) I am not saying that Airbag is the ultimate track for demonstrating this, it just happens to be an example I listened to recently. The second point is what all this might mean for those using digital room correction. If one were to have a system that exhibits a -6dB 2kHz cut at the listening position, then if you room measurement microphone was so located when taking room measurements, then the resulting room correction / convolutions would add a 6dB boost to correct. This would actually correct for the partially cancelled 2kHz at the listening position, but the listening off-centre, you would actually experience the boost. When creating measurements for room correction software, very often the recommendation is to take an array of measurements, both at and around the listening position. When using this approach, some measurements will be subject to the 2kHz cancellation, whilst others will be outside the zone that is affected. So I am guessing that this typical approach will generate an averaging of readings, so any cancellation effect at 2kHz will be partially mitigated. It does mean though that if you are listening at the dead centre position, you might then be experiencing a small cut at around 2kHz, due to the measurements takes outside of the “zone”. @mitchco - Any thoughts on this? I did think of posting this in the “Objective-Fi” forum, as this is an easily measurable phenomenon, but I then considered that there is likely a lot of subjective preference around all of this, with frequency cuts at 2kHz providing a more likable or listenable sound, or a boost giving the impression of more detail. So considerations of all the above my be subject to individual preferences. Give the clip a try on you system if you can, I would be interested in the thoughts of others.