

In late October, I had a week of vacation, and an opportunity to take the hamshack computer to Milwaukee. Two DTV stations are operating there: WMVS-DT (PBS) on channel 8, and WTMJ-DT (NBC) on channel 28.
The reception was from my mother's house in Menomonee Falls, 18km (11mi.) west-northwest of the two transmitters. Both stations' NTSC analog transmitters operate at maximum allowable power from the same towers as their DTV operations. WMVS analog is on channel 10, and WTMJ on channel 4. WMVS-DT's power is 25.1kw@354m which I believe is the maximum permitted for this channel. The antenna is actually higher than their analog antenna. WTMJ-DT's power is 1000kw@273m which is definitely the maximum for UHF DTV, though the antenna is 32m (100') lower than their analog. Strangely, I was not able to locate their DTV antenna while driving past the site. The antenna used was a set of rabbit ears - far from optimal for channel 8 and even further from optimal for channel 28.
First tests were made with the rabbit ears connected to a balun and then plugged directly into the back of the computer. I was able to capture a brief still frame from WMVS-DT, but no consistent signals. Looking at analog signals, it became apparent why: massive computer interference.
Luckily, there's a Radio Shack in Menomonee Falls. I bought a 6-foot RG-59 jumper cord and a F-connector "barrel". Couldn't remember how to get the diagnostics screen up, so I just manually added DTV channel 8 and began moving the antenna around the room until WMVS-DT audio appeared. It turned out the antenna could be set atop a cedar chest at the foot of the bed and give consistent reception. But a different program would come up each time the signal locked in! More on this later. I couldn't get anything at all on WTMJ-DT, but it would later prove there was a reason for this.
Eventually, I remembered how to bring up the diagnostics menu. (hold CTRL while holding the right-hand mouse button down for at least a second) This menu also shows which virtual channels the station is offering, and allows you to select one. WMVS was running four virtual channels, which explains why the program kept changing! The channels were titled "WMVS SD 1" through "WMVS SD 4". For up to maybe 30 seconds after acquiring lock on the WMVS-DT signal, some of the virtual channels might be missing from the list. There did not seem to be any pattern.
During the day, none of the four virtual channels on WMVS-DT matched the program on their analog signal. PBS provides a pre-encoded DTV stream (complete with virtual channel numbers 80-1 through 80-4; the WinTV-D doesn't display virtual channel numbers) on their satellite, and some PBS stations have been known to simply pass this signal through to their transmitters unmodified. That seems to be the case here, as no local promotions or IDs were seen. I guess the FCC is looking the other way on the legal ID and EAS rules.
The WMVS-DT virtual channels included two channels of children's programming - mostly cartoons during the day, live action at night. There was also a channel of information programming and one of cultural programming. Whatever you think of cartoons, they are a very good test of the color rendition and resolution of a TV system. The cartoons on WMVS-DT looked FANTASTIC. The live-action programs didn't look half bad either. I couldn't evaluate the audio properly as I was using regular computer speakers.
During "Teletubbies", a strange black strip with blinking white lines appeared at the left of the screen, going down about as far as the bars in color bars. I've heard that interactive Teletubbie dolls are available which will react to events in the TV show; my guess is this strip was the data the dolls pick up on. The DTV card doesn't overscan unless you select full-screen mode (even then, it doesn't overscan much) so things that won't show up on a real TV did show up here. Including the Line 21 captions on analog signals.
After going out to dinner and a few other things, I tried WTMJ-DT again - and now it worked. They were in the 6:00 news. Video quality, especially color depth and resolution, were noticably poorer than WMVS-DT. My presumption is that they were upconverting analog signals to DT. Unfortunately, I neglected to ask about this when I called the station on Wednesday morning. They did confirm that WTMJ-DT is not operating during the day; they only operate after 5pm. WTMJ-DT has only one virtual channel, "WTMJ Digital".
The presence of the Line 21 captions at the top of the DTV picture would tend to support this presumption. My understanding is that if you connect a NTSC analog TV to a DTV converter, the converter is supposed to automatically pull the closed-captions from the datastream and stick them in Line 21 of the analog output. It is not supposed to be necessary to transmit captions on Line 21 of a DTV signal. (but of course, if that DTV signal is upconverted from the same video used to drive the NTSC transmitter then the captions will be present)
Unfortunately my relatives are early risers, and I was unable to stay awake late enough to determine whether WTMJ-DT was passing the HD feed of Leno. The WinTV-D card downconverts everything to 480p anyway (though Hauppage does now have a card that does HD) but I would have expected better color and sharper resolution.
Two questions I wanted to answer were acquisition time and delay time through the DTV system. Acquistion is quite quick; it took less than a second to begin displaying a picture. Sound came even faster. This was definitely a DTV decoding delay, as the card acquires analog signals instantly. I also had an opportunity to compare the audio coming through the DTV system to the same station's analog audio. (admittedly through the Menomonee Falls cable system which may have added some delays of its own) The DTV coding delay appeared to be on the order of 1/2 second. Much better than DirecTV!
Signal breakup could lead to one or more of several symptoms:
I2C would be found almost instantly if there was any signal at all. It
didn't seem to indicate much. Sync Lock would come next, and only on DTV
signals. If the diagnostics window was brought up while tuned to an analog
station, this window would say "ANALOG". Finally, the Eq. Lock would come
on, and the audio and video would decode. Again, it would display "ANALOG"
on analog signals. A minimum SNR of roughly 20dB seemed necessary for lock.
Here are some pictures of TV reception on this card. Unfortunately the JPEG
conversion process seems to have introduced some noise in the digital signal
- and cleared up some of the noise in the WMVS-NTSC analog signal - causing
things to look more similar on the Web than they did in real life.

(unfortunately, the computer failed to capture the whole screen but this is
the important part)
WMVS NTSC analog
WMVS DTV
WTMJ DTV