Good news! Keep us posted. The increase in dynamics comes very likely from the phase correction, the better the time alignment, the more impacts are “coherent” across the spectrum, the higher the levels.
The sound stage benefits a lot from the reverberant field correction (which helps your brain to "disconnect’ from the unnatural tones of your room), and in some cases from the x-talk reduction, depending on the type of recording. I had a funny experience this week-end listening to Bill Evans Trio / Live at Village Vanguard, all filters activated : in between songs I could hear very clearly the waiter serving drinks at the very far left… about 5 meters beyond the walls of my apartment, as he was floating in the air (I live at the 9th floor…).
For the intellectually curious like me : here are for illustration the HAF filters for my left channel. It’s easy to import them in REW : File / Import Impulse Response. For a given speaker (left speaker in this case), the first channel is for amplitude and phase correction, the second channel is for the reverb tuning (+ X Talk in option).
The first graph shows the first channel of the filter : amplitude correction in dark red, phase correction in light red.
Please note that what we see here is the filter slope: a low in the filter corrects a peak in room/speaker behaviour, a high in the filter is here to boost some frequencies. You can see I have a very strong room mode at 55Hz, and a peak in my tweeter around 6-7kHz, and many other things happening in my room (yours is probably not any better if it’s a family living room…). The phase correction is not as tortuous as it could be for others because my speakers are conceived to be well time aligned thanks to a specific time delay filter on the tweeter side. Don’t mind the crazy behaviour of the phase around 55Hz and 3.2Khz, these are areas that are non minimum phase and can’t be corrected.
You would get similar filters using Accurate, REW/Rephase or Dirac (except that with Dirac you can’t access the phase correction and cannot import in REW as one need their `ù$^àç audio driver, sorry for cursing).
The second graph shows the “reverberant field” correction : there is a clear boost between 1 and 6Khz, with a plateau between 2 and 5, which is the crossover area (quite large as my filters are 12dB/oct only). What is remarkable is that these numbers matches with what the theory of HAF predicts : maximum correction around the crossover, with typical mid woofers and tweeters. And Thierry from HAF had no clue what was my crossover upfront! All the info was computed from the initial measurements only.
