Reducing music in videos

The circuit you link to uses back-to-back diodes that will limit ALL audio above a certain level - but the 'limiting' will be in the form of clamping and will severely distort the resultant output.

Since the speech is actually part of the music you can't reasonably or easily separate them - some form of digital processing may be capable of doing this but unless there is a distinct separation of music and speech then what you ask for is simply too difficult for most people to tackle. Indeed if such a facility was available then many, many people would be after it!

 

I doubt any analog solution is going to be enough. You could do a band pass filter with the assumption that human voice is mid-range but music typically has a lot of components in that range too, and certain males and females are at the opposite extremes of that range, and there's bound to be ambient noise that falls outside of midrange and if muted would make a very sterile, unimmersive presentation.

Frankly if I have to strain to understand what people are saying I don't enjoy it and just watch something else instead. At least with larger LCD TVs you have a better chance to read lips.

 

You could use the text facility on modern TVs to 'read' the lips!

 

The problem of loud music in movies is caused because the movie has many "takes" and the actors and soundman learn the dialogue and think the dialogue without even hearing it. They do not realise that the loud music covers up the dialogue.

 

There's a really great book by Oliver Sacks called Musicophilia, looks at a range of Audio type ailments. Audio hallucinations, earworms to audio processing disorders. I for one can't separate sounds very well at all and found the book useful. In terms of films I usually find versions with subtitles.

 

Subtitles work for me, too, most of the time. Listening through a pair of earphones also seems to help. There is a lot of "magic" that can be done with digital audio post-processing, and some of it is available online for consumers to play with. I am absopositively sure the NSA uses the latest and greatest stuff to gather intelligence from the immense quantity of "chatter" it continuously collects. There aren't enough trained analysts to listen to all that crap and decide what is actionable, so computers and software help filter the chaff from the wheat. Same-o same-o with digital imagery for sorting people in a crowd or identifying "changes" over time. Amazing gadgets, computers, and getting "smarter" all the time. @HarryA should look into that. Analog processing is NOT the way to go with this.

 

For some reason, British productions seem to be particularly bad in that respect. The lovely wife Morticia clued me in that if you watch the lips (when possible) it is easier to understand the speech, and she is right even though I would never say that I was doing "lip reading."

Bob

 

Just an after thought here in that it might be worth looking at the acoustics of the space you are using. I suspect it might be fine with as being at home i'm assuming with lots of soft things to dampen. I Have to avoid locations with bad acoustics, or good as some think. For me the sound bounces around everywhere and i can't distinguish anything.

 

(A): Record the programme to your PVR / VCR.
(B): Download the original music used as backing.
(C): Using a video editor or audio mixer, align then mix the original soundtrack with inverted audio.
Potential Problem: Finding exact music score used, and matching fadeins / fadeouts / studio edits.

 

Just incompetent audio engineers, trying to be impressive and impacting in their work, trashing the production.

I called/messaged three or four times to the production companies shown at the credits end of documentals/movies to spank them and about half of times has worked, with more considerate background music and audio effects results in newer productions.
If I remember, targeted programs like 'Nature' and other science documentaries.

Digital signal processing can do wonders, but when you end tweaking controls of expensive equipment, the program is already ending...

 

I had a similar problem with movies, the music always overloaded the dialog, or the dialog would whisper and get too low to understand, but I know what you mean about the Nature documentaries, drove me crazy.

What I ended up doing, since I use a desktop box to stream movies, was to get a cheap pair of Dayton B652s (like 50 bucks a pair), a little Pyle PCA1 15w a side amp, a Berringer Mini FBQ equalizer and the one thing that really made all the difference, a DBX-266xs compressor-limiter.

The compressor is about $150 US and I think the EQ is around $60 or so.

It was worth the investment for me to be able to hear everything clearly. That was three years ago, haven't had anything I couldn't hear since.

If you decide to take the plunge, here's my DBX settings:

Code:

threshold: -30
ratio:      3:1
attack:     fast-75% ccw
release:    fast-75% ccw
gain:       +10
overeasy    on
auto        off
bypass      off
stereo-couple  on

You only need to set the left channel, the right will track the left with stereo-couple on.
You can start there and mess around some more until you get the hang of it, that's just what I settled on.

Berringer makes a similar compressor, but I stuck with DBX from past experience, it's a better unit.

The EQ is to your taste, but you might want to drop some below 500htz and/or boost your highs because one of the side effects of compressors is to reduce the perceived highs (it's a psycho-acoustic thing).

 

Here's the rig, kinda jack-legged in but I can dig any dialog out from under sound track or background with it, never run into any media it couldn't handle (but I use it mostly for streaming movies).

The only thing better would be to replace the graphic EQ with a multi-band parametric EQ, but that would get way too 'fiddly' for general use.