71 dB
Headphoneus Supremus
1. The things that have superseded crossfeed are rare. Even less people use them than crossfeed and you do hold the lack of commersial success of crossfeed against it. That's because they are expensive and technically demanding. Also there are also scientific justifications against crossfeed against headphone sound as it is, the original reason I discovered crossfeed.
2. My spatial hearing expects some sort of resemblance of natural word to be fooled in a way that makes sense. If the spatial parameters don't make sense, the whole concept of stereophony collapses. I believe the assumption of natural world is inherently baked into stereo sound. Stereo works because of how our spatial hearing works to process sounds from natural world. The same principle applies to 3D movies. The picture for left and right eye need to follow the principles of stereo vision or the concept collapses. With speakers we are dealing with natural world, because we are hearing two speakers playing in a room. It's natural to our ears even if the sounds are "unnatural test signals", because the room acoustics render the spatial cues the same as if it was a person speaking or a animal making sounds where the speaker is. That's the reason why speaker stereophony works. It appears natural no matter how crazy the spatiality on the recording is and our spatial hearing is easily fooled to hear what sound engineers/artists intented. Headphones remove this naturalization process and things go south at least for me, but with crossfeed I can have a coarse simulation of one aspect of the natural processes and my spatial hearing is fooled enough. I think I am lucky my perception allows something as simple as crossfeed to deal with the problem of headphones spatiality. That said, we are again walking on a minefield of semantic interpretations. What is natural world and what isn't? Where is the line?
3. I don't think the problems of too large ILD at low frequencies is cherry picking or half-truth. 9/10-truth maybe. I don't understand why the intention of sound engineers and artists would be large ILD at low frequencies. My intention as a sound engineer would be to produce recordings that sound as good as possible and that is achieved to my knowledge doing completely different things than creating large ILD at low frequency. That's filtering tracks carefully possibly filtering out unwanted resonances, compressing tracks, adding warmth with tape/tube saturation plugings, creating the sense of depth using reverberation differently to track, using side-chain compressing, using glue compressor to glue tracks together into something coherent and so on whatever it is the track under work requires. That's the stuff I encounter when I watch Youtube videos about mixing. Never has any sound engineer in these videos said having a large ILD at low frequences is a positive thing. No, often it is recommended to mix bass MONO! I think that's unnecessorily simplistic and personally my target ILD at low frequencies is 0-3 dB, the range we see in HRTF-measurements. To me that makes scientific sense and is a rational intention in music production. Amplitude panoration at low frequencies is a stereophonic brain fart. Spatial hearing in primarily based on ITD at low frequencies, not ILD so much. You want bass sound for some reason on the left and right in your mix? ILD = 3 dB, ITD = 600-700 microsecond. That's how it's done, how you do omnistereophony that works for both speakers and headphones. On headphones I perceive sounds like this on one ear only even if the sound on the other ear is just 3 dB quieter! The ear where the sound arrives first or is louder "masks" the other ear so that it's as if the other ear hears nothing, but it's all natural. If ILD is larger, say 10 dB, some bass energy is "missing" on the contralateral ear and the masking process overcompensates and I think I actually hear the difference as an out of phase version of the sound. So, if the ILD at low frequencies is too large, I hear "virtual" out of phase sound sources that are modulated in and out of existence by the ILD level! Now you perhaps understand why I experience large ILD at low frequencies distracting and annoying as hell. When the ILD at low frequencies is small enough (a few decibels only), my perception thinks there's nothing missing on the contralateral ear and no virtual out of phase sounds pops out. Everything just sounds natural and pleasing. It's important to understand how the importance and function of spatial parameters vary with frequency:
20-800 Hz: Spatial information is strongly encoded into ITD and weakly encoded into ILD (assumed small) and ISD (assumed small).
800-1600 Hz: Spatial information is encoded into ITD, ILD and ISD, but in a messy way so that spatial hearing doesn't rely much this transition octave.
1600-20000 Hz: Spatial information is strongly encoded into ILD and ISD and weakly encoded into ITD.
Classic methods of creating spatial information to recordings such as amplitude panoration take this all into account poorly so that with speakers it works thanks to the naturalization process of acoustic crossfeed, ER and reverberation, but on headphones it just falls apart, at least for me. Of course these days things are much more sophisticated, but it doesn't help with the older stuff. It is what it is and even the new stuff isn't always perfect, just better…
4. What is this model you are talking about? HRTF convolution? Everything the science said about crossfeed in the 70's is still valid. It never said there can't be anything better in the future. It said crossfeed is an improvement in headphone listening and that's still true today, at least for me. The fact that even BIGGER improvements now exist doesn't nullify anything.
