Dolby Atmos vs SMPTE 2098-2 / IAB format

[João Serra]

Hi guys,

Quick question about the SMPTE ST 2098-2 / IAB format that is supposedly compatible with Atmos:

Will a IAB compliant IMS/Processor that supports IAB (say, a GDC SR-1000 IMS or a SX-4000 plus a XSP-1000 processor) properly decode and render a Dolby Atmos IAB bitstream?
I know the standard, in theory, says yes, but it is only "on paper" or it works in practice? or in the real world i'll need a CP850/CP950A to pair with the GDC in order to properly decode Atmos?

If it really works, has anybody tried it yet? I'm saying this, because while standard defines the IAB bitstream format, one thing is the format, another is the rendering algorithm/code that Dolby actually uses in rendering the audio.


Thanks!

[Joao Lopes]

Simple answer: If you have a decoder/processor (Atmos / DTS-X / whatever) updated to play IAB. You play it. From now on forget "Atmos" or any other immersive sound system. IAB is universal and not brand/system proprietary dependent. You are free to choose any immersive processor that best suits your interests and budget.

[Harold Hallikainen]

I agree with Joao. I spent a lot of time working on the standard as part of the SMPTE committee. ST2098-2 is based on an input document from Dolby. It was then EXTENSIVELY reviewed and edited to make sure it was concise enough that someone could implement the bitstream based on the document and not some hidden knowledge. During the editing process, several people implemented renderers and encoders based on the draft standard and tested them. ISDCF also did testing of renderers. The input document and the standard include features that were not supported by Dolby at that point (and may still not be). To ensure bitstreams would play everywhere, ISDCF did extensive testing and came up with RDD57 (https://ieeexplore.ieee.org/stamp/st...number=9452728) which constrains ST2098-2 so encoders will not use features of ST2098-2 that are not implemented in renderers.

Finally, ST2098-2 describes the bitstream and the desired result (like "make the sound appear to come from this XYZ location"). The implementation of the panning algorithm and other audio characteristics vary system to system, but the overall result should be very close. This leaves room for different manufacturers to use different rendering techniques as each tries to get the best result.

Harold

[João Serra]

Thanks guys for the answers!