junh1024 (Customer) asked a question.
Dolby Atmos technical questions - codecs, objects, and signal processing
NB: I'm asking here since Dolby support redirected me to these forums.
1. When using binaural monitoring compared to stereo monitoring in the Dolby Atmos Renderer, the peak levels could be increased. Does Dolby Atmos analyze the audio and enhance transients? Or is this simply due to that the signal is changed, and hence the peaks are changed?
2. When I encode my mix to DDP Atmos, the peak levels could be increased. Does DDP Atmos analyze the audio and enhance transients? Or is this due to the lossy nature of DDP ?
3. Does (lossless) Dolby TrueHD enhance transients? Or not?
4. About the decorrelation algorithm for objects above a certain size. Does it analyze the audio of the objects and adapt the decorrelation depending on the audio content, or is it just a simple digital phase shift?
5. Does Dolby apply additional audio enhancement to objects compared to bed, so that objects "stand out" from the mix? Or are level differences between objects & bed more related to panning & rendering?
6. Disregarding B-chain factors, Is audio output from a stereo rerender different to stereo monitoring? Or should it be very similar?
7. Disregarding spatial coding, object size, and codecs, Do objects interact with each other? Or not?
These are some conceptions that other people may have about Dolby Atmos. And therefore, it would be good if these conceptions were clarified.
1. When using binaural monitoring compared to stereo monitoring in the Dolby Atmos Renderer, the peak levels could be increased. Does Dolby Atmos analyze the audio and enhance transients? Or is this simply due to that the signal is changed, and hence the peaks are changed?
2. When I encode my mix to DDP Atmos, the peak levels could be increased. Does DDP Atmos analyze the audio and enhance transients? Or is this due to the lossy nature of DDP ?
3. Does (lossless) Dolby TrueHD enhance transients? Or not?
4. About the decorrelation algorithm for objects above a certain size. Does it analyze the audio of the objects and adapt the decorrelation depending on the audio content, or is it just a simple digital phase shift?
5. Does Dolby apply additional audio enhancement to objects compared to bed, so that objects "stand out" from the mix? Or are level differences between objects & bed more related to panning & rendering?
6. Disregarding B-chain factors, Is audio output from a stereo rerender different to stereo monitoring? Or should it be very similar?
7. Disregarding spatial coding, object size, and codecs, Do objects interact with each other? Or not?
These are some conceptions that other people may have about Dolby Atmos. And therefore, it would be good if these conceptions were clarified.
A lot of questions there. I don't have answers for all of these and not sure what you mean by enhance in this context. I've tried to answer inline below.
1. When using binaural monitoring compared to stereo monitoring in the Dolby Atmos Renderer, the peak levels could be increased. Does Dolby Atmos analyze the audio and enhance transients? Or is this simply due to that the signal is changed, and hence the peaks are changed?
The binaural renderer is a completely separate renderer than the renderer that is rendered->downmixed to stereo. The binaural renderer is consistent a bit louder due to the separate rendering process. Every is louder. Peaks are not enhanced.
2. When I encode my mix to DDP Atmos, the peak levels could be increased. Does DDP Atmos analyze the audio and enhance transients? Or is this due to the lossy nature of DDP ?
Again it depends on how you are decoding the Atmos. Dialnorm metadata and the playback device determine the reproduction levels so it isn't really a matter and anything being deliberately modified or enhanced.
3. Does (lossless) Dolby TrueHD enhance transients? Or not?
No. See 2.
4. About the decorrelation algorithm for objects above a certain size. Does it analyze the audio of the objects and adapt the decorrelation depending on the audio content, or is it just a simple digital phase shift?
I'm not entirely sure but this method has been updated with the last version of the Renderer. I will check into this.
5. Does Dolby apply additional audio enhancement to objects compared to bed, so that objects "stand out" from the mix? Or are level differences between objects & bed more related to panning & rendering?
I'm unclear what you mean by the word "enhancements" in this regard.
6. Disregarding B-chain factors, Is audio output from a stereo rerender different to stereo monitoring? Or should it be very similar?
The stereo re-render and output should be identical unless the output limiter is kicking in.
7. Disregarding spatial coding, object size, and codecs, Do objects interact with each other? Or not?
Not sure what you mean by this. Objects are rendered independently of each other but rendering always involves a degree of summing so they do affect each other in that regards.
Hope this helps.
Best,
Adam
These are some conceptions that other people may have about Dolby Atmos. And therefore, it would be good if these conceptions were clarified.
4. Please get back to me about objects decorrelation when appropriate.
5. To rephrase my question: When comparing object panning to bed panning, the Peak levels are sometimes different. Is this the result of Dolby Atmos changing the audio content of objects compared to bed (like enhancing transients or harmonic content), or is this due to rendering/summing/panning differences?