Measured at the U-verse gateway, AT&T's HD streams for video are approx. 5.7 Mbps. This makes sense as on the 32/5 U-verse profile (this is the one assigned to you on AT&T's fastest Internet speed offering, 24/3 Mbps) - one's Internet connection speed drops to 20 Mbps when viewing two HD channels at once - 14 Mbps when viewing 3 HD channels - and just 8 Mbps when viewing four HD channels. So yes, Internet speed loss on U-verse when viewing 2 or more HD channels definitely correlates with the 6 Mbps of bandwidth per HD channel as well.
I had U-Who and DirecTV side by side.
Since I'm so close to the VRAD, by profile is 32.2/5.04, yet my modem reports the max downstream rate is 64 Mb/s. Whether U-Who will use this overhead to feed TV, I don't know. My internet is 12 Mb/s.
I setup 3 HD recording and had a forth HD channel live on my other receiver. While these were going on, I tested my download rate by finding large files to download. I saw no speed reduction in downloading.
The point of the comparison was to illustrate how bad AT&T U-verse looks at that bitrate - that a recording from Verizon FiOS of the same channel uploaded to YouTube still looks better than AT&T even after YouTube molests it. It's that bad.
FiOS is MPEG-2 correct? You can't compare bit-rates between MPEG-2 & MPEG-4, as it's apples and oranges.
You can compare PQ, but it's subjective and effected by the scaling in the display. My Sony XBR can make a fairly crappy SD program look "fairly good".
When I tried U-Who, I figured the PQ would suck so bad it would be the biggest complaint. Well it wasn't to my surprise. I did see skipping/dropping of frames, so it does have limitations.
As for whether or not DirecTV looks any worse with the 25% bitrate drop, I'd take opinions without some screenshots to back it up with a grain of salt.
Where are you getting this?
The average bit rate reduction from MPEG-2 to MPEG-4 is close to half.
MPEG-2 is fairly constant as it sends every bit needed for every frame. MPEG-4 only send the bits for what changes in the next frame. This means bit rates vary much more with MPEG-4, where they can be as low a only a couple of Mb/s, and then peak at 16+ Mb/s with DirecTV, as the frame content changes.
UV Realtime allows monitoring of the router ports.
I setup the same recorded show from U-Who and DirecTV, and sync'd them within a sec of each other. I then monitored my router port usage over about 10 mins to compare MPEG-4 to MPEG-4. U-Who's bit-rates were 66% of DirecTV's, with the main difference being around an 8 Mb/s limit on U-Who, where DirecTV was averaging 12 Mb/s. When DirecTV was below 8 Mb/s, U-Who's rate was much closer.
Really the only way to know scientifically is for a thorough identical frame screenshot comparison to be done like the AT&T vs. YouTube comparison above: screenshots need to be taken of a program before it gets dropped down to 6 Mbps and then a repeat of the program after the channel is dropped down.
That may be what you need to do, but the act of taking a screen shot has a variable that we don't know.
Actual bit rates, monitored over time, of the same program, and the same type of encoding, seems more useful to me, which I've done.
DirecTV PQ: I've also done comparison between OTA MPEG-2 and the DirecTV MPEG-4 of the same program. There is a slight difference, but it's so small that it might fall under "I want to think" there is a difference and so I think there is. It's that close.
Edited by veryoldschool, 04 July 2012 - 09:30 AM.