dw wrote:
On 22/09/2013 12:51, Andy Furniss wrote:
Hi
I do not have a 7.1 sound system so can't actually test this, also
as may become apparent I don't know much about sound :-)
I would be grateful if someone could correct/confirm the
following.
If I were to mix down a digital 7ch to mono I have to reduce by 1/7
amplitude to prevent clipping, so one channel will be -16.9dBfs.
This assumes that all 7 channels are driven with the same maximum
level, in phase, and that this would get past the mastering stage,,
I'm sure they bear in mind the possibility of stereo, or mono
mixdown.
Yea I guess all channels full at once only really happens on the front
three in practice, mine was just an example to make the figures as
different as possible.
I don't quite understand the "in phase" though, are you saying that they
artificially adjust phase for the same sound that comes out of more than
one speaker to affect the mixdown?
Level wise, I don't think my method is any different from and software
decoder, except of course various channels get different weighting
before normalisation.
Peaking at 0dbfs is evident on a couple of film soundtracks, disk and
BBC broadcast I've just looked at (without compressing using the DRC
metadata, of course).
Of course in practice consumer 7.1 currently means TrueHD or DTS MA
which seems to be mixed up rather than down - so there is a studio
stereo mix there in the stream to be decoded directly.
If I play the track over one speaker the volume difference between
one and all channels will be 16.9dB.
If I were to measure the volume at listening position (assuming
anechoic and equal speaker distance) with a real 7 speaker setup
then the volume difference, because the speakers are not close,
would add up using power not amplitude so the difference
heard/measured between 1 and 7 at full power would only be 8.45dB,
so there is quite a large dynamic range discrepency?
You would add powers if the 7 channels had random relative phase
(which is less likely with "anechoic and equal speaker distance"
Ok, so would infinite distance planewaves playing the same sound add as
if the speakers were coupling (I thought you needed < 1/2 wavelength for
this) to gain twice as much as power alone even though they are not in
the same direction? Or is this a just a mathematical perfect position
effect that wouldn't really happen in a real 7.1 setup even if the
speakers were playing the same sound?
The real reason for this question is more to do with simulation
than real life, so perhaps that will make a difference - if the
speakers are infinitely far producing planewaves and a soundfield
is at listening position would that change anything for what levels
the virtual omni would "hear".
Sorry for all the questions - I would just like to understand if I use
supercolliders 7.0 ambisonic encoder so I can then do either personal
hrtf or uhj stereo decode, whether the simulation it uses is correct in
the sense of the levels I would get with a real soundfield measuring
real speakers.
Thanks.
Andy.
_______________________________________________
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound
