Recording Sound for Internet Distribution

You can't fix it in post-production

Sound is all there is

Wow, that's a big sweeping (honking) statement. Turn off the sound on most movies and see what you get out of it. Typically not much. The single most important element of most movies is sound. Most of the "information" passed to the audience is the result of dialogue. No matter how much you emphasize the visual element, even if you have Spielberg as your director, nothing changes. (Exceptions: Some movies intentionally use little to no dialogue.) So, this is where you grab your sound guy by the ear and say, "I know you know everything there is to know about sound, but Internet distribution is different," as witnessed by a host of movies that have nearly unintelligible dialogue when distributed over the Internet.

There is a reason why you can't fix sound in post-production... except at more expense than you pay the sound crew. When one voice is low, the difficulty is, the background noise is already mixed in, so you can't raise the gain on one voice without raising the other, which solves nothing.

Another problem is the acoustics of the environment is difficult to duplicate. Every room or aread reverberates differently. A good sound guy will record several minutes of background sound so if you need to dub lines later, you can dub the errant line over the background... but that doesn't create the reverberation of voices. To fix that problem requires dubbing all voices in the scene.

If you try to fix one actor's lines with dubbing, it doesn't work. The two voices sound like they are in two different places. So you have to bring both actors in and dub all of their lines for the entire scene. Well, sometimes you just can't get the actor's lines in a noisy environment and have to dub the entire scene anyway. But dubbing is expensive - why set yourself up for that expense if it isn't needed?

There is another reason why sound doesn't get fixed for Internet distribution in post-production. Most post-production work on sound emphasizes the theater experience. In a theater, the sound is played at a very high level, so people can usually pick out the dialogue from the background sound and music. Few movies on Internet distribution are shown in theaters. The majority are shown in homes or on mobile devices, on 20 to 30 watt speakers, or 3 watt speakers, or less, and the sound environment is very different. If people played the sound at theater levels, they would be arrested for disturbing the peace. But the quest in Internet video is becoming a drive for best sound. Streaming Media Extra article: Audio Catches Up: Online Video Rocks with Multichannel Audio.

The point here isn't to present some tutorial on sound - leave that to the sound guys who really already know this stuff. The point is, the things that you might have gotten away with in theater, don't perform that well in other environments. Sound for Internet distribution is more exacting. It has to have the versatility to perform well from devices with poor sound - mobile devices to home theater... and even theater. For more information on controlling sound, a BSS Audio equipment manual gives a good explanation: DPR-402 (Dual compressor, de-esser, limiter).

One simple technique is to boost the bass and suppress the treble, for small devices with small speakers. Another technique is to use surround sound techniques.

Journal of Sonic Studies article, EDITORIAL: Rethinking Theories of Television Sound

Surround Techniques

The strange human ear

Just a little explanation will help. Most 20 something sound guys will hear the sound they are monitoring very well. Well, it isn't that simple.

The human ear passes information to the brain at a wide range of volume levels. We don't really notice big changes in volume. It actually takes twice the power (3db change) for the human ear to even notice a difference. So we hear crickets chirping in the woods outside the window, a frog adding to the serenade, the clock ticking on the wall, and the child screaming in our ear. We hear these just fine and are able to differentiate between them. We may even know that the frog is not a cricket while the kid is screaming... but probably won't care.

One major difficulty is age and noise damage. The ear changes as people age. They lose the ability to hear higher frequencies, and voice frequencies. People in their sixties, or those who have listened to a lot of loud music, often lose the ability to differentiate a speaking voice from the background sound, which means they can't understand what people say if other sound is present, often even with a hearing aid. This makes the inherent problems of sound even worse. Over the next twenty years, each age in our population will be equally represented. There will be as many twenty-year olds as there are 80 year olds. People who can't understand what actors are saying simply won't watch - that's a large portion of the potential audience.

On TV, with all sound blaring from the speakers, the way it is edited for theaters, means that we hear all sound at around the same level, competing for our ear's attention, and we miss a lot of detail. For example, morning TV news programs often have music in the background as the hosts speak. You really can't decipher what the host is saying because the engineers have the music level too high and it drowns out the speaker.

Next page: Best practices for recording sound on set.

  1  Page go right 2

Or click a Page selection on the top menu.

Best practices for recording sound on set

Following are some recording and editing practices, for Internet distribution, that will serve you well, even for theatrical distribution and TV:

  • Know your actors voices, and coach your actors. Many actors vary the volume of their voices considerably. One moment they may be shouting in a piercing stage voice that can be heard in the next town. The next moment they may be speaking in hushed tones or whispering... and not a stage whisper, an intimate pillow talk whisper intended for their co-star only... which you can often hear in a theater, but not on the TV or mobile. The sound guy's job is to capture that all important sound. Get the microphone as close as you can to the source - if necessary hide it in the pillows, use a directional mic (which may cause fluctuations in background sound), lapel mics, a shielded mic, or push the boom mic into his... armpit.
    The most difficult problem is two voices in noisy environments, when the co-star is talking in a normal voice, and the star is whispering. Riding the gain control to boost volume when the low speaker is talking won't work - it raises the noise level at the same time as the voice. It sounds like the actors are in different rooms. The actor has to speak up.
  • Loud music, when there is no voice, is not a problem.
  • Voice level should be recorded at -12db (especially if not using clipping). Using -12db allows for normal fluctuations so that volume doesn't rise above 0db, unless there are very loud noises on set like a train going by.
  • Music should be recorded at -6 to -9db below voice, (meaning at -18 to -21db) so that voice is always intelligible above background noise and music. If not using -12db as your reference, such as using -20db, just make sure other sounds are -6 to -9db below voice. When there is no voice competing for attention, music volume can rise to -12db, although fluctuations between dialogue would be weird. Most Internet enabled devices will have the best volume control if the movie was recorded at -12db, not -20db.
  • Constant level devices should be avoided. They will usually make background noise flood the audio when there is no voice, and background noise will vary between actors. Voice and other sounds should never be recorded above 0db. It causes sound distortion in analog, or sound that is too loud in digital plus distortion.

Next page: The encoding format.





1 Page go left  2  Page go right 3

Or click a Page selection on the top menu.

The encoding format

Encoders are used to place your movie video and audio into a compressed format for distribution. Compression, including MPEG, is almost always "lossy," meaning that some information is discarded because it is not essential enough to the end product. The higher the compression, the more lossy it is, and the more noticeable it is.

Codecs vary considerably as to the sound encoding that is used. Inexpensive codecs usually do not offer quality sound because it means paying a license fee in the product. That isn't to say that the sound isn't good... it just isn't the best sound.

It is impossible to know what is wrong with the sound quality on some movies. All you can tell is that much of the dialogue is unintelligible, and the likely conclusion is that either the sound guy did a horrible job (probably not), or they messed it up in post production (nah), or most likely they used an encoder with a poor quality sound encoder.

The Internet market, which delivers movies to any Internet enabled device - gamebox, TV, mobile phone, etc. - is beginning to emphasize the quality of the sound. This means that encoders will likely get a little more expensive to accomplish this.

The H.264 format, which is becoming the Internet standard, has a leg up on this need. The H.264 encoder uses Dolby Digital AC-3 sound. This is very good sound quality. It will be up to VC-1, the older but very good video format used by Netflix and Windows Media Encoder, and WM, the new open source (no license fee) video format being developed by Google, and used to the exclusion of other formats in Google Chrome, to match the quality of H.264.

To be distributed on FlixStreamer, the encoding must be in H.264.

See also: How to Get Perfect Audio in Adobe Premiere and Audition

Have you produced a good independent movie or short? Get more attention for your shorts on Rent your movie on for more profit.

- Dorian

End - or Blog on next page.

2 Page go left  3  Page go right 4

Or click a Page selection on the top menu.

Blog - What do you think?

4 Page go left  5  Page go right 6

Or click a Page selection on the top menu.


Companies mentioned, or commonly mentioned on this Web site:

NETFLIX® is a Registered Trademark of NETFLIX®, INC.
BLOCKBUSTER name, design and related marks are trademarks of Blockbuster L.L.C.
Redbox® is a registered trademark of redbox Automated Retail, LLC.
YouTube® (YOUTUBE®) are registered trademarks of Google Inc.
Disney is a business name of Disney Enterprises, Inc.
HULU® is Registered Trademark of HULU®, LLC.
AOL® is a Registered Trademark of AOL Inc.® is a Registered Trademark of, Inc.
Barnes & Noble® is a Registered Trademark of Barnes & Noble, Inc.
Any trademark not listed out of oversight is a Trademark or Registered Trademark of it's respective owner.

Mention of any business or movie in this article is not intended to endorse, disparage, or favor any business.

Movie names that are mentioned are not given reference citations. This is because numerous studios are involved in production, and they then assign distribution rights to multiple distributors, and these rights can be sold to other distributors. For production and distribution information on any movie mentioned, consult the Internet Movie Database, or other authoritative listing.

5 Page go left  6  Page go right 7

Or click a Page selection on the top menu.

6 Page go left  7  

Or click a Page selection on the top menu.

In this article: Webutation

MSP Insider Articles

  • Recent
  • How To Video Stream - Specifications
  • Make Your Own Movie: Part 1, Production
  • Make Your Own Movie: Part 2, Distribution
  • Can There Only Be one? Regional production
  • What is the future of HEVC?

    Movie Stream Productions and MSP Insider are productions of TechGenie Media, LLC -

    Copyright © 2011 TechGenie Media.

    Home and About us.