Navjyot, interesting questions and I especially like the effort you’ve taken to search and post various renditions of the same song over the years and with different settings to illustrate.
I see this change as a combined effect of multiple phenomena. Let me list them separately.
1. Firstly, the difference between the two Lata renditions is easily attributable to one being a film play back and the other live performance with an orchestra. Live performances will usually have more/heavier instrumentation as it’s played in an auditorium or large ground where vocals would struggle to have the same impact as in an intimate listening session to a vinyl/CD/even radio in one’s room. Also the fact that it’s an orchestra - the musicians are also performers here and in most cases adding something to the original song means they deserve a fair share of the sound. You’d also see that these orchestras have many pieces of each instrument unlike in the original song (with some exceptions) thereby increasing the body of the instrument in the live performance - not unlike in a western classical orchestra (which doesn’t use amplification most times) - it’s in the DNA of orchestras.
2. The evolution of Hindi film music over the decades. I’d say there were two waves - first was the RD Burman wave - where innovative instrumentation started sharing almost equal space with the singer in the playback. No longer were the instruments just backing - but you had guitar or sax solos in prolonged openings or interludes... and then multi-instrumental harmony bringing in western idioms to Hindi film music.
This went on from late 70s till the mid-90s when the second wave - the AR Rahman wave took over. Rahman’s songs if you notice, have a limited tune which is repeated over and over again. Surely it’s an excellent and original tune, but is short to carry an entire 5 mins song on its own. Rahman, also with his western classical/orchestra initiation (just like Illaiyaraja before him in the South), added entire orchestra to his songs. This was liked by the audience and over the next decade (the noughts) became de rigueur as directors like SEL, Amit Trivedi et al followed it.
3. The downfall of lyrics in Hindi film music. This started in the 70s and continues unabated till now. One can argue if it’s due to changing taste of the listeners or paucity of good lyricists, or both. But the fact remains that not too many lyrics written today are strong enough to shoulder the song entirely/largely on vocals. A higher vocals focus would need quality lyrics - like for example in country/folk music in the west (in contrast to say pop or heavy rock).
4. Quality of singers. The videos you’ve shared is a testimony enough on this. While Shreya is considered better (more tuneful) among the current lot of singers, one can easily see the wide chasm between her and the older greats such as Lata, Asha, Geeta Dutt etc - Shreya’s voice always feels more strained in comparison. And when you further go onto Akriti, the point gets absolutely clear. Also the world of auto-tuning, anybody can sing. You can make a singer out of almost anyone. But can they carry a song purely on their singing? Highly unlikely.
5. While I see the above four as the prominent reasons, this one might also have contributed along the way. The improvement in recording technology (as well as electronic production of music) over the years mean that you could have more content in the especially the lower frequency range than before. No wonder there’s so much bass in modern film songs - you can listen to most of it through even streamed digital music with decent set of earphones. Vinyl would struggle to go that deep and yet resolve. And it’s easier to put repetitive bass beats (electronically produced at that) to make a song catchy. If you see, a large percentage of songs today are what can be played in a dance party. This is also a social change - back in the 60s you’d have had a social scene where the songs were played in a club - and vocals and/or acoustic music suited that setting better. Today you’d need a lot of low end for the DJ situations (pubs, discotheques) that these songs get played in.
Having said all this, there’s always a counter/retro trend in every mega trend. So you also have acoustically backed largely vocal music like that of Prateek Kuhad also growing an audience outside the film scene - with some crossover into films also happening.
Let’s come to this question of yours now. The only two ways to reduce the instrumental music (amplitude wise) without reducing the vocals that I know:
1. Select equipment, especially speakers that are mids-specific. Avoid floor standers and subwoofers - go for bookshelves known for their mid-range quality. Best option? Full range drivers - but they’d be too costly... find the next best in two driver speakers v auditioning.
2. Use equalisers - whether software or hardware. You can then control the amplitudes of the frequency ranges.
Both these will reduce the instrument music in the higher (treble) and lower (bass) ranges, but not have much effect on the instruments in the mids (which are closer to the frequency of human voice) such as flute, brass, sarangi or violin.
Personally I am not a fan of the second option because you are modifying the music too directly. Also it’s a headache to do so from recording to recording. Option 1 works better if you know what kind of sound signature you like and which speakers reproduce it the closest to your liking.
I like this socio-economic perspective provided by
@alpha1 on the subject though I don’t agree with it. It’s the radio (following recorded music itself) that took music to the masses and radios have been playing in every small/big house since the 50’s and 60’s. In fact, more households listened to music on radio and cassette players back then than they do on BT speakers today. But that didn’t lead to loss of melody.