Sound Design Stack Exchange is a question and answer site for sound engineers, producers, editors, and enthusiasts. It only takes a minute to sign up.
Sign up to join this community
I just came across this article (Atlanta Braves Radio in Surround):
http://radiomagonline.com/digital_radio/hd_radio/atlanta_braves_surround_sound_0907/
And checked out this link:
http://www.dts.com/DTS_Audio_Formats/DTS_Neural_Surround/DTS_Neural_Surround_UpMix.aspx
How in the world can you encode a 5.1 (or even 7.1) mix into a Stereo mix?
@Chris Hi Chris, combining multiple channels to stereo L+R or Mono is relatively easy. It's called "Downmixing. This is very common; radio, tv, dvd, etc. They key issue with downmixing are phase cancellations or timbral shifts due to constructive/destructive phase. If the source is recorded in surround using a coincident-array mic, e.g. Soundfield then phase will rarely be an issue when downmixed.
All multiple mic recordings have these problems when mixed to stereo/mono, and it could be argued that any multiple mic recording could be used to create a "surround-mix" {how well it achieves the recreation of a particular space is explored here, "Informal Comparison of Surround Techniques"
I think the more crucial question is the other way round. How on earth can a "near discrete" surround format be created from a discrete stereo source? The answer lies in the complex algorithm in the DTS upmix plugin and can be pointed to by considering the 3D sound-field possibilities with Binaural recording
Consider this, two slightly modified omni mics, i.e. baffled near-concident, can capture all the data from a recording space that would give all the time-domain, frequency-domain and amplitude cues required to create "3-D" surround, i.e. X-Y-Z planes. The weak link in the chain is the reproduction process, not the recording.
Hope this helps. James.
The stereo encoding does not yield true surround playback. Like James said, the DTS decoder detects the presence and character of the sounds in the left vs. the right channels, and places them in the surround field accordingly (for instance, sound material that appears at equal level and in phase in both the left and right channels of the stereo file, when decoded, is sent to the center channel). It is by all means an approximation of the original placement of the sound in the surround field, but it gives a convincing illusion (especially when concerning a professional sports broadcast, when most of the surround material will just be crowd sounds anyway). They even say so on the DTS link you posted: It "delivers a near-discrete 5.1 surround sound experience."
5.1, 7.1, 13.1 ... It's all the same concept spread over more speakers.
It is accomplished by "matrixing" and the product is called "matrix stereo". From what I gathered, Neural is to DTS what LtRt is to Dolby (both downmix a surround mix to stereo, the differences in algorithms might make them more different than they appear but to my innocent eye they are merely competitors products).
I reckon it might be equally interesting to know how a box we have at work can take 6 discrete channels in input and can emulate a 6-discrete-channel playback in your everyday pair of headphones (there was a large-scale headphones shootout to find the pair that would yield the best results, DT 770 Pro's came out glorious). There must be some crossing-over in the research for these technologies.
