Sure.
I downloaded the file from the
@Kosorm user, organized it, and delivered it to them.
There’s nothing particularly special, but there might be parts that could be helpful for new users, so I’ll leave a record of what I did.
I opened the front-FL,FR.wav file first
(Right Channel +3.9db , Some EQ)
I noticed an imbalance in the channels, but it seemed to be due to the microphone input gain and slight wearing deviations.
To be more certain, I also checked the headphone file.
As a result, I chose post-correction rather than re-recording.
This isn’t a big deal, but it’s a fairly common case.
When recording or advising someone, I personally recommend testing with a speaker placed in the center (mono) to adjust the insertion depth for both ears and balance the input gain.
Yes, post-correction can fix everything, of course.
However, as such issues accumulate, they can become increasingly cumbersome to deal with.
(For example, what if the left speaker has lower gain, but the right input microphone has lower gain? )
Therefore, performing a few quick tests to check if the responses from both sides are similar, verifying if the resonances are properly aligned, and ensuring that the ultra-high frequencies beyond 4kHz are not lost can usually eliminate most microphone-wearing issues.
(While proper insertion depth is important, sealing is equally critical during plug-pipe measurements. If the seal is incomplete, resonances may diminish, low frequencies may be lost, and high frequencies may also be measured inaccurately.)
The basic response was not bad.
Since his room had a slight reverb, I applied a limit to the overall reverb, restricting it to around 200ms, and then I synthesized the low frequencies afterward.
Of course, the Decay parameter function in Impulcifer can stabilize things to some extent, but it cannot eliminate spatial errors.
Additionally, while some issues can be corrected with EQ, others cannot. In such cases, effectively managing bass—such as with a quad-subwoofer setup or a DBA—can be quite expensive.
Therefore, it is more practical to synthesize frequencies below approximately 60Hz to 150Hz, depending on the situation.
His speakers are decent, but I also created a file with some corrective EQ based on Spinorama data. (500~8000Hz EQ)
Generally, frequencies below 200Hz depend on the room, frequencies between 200–500Hz interact with both the speaker and the room, and frequencies above 500Hz are mostly determined by the speaker’s own response.
Of course, EQ can’t correct reflections (although time-domain EQ can be applied in BRIRs, but that’s not the point here). When applying EQ, the intended results may be mixed with the actual room-compensated sounds, and the resulting perception might vary. However, at least from my listening, it didn’t sound bad.
His impulse response was decent, but experimentally, I delayed all the reflections from 250Hz to 12kHz by 10ms after 4ms.
This approach has been explored in some research papers, where some users reported a distorted sense of distance and space compared to the original BRIR, while others found improvements.
Generally, if early reflections arrive too quickly, they negatively affect the timbre and fail to contribute to the sense of space and distance.
Therefore, this is the corrected outcome we can observe.
It’s not always beneficial to apply a lot of delay, as the limitations varied depending on the space.
Even in HRTF scenarios with distances over 2 meters, there are subtle changes, but since the ITD/ILD changes are small, applying delay alone can cause confusion in the brain.
Looking at the frequency response (FR), you can clearly see that it has become much cleaner.
https://www.head-fi.org/threads/rec...r-speaker-virtualization.890719/post-18402913
It was fine when applying the logic of Dolby Pro Logic IIx for upmixing.
However, to add more interest, I added concert hall reverb to the Side and Rear channels.
This is how it works.
Of course, applying reverb to all channels would make the sound very muffled.
However, since reverb is only applied to the upmixed, separated information, it’s somewhat different from typical experiences.
