Formats & Standards
DTS Neural:X
DTS Neural:X is an audio upmixing technology that extends stereo, 5.1, and 7.1 channel content to height and overhead speakers via an AV receiver's digital signal processor. Unlike the native DTS:X codec, Neural:X operates on legacy channel-based audio by analyzing spatial relationships to determine height channel placement without requiring object-based source material.
Mechanism: Upmixing vs. Native Codec
DTS Neural:X is an upmixing simulation technology that operates entirely within an AV receiver's Digital Signal Processor, distinct from DTS:X, which is a native audio codec embedded in mastered source content. This distinction is fundamental: Neural:X cannot enhance discrete object metadata that doesn't exist in the source file. Instead, it intelligently extracts spatial information and sonic cues from standard channel-based audio (whether stereo, 5.1, or 7.1) and remaps those signals to height and overhead speakers.
According to one enthusiast analysis, Neural:X is described as using a spatial remapping algorithm that determines height placement from spatial relationships already present in the source mix. However, this characterization comes from a single non-technical source and is not documented in official DTS technical specifications. The underlying mathematical model of this algorithm has not been publicly disclosed by DTS.
Input Formats and Speaker Configuration
Neural:X can process mono (1.0), stereo (2.0), 5.1, and 7.1 source material, making it applicable to a wide range of legacy content. The minimum speaker configuration required for Neural:X height processing is 5.1.2 (five main channels, one subwoofer, two height speakers). This baseline enables the decoder to generate the four distinct height channels that Neural:X creates: front-left height, front-right height, rear-left height, and rear-right height.
Neural:X processes up to a 7.1.4 / 11-channel configuration. The system operates within DTS's 11-channel maximum total channel count, which can constrain setups with more than nine ground-level (bed) speakers. This ceiling reflects the broader DTS:X specification, not a specific limitation of the Neural:X upmixer alone.
Comparison to Dolby Surround
Both Neural:X and Dolby Surround are post-processing upmixers designed for legacy audio, but they employ different signal-processing approaches. Dolby Surround uses a multiband decoder that processes timing delays and frequency-band information independently. Neural:X, by contrast, operates entirely in the spatial domain, attempting to preserve spatial relationships from the original mix.
In practical listening tests conducted by High-Def Digest, Neural:X operated approximately 2–4 dB more aggressively with height speaker levels compared to Dolby Surround in their test setup. However, this figure represents one outlet's measured finding from a specific configuration, not a universal specification.
The two upmixers handle content types differently. Neural:X pulls discrete sound effects upward to height speakers while leaving ambient sounds in the bed channels. Dolby Surround takes the opposite approach, excelling with environmental ambiance. One hobbyist reviewer concluded that Dolby Surround is superior for stereo content, where Neural:X exhibited comb-filtering artifacts in their subjective testing, though this assessment is not independently corroborated by measurement sources. For multichannel 5.1 and 7.1 content, both excel in different ways: Neural:X provides pinpoint object placement, while Dolby Surround excels with environmental ambiance.
Behavioral Differences Across Source Formats
Neural:X behaves differently depending on whether the source is 5.1 or 7.1 channel audio. According to one enthusiast reviewer, 7.1 sources yield greater placement accuracy than 5.1 inputs. This suggests that the availability of four discrete surround channels (rather than two) provides more spatial information for the algorithm to extract.
Practical Limitations
The spatial remapping approach underlying Neural:X cannot create discrete object data that does not exist in the source. It works best with content that was mixed with surround and height information already embedded in channel positions, less effectively on content that was originally stereo-centric or mixed without spatial depth intent. Additionally, the 7.1.4 channel ceiling means that high-end setups exceeding this configuration cannot benefit from Neural:X's full capabilities.
Sources
- [1]
- [2]
- [3]
- [4]
- [5]
Related
Reading