With the latest product announcements for the consumer, it looks like surround-sound broadcasting and HD Radio are making more headway in the radio broadcasting arena. However, there are still many questions to be answered and many tests are underway to address the concerns of surround, which include the challenges of downmixing and the concerns about the three major proponents that are vying for acceptance. It's too early to address the proposed MPEG surround technology, as its standardization is not yet complete and hardware-based encoders/decoders will take some time thereafter to develop.
As you may know, the three major proponents (Dolby Prologic II, Neural Audio 5225 and SRS Labs Circle Surround) take discrete multi-channel surround content and encode the multiple channels of audio into a stereo mix for transmission on standard two-channel delivery systems. The surround-encoded two-channel stereo mix is called a downmix and referred to as Lt/Rt. Conversely, upmix is the term used to describe the surround material that is decoded from the two-channel stereo (Lt/Rt) mix. Because of this downmix/upmix process, none of these systems create a true discrete multi-channel upmix identical to that of the original content, however, some subjectively do a better job than others depending on the content and can complete the task of delivering the surround content quite impressively. Additionally, unlike the older versions of these systems, all of the proposed systems provide full bandwidth and stereo imaging to the rear channels.
Two of these technologies, Dolby's and SRS Labs', use a matrix encoding/decoding method that alter channel levels and phase relationships to create the downmix. Neural Audio's 5225 method implements many of these matrix techniques to offer legacy system compatibility, but it also embeds steering information into the downmix to create a reliable upmix. The question that remains is whether or not all these systems are cross-compatible as is in practice today; allowing one to encode with one system and decode with another, providing the consumer with a satisfying surround experience that was originally intended. Last fall, thorough subjective listening tests began to answer this question, with further tests in the next few months to verify the preliminary findings.
Regardless of which system is used, including the MPEG system, creating a downmix will be inevitable. Though there are valid arguments to use stereo artistic mixes alongside the surround broadcasts, many obstacles would have to be overcome. These include the proliferation of the matrix decoders already in the market and the delivery constraints that will continue to exist. Then there is an often-overlooked complication that the song structure of many of the commercial surround-sound discs simply do not match their artistic stereo counterparts. Bundle that with the time alignment issues of content and blending of analog to digital signals, and it's obvious that downmixing will be in place for some time.
Because of this, carefully watch what occurs with the resultant downmix, regardless if one takes content from commercially available discs or creates his own. Because these systems are taking 5.1 (or sometimes 6.1) channels of discrete surround material and encoding it into two channels of audio, the phase relationships, delays, equalization and other effects that exist in the discrete-channel surround mix can cause comb filtering and in turn, loss of definition of instrumentation and vocals in the downmix. If you are working with classical material that has mostly ambience in the rear channels, the resultant downmix should have few problems. However, many surround titles include full instrumentation directed to all speakers, immersing the listener in the multi-channel experience, it is possible that the downmix may have problems.
Two things to avoid are delaying the rear channels of any material before passing them on to the encoder, and placing the vocal in the front and rear channels. The Neural Audio 5225 and the upcoming MPEG Surround system use special algorithms that alleviate many of these problems in the downmix quite effectively.
|Note: LFE and bass management are not the same! The LFE channel is part of the playback functions of recorded media such as DVD. Its function and purpose is often mistakenly used for bass management control. Bass management is a separate function that takes bass frequencies from the main surround audio channels and redirects them to the subwoofer, as many smaller monitors are unable to produce low frequencies.|
The use of the low frequency effects (LFE) channel is often overlooked. In most downmix scenarios — especially music — the LFE signal is used sparingly if it used at all. The LFE was originally created by the movie industry to provide dramatic effects to the listener. Therefore, if you transfer material and there is extensive use of the LFE channel, monitoring the downmix and even the upmix is crucial. Overuse of the LFE can create too much bass, loss of bass or other undesirable effects, especially on poorly calibrated home theater systems. The LFE channel should not be used for the bass content of the main speaker channels, but rather allow the bass management of the playback systems to distribute bass appropriately. There are guidelines available from Dolby's and SRS Labs' websites as well as from the Producer's and Engineer's Wing of Recording Academy at Grammy.com.
There are still other issues of debate concerning surround sound encode/decode solutions for HD Radio as well as standard analog broadcast. For instance, the increased amount of L-R (left minus right) energy that the systems are producing and the effect it may have on the analog side of the FM broadcast chain.
Next time, I'll review tests that are underway to address this, provide analysis of the amount of L-R energy each system produces, and a comparison of the different surround encode/decode systems proposed for HD Radio.
Kosiorek is the audio recording, mastering engineer at the Corbett Studio at Cincinnati Public Radio.