In an Optical medium, The Dolby Digital print has a 5776 dots each representing a bit of data in one perforation. Therefore 4 perforations per frame and 96 perforations in one second. So in one second dolby is accommodating 554496 bits of data. However the data in an uncompressed wav in second is 6912000 bits of data is stored.

What kind of compression does my audio go through when the AC3 codec is used ? How come there is no drop in quality and while there is a huge, drastic drop in data ?

  • "no drop in quality" - you think?
    – user49
    Commented Nov 30, 2011 at 6:01
  • 7 million bits assume the sound is 48KHz 24bits and with six fullrange channels, but AC-3 in 5.1 only admits 48KHz 16bit for five speakers and the LFE-channel limited to 240Hz 16bit, meaning only a little less than 4 million bits (3 843 840). But as the bitstream also has start and stop-bits, as well as some simple error-correction, meaning the compresion ratio is about 10:1, roughly the same as fairly decent MP3-file. Though unlike MP3 with a fixed compression-method, ofcourse :-) Commented Dec 1, 2011 at 8:37
  • True. It's a very drastic compression ratio. But In the Optical medium , The loss of data is not as audible as it would be in digital. Commented Dec 1, 2011 at 17:14

4 Answers 4


It's going through a perceptual (think mp3/aac) and matrix encoding process. And there is a drop in quality, but most people aren't listening on speakers where they would be able to clearly hear the difference (at home anyways). That's how it is with any of the flavors of Dolby or DTS encoding formats. Though it's interesting to note that Blu-Ray have started shipping with unencoded, discrete, multi-channel audio on them.

  • I am not talking about Listening Audio at Home, We have a Dolby 7.1 studio in our institue. I have heard sound on both the A-Chain and B-Chain, I mean literally toggled between The Pro Tools Session and the Optical Digital Track playing simultaneously. I was not able to spot the difference. Audio is going through that compression but the quality loss is not at all noticeable. Both were being played on Martin Audio Speaker Set up. Commented Nov 30, 2011 at 15:28
  • @Audio Union - That's the perceptual (psycho-acoustic) codec at work. I believe it analyzes across all channels. The more audio information there is spread across the channels, the more the codec pulls out masked frequencies...i.e. frequencies from the front can mask non-fundamentals in the rears. You still get a reasonable spread of frequencies, but you're not getting all from every channel. Our brain fills in the gap based on what is available around us. If there's little going on (i.e. just dialog), not much compression needs to be applied, if any, preserving the quality. Commented Nov 30, 2011 at 15:47
  • @Shaun Farley - Thanks That Answers My Question. Commented Nov 30, 2011 at 17:40

The way To clearly hear the destruction is to feed strong complex signals to the fronts and some nice sounding orchestra music at lower levels to the rears. Mute the fronts and only listen to the rears... Make sure to mute AFTER the DMU and that you monitor through the ac3 process.

Sometimes bass heavy sounds gets seriously messed up as well. It used to be a big issue, but Definetly have become better the last few years.

The perceptual coding and data reduction for film prints is crazy. There's way more data on a DVD or bluray compared to what's available on a 35mm print. Otoh it's quite a feat getting it to sound as decent as it does with the limited data rate available.


Actually there's a great drop in quality! If you ever listened to a mix while making the Dolby optical disk (so, in the same room where you mixed) you clearly hear drop of quality, definition, dynamics. One of the great things of DCP projections (or Blu-Ray discs), is the uncompressed audio format. I think that abandonment of AC3 for uncompressed audio and better calibrate theaters are the most prominent improvement we could have in the next few years in sound.

  • Some Audio and Video Production Softwares are providing uncompressed Multi Track audio data. WAV (Poly) softwares Like Logic Pro is one of them. I understand that. There is a major compression of Dynamics in Dolby SR, which Analogue Optical tracks, Dolby Digital which is 5.1, 6.1 and 7.1 are Digital Data Tracks. The data while exhibition is decoded in the CP650. AC3 is an Digital codec, there is no Audio compression happening but data Compression. That data compression is not loosing out on that much of dynamics because it is measured at the mixing stage and during the mastering stage. Commented Nov 30, 2011 at 15:38
  • 85 dbSPL in LCR and 82 dbSPL in LS/RS/BS.This limitation is set by the Dolby Laboratories Inc. All theatre's are aligned to this specification. All Audio is reproduced under this specification. I personally cannot hear any loss of dynamics or definition. Commented Nov 30, 2011 at 15:44
  • Remember I am not talking about a small medium such as DVD's and Blu Ray's . I am talking of Theatre projection and Dolby application in Cinema. Commented Nov 30, 2011 at 15:46
  • I'm talking about data compression and the Dolby Digital track. While mastering the optical disc with DMU, try to switch between the output of the console and the DMU output (not LtRt): I can clearly hear differences, maybe not in a dialogue sequence, but on music and dynamic sequences it's evident. I don't know how well calibrated are movie theaters outside Italy, but here in Rome, only a bunch of theaters sound really well to my hears, and only a few have the CP650 set to 7. Commented Nov 30, 2011 at 16:33
  • Actually It's not that different in India. CP 650 Ref.7 is is barely kept usually they blow out their amps and keep Ref. at 4.5 - 4.9. I guess you guys do have valid points. Commented Nov 30, 2011 at 17:25

I too think the loss of both details, definition, and transients is humongous in Dolby Digital, though it is by all means a good format considering what it's made for. Personally I don't like it anymore though, but I do respect it.

Most people doesn't think much of it as they don't really know what it could have been, but they react to it in a very physical way as the human conception of reality is very much based on small sometimes barely noticeable details and imperfections which helps us distinguish between different sound-sources in a more realistic way. Dolby Digital, like MPEG Layer 3 and such for example, is losing most if not all that small pieces of air in the process of reducing the bitstream. That doesn't mean the mix gets totally bollocks, it just means the audience doesn't get as into the illusion as they would had had the soundstream been richer.

For me, DTS is da shit! It's still lossy, but it is MUCH more transparent to the material due to both lower compression ratio and frankly a better algorithm.

It's difficult for me to explain exactly how it works, but it is not based upon samples but in frequency response. By eliminating stuff the codec doesn't think is perceptible, or at least unimportant, the codec can keep the soundfile small. One of the things lost in this kind of compression is the extreme highs, and in at least MP3, the extreme lows. I'm not sure whether or not the fullrange speakers in AC-3 have much sub-bass. I do know the LFE does (well duh! ;-), but I actually hasn't checked a Dolby stream in an analyzer yet as I has with MP3 :-) A more direct effect is that the sound gets more dull and compact, with more transient sounds risking to drown in the mix, at least on lower settings. Some people like this effect though!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.