Join Date: Sep 2002
Location: montreal, canada
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Last November an article was published on HOMETHEATREHIFI.COM that claimed "There is a bug in the Silicon Image HDMI transmitter that pops up when converting YCbCr to RGB."
This reputed bug applied in cases where a HDMI transmitter was talking to a DVI receiver.
The argument was that DVI is 8 bit RGB while HDMI can be 8 bit RGB and also 8 bit, 10 bit, or 12 bit YCbCr as well. If the broadcast studio was sending the cable box a compressed signal that is ultimately presented to the HDMI transmitter in the YCbCr format, then the transmitter, knowing that it is talking to a DVI display device, is required to convert this format to RGB. Apparently, this conversion results in some bits being truncated in the calculation and the transmitted signal does not have quite the original video dynamic range, so to speak.
I do not know how a reduced video dynamic range manifests itself in terms of picture quality , but I assume that black is not as black as the original, and white not as white as the original.
I plan on buying a SA8300HD with a HDMI output for use with a Sony XBR800 with a DVI input, so this question is of particular interest to me.
This is the only case where it appears that in practice HDMI is not always backward compatible with DVI, although on paper it may be so.
Does anyone know if SA8300HD uses a Silicon Image HDMI transmitter chip, and if so, has this bug been corrected?