Quote:
Originally Posted by psgcdn  That was the case with Dolby Digital on DVD, that Dynamic Range Compression (DRC) (similar to midnight-mode on AVRs) would engage when fewer than 5.1 speakers were used to avoid clipping the analog output of channels when signals are combined.
|
For DVD and DD, the compression was done digitally in the decoding process. The intent was to avoid digital "overs" when downmixing caused by sending to much data to the processor. For example, in Phantom CC mode, the CC info is added to the L & R channels. If the CC content is high and the L& R channels are already high, the added content could overload the processors, causing errors. To avoid this, controls are placed in the "Metadata" to avoid overload. From the Dolby document labeled:
Dolby MetaData Guide, Vol. 2 (attached)
Quote:
In addition to the DRC profile, metadata can limit signal peaks to prevent clipping during downmixing. This metadata, known as overload protection, is inserted by the encoder only if necessary. For example, consider a 5.1-channel program with signals at digital full scale on all channels being played through a stereo, downmixed line level output. Without some form of attenuation or limiting, the output signal would
obviously clip. Correct setting of the dialogue level and DRC profiles normally
prevents clipping and unnecessary application of automatic overload protection. |
Quote:
Originally Posted by psgcdn I wondered if they were being smarter with lossless audio. The obvious way to do it without losing any dynamic range would be to decrease the gain of all channels. It sounds like they are still decreasing dynamic range when phantom center is engaged, which would be a huge reason not to use a phantom center.
|
For TrueHD, the process is different. The following is from a Dolby document entitled:
Dolby® TrueHD Audio Coding for Future Entertainment Formats (attached)
Quote:
Channel Extensions, Downmixing, and Dolby TrueHD
One channel extension technique is the method by which MLP Lossless, Dolby TrueHD, and MPEG-2 LII deliver compatible downmixes for soundtracks with expanded channels. In these codecs, a 7.1-channel soundtrack is first downmixed to create a 5.1 mix, which is supplemented by a two-channel extension (which well call extension B). The 5.1 mix is then further downmixed to a two-channel stereo mix, and another supplemental stream is created that carries the 3.1-channel extension A.
So the 7.1-channel program is delivered in three separate components: a two channel mix, the 3.1-channel extension A, and the two-channel extension B. The total payload is still 7.1 channels, with preconfigured subsets to create two-, 5.1-, and 7.1-channel presentations. If a listener desires a stereo presentation, the decoder plays only the two-channel downmix, thereby minimizing DSP resources for the simplest hardware productsa useful idea. If a listener selects a 5.1 presentation, the decoder reconstructs it from the two-channel downmix plus the 3.1-channel extension A substream by means of rematrixing. If a listener wants a 7.1 presentation, the decoder reconstructs it by rematrixing the reconstructed 5.1-channel program with the final two channel extension B substream.
This all works nicelyon paper.
However, when used with lossy codecs that rely on psychoacoustic principles such as noise masking, this rematrixing can reveal coding artifacts that were otherwise inaudible. Its not that the coding artifacts have increased; instead, they become physically separated from the sound that originally masked them. As a result, the main sound and the coding artifacts may be directed to different loudspeakers, taking different acoustic paths to the listener and resulting in a phenomenon called coder unmasking. |
Bottom line, depending on the source material, there will be some sonic impact from using "Phantom CC." However, it will be a different cause/effect scenario to blame depending on whether you're using DVD DD or BlueRay TrueHD.
BTW, DTS/DTS-MA doesn't use MetaData, and does not have this "issue."
Craig
Dolby Metadata.pdf 120.4482421875k . file
TrueHD_Tech_Paper_Final.pdf 149.51953125k . file
Your FR is outstanding, although I usually use 1/6 smoothing since it is often cited as the finest resolution we actually hear. When calibrating I will often go with a higher resolution (1/24) to get a little better picture of what's going on, then retake a shot at 1/6 to be sure I got everything that matters.
