Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: dc bias (Read 7890 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

dc bias

hi,it is said that when we record speech ,the signal is subject to dc bias ,(distortion) i want to know  what does this mean (what is dc bias )

dc bias

Reply #1
Code: [Select]
.                .
  .            .   .
   .          .     .
-------------------------------
    .       .         .
      .   .            .
        .               .


A sound wave basically looks like what I drew above.  It's alternating current, which passes through the 0 volts zone every once in a while.  The signal extrema (the top and bottom of the signal) have to be within the limits of what your amplifier can produce / soundcard can produce, or else you'll get clipping and it'll look like this :
Code: [Select]
..             .....
  .           .     .
    .       .         .
      .   .             .
-------------------------
        .                .

Basically, you added a constant component to the sound, making it extremely loud (this is simplified) on the positive side and not very loud on the negative side.  It's too loud on the positive side for your sound card to produce, so it clips  As such, you'll hear the sound cutting off and losing parts.  It may also break your speakers or headphones when you play it back if you don't remove the DC first.

dc bias

Reply #2
Speech is normally not symmetric around zero. When you speak air goes in and out of your mouth. The out flow is usually at a higher pressure, therefore it moves the microphone diaphragm further. This is totally normal, the way the universe works. That greater pressure on the outflow will generate more signal and displace the wave trace either above or below center, depending upon how the microphone is wired into the system. There is a general standard but either way works the same.

If the recording is done properly it is not clipped. There is no difference between recording speech and recording any other audio in this respect. The recording can be normalized to 100%, just like any other audio, and it still will not clip. There will be no problem with it unless someone gets antsy and tries something screwy to "fix" what isn't broken.

Actually, there are processes to shift the recording towards center that are often used in broadcasting where much loudness and little dynamic range is the goal. With the recording centered it is possible to apply more compression, limiting, and amplification without clipping.  These are not a part of the editors I've used; I can't point to any references as I've never dealt with such stuff.

DC offset is a matter of having a non-zero DC voltage somewhere in the recording chain. That shifts everything (not just speech) either positive or negative by that DC voltage amount. This should not happen but too often does.

In a wave trace this looks similar to the non-symmetric result of speech by itself. Generally the offset is small, well under 1% to a few %. It can interfere with various kinds of processing one might do in editing and mixing. It is generally correctable in software but doesn't absolutely have to be corrected.