Normalize audio inputs

Thread Starter

gkhnoisgtht

Joined Jan 26, 2011
1
I am creating a project where I am taking A/V inputs from a DVD player, laptop, insert other a/v device and outputting an analog signal (HDMI in the future) where the audio is at a constant level between input devices.

I figure that this would simply consist of some sort of volume control that bases its volume off of the output from a peak detection circuit with two voltage match circuits (one for over the threshold, one for under) to keep the peak volume even at 100db. I would like to keep everything in analog so its faster and wouldn't rely on programming but its not a requirement.

Of course I could also use a DSP or microcontroller but i'm not sure if the delay would offset the sound to the point where the lips are not syncing correctly.

Does anyone have any suggestions on how to complete such a circuit? I would prefer to use TI parts, but i'm up for just about anything to get me started.
 

Audioguru

Joined Dec 20, 2007
11,248
Diodes used as limiters cause bad distortion to audio.
You need to use a voltage-controlled-amplifier that has its gain controlled by rectifying and filtering the peak or average level of the audio.
 

bribri

Joined Feb 20, 2011
143
Diodes used as limiters cause bad distortion to audio.
You need to use a voltage-controlled-amplifier that has its gain controlled by rectifying and filtering the peak or average level of the audio.
...okay in that case you could control the gain with a diode.. or rather a led.. or rather a vactrol type thing.
 

Audioguru

Joined Dec 20, 2007
11,248
...okay in that case you could control the gain with a diode.. or rather a led.. or rather a vactrol type thing.
A sine-wave generator also needs a device to stabilize the amplitude. Years ago a light bulb was used because its resistance changes with signal level through it but its response was very slow resulting in the amplitude bouncing.
Then an optical isolator using a light bulb or LED plus a light dependant resistor was used but it is also slow to respond.
Then an optically isolated Jfet was used but it caused distortion.
You need a voltage -controlled amplifier (VCA).
 

bloguetronica

Joined Apr 27, 2007
1,541
The LM series has a few interesting companders that might be of interest to you. I recall having references to those ICs in a book, but for the moment I don't have the book or the references.
 

Adjuster

Joined Dec 26, 2010
2,148
Bear in mind that, even if the harmonic distortion performance is acceptable, by its very nature any compressor system modifies the dynamics of the signal. This is in itself can be considered to be a form of distortion.

If applied to an excessive degree, the effect of compression compression can be quite objectionable - for instance during quiet passages during a film, quiet sound effects like rustling leaves may be amplified to the same level as the loudest shout. Sometimes a long response time is employed to minimise this effect, but that can give rise to disturbing variations of signal levels.

It may be better to obtain the best possible level match between the different program sources before compression, so as to minimise the degree of compression required
 

Audioguru

Joined Dec 20, 2007
11,248
Signetics and Philips made the NE570, NE571 and NE572 compandor ICs many years ago.
Newark still have some NE570 ICs in a surface-mount package made by ON Semi.
 

bribri

Joined Feb 20, 2011
143
Then an optical isolator using a light bulb or LED plus a light dependant resistor was used but it is also slow to respond.
my portable recorder employs an opto-isolator based limiter. it's fan-friggin-tastic at doing what it does and sounds great.. i'm pretty sure that slower response times are part of the reason for that.

It may be better to obtain the best possible level match between the different program sources before compression, so as to minimise the degree of compression required
absolutely.
if it's just two 'program sources' then maybe just having a limiter for the sake of aesthetic safety. it really depends on your what you're using it for... perhaps harsh compression thumping is to your liking.
 

Audioguru

Joined Dec 20, 2007
11,248
Slow compression or limiting of audio results in the first sounds or words blasting loud and probably severely distorted then a noticeable drop in level, or very low level first sounds or words followed by a noticeable increase in volume.

My favourite TV news station has blasting severely distorted first sounds or words followed by a noticeable drop in volume. A faster compressor or limiter will sound much better.
 

Adjuster

Joined Dec 26, 2010
2,148
It's not uncommon to hear fast gain reduction (presumably to minimise clipping), followed by a much slower recovery of gain during quiet periods.

This may avoid blasting, but still sounds pretty awful, in my opinion. In addition, this system will naturally go "deaf" for a while whenever there is any loud sound, so it is badly affected by any sudden background noises.

The sound-track on some video clips seems to show this effect - I wonder if this is usual on the small cameras which have become so common recently?
 

bribri

Joined Feb 20, 2011
143
Slow compression or limiting of audio results in the first sounds or words blasting loud and probably severely distorted then a noticeable drop in level, or very low level first sounds or words followed by a noticeable increase in volume.
granted, but there is attack and decay to consider. optos can be plenty fast enough for audio applications. maybe it's the response curve rather than sluggishness which people like.
 

bribri

Joined Feb 20, 2011
143
The sound-track on some video clips seems to show this effect - I wonder if this is usual on the small cameras which have become so common recently?
video camera AGC is just really horrid in my experience. i haven't really had a look at newer cameras so much... but i'd wager that (data)compressed audio such as AC3, which seems to be in vogue with newer cameras, isn't really helping matters. but then again, i had a mini-disc recorder which just made way better recordings (data compressed) than any miniDV camcorders (uncompressed linear PCM) ever could hope to achieve.

but who knows...maybe the point of this thread is to find a way to reduce the perceived loudness of commercials in relation to other programs.
 
Top