Half-D1 from Premiere CS4

[nharikrishna]

Hi all,

Q1: I digitized VHS tapes using Canopus ADVC-110 into DV format. Since many have said that Half-D1 is a sufficient (and better) resolution for VHS captures, how do I get my video encoded to Half-D1 in Premiere CS4? It comes with MainConcept encoder built-in, and there is no option to choose Half-D1 there, and ther are a lot of 'presets'. MPEG-2 DVD option doesn't give me Half-D1 as one of the resolutions to output in. I know I can use HcEnc separately, but I want to stick to MainConcept in CS4.

Q2: If I capture SD Video from Satellite box (using ADVC), is Half-D1 the right resolution to put on DVD? or I need to maintain full resolution?

Q3: If I capture HD Video from Satellite box as SD (through analog composite, S-Video or component connection), then would I benefit from full resolution (since input is HD) or I can still go with Half-D1?

My output format is DVD-Video.

Thanks.

[lordsmurf]

Start from the Generic MPEG-2. I can't give further help right now, as my CS4 system is encoding.

352x480 is sufficient for tapes or broadcasted SD. Only DV is truly a max resolution SD 720 source. Some of the satellite is between 352-720, but resolution isn't all that affects detail anyway. Some of those stations are very compressed.

Don't capture 16:9 as Half D1 -- it looks bad. Use 720x480 Full D1.

[nharikrishna]

Thanks LS.

Start from the Generic MPEG-2.

I selected "Generic MPEG-2" as opposed to "MEPG-2 DVD", but am confused with a host of settings that appear under Generic MPEG-2.

352x480 is sufficient for tapes or broadcasted SD. Only DV is truly a max resolution SD 720 source. Some of the satellite is between 352-720, but resolution isn't all that affects detail anyway. Some of those stations are very compressed.

Ok, understood.

Don't capture 16:9 as Half D1 -- it looks bad. Use 720x480 Full D1.

Ok. That answers my HD capture question. So the rule is, capture and encode at 720x for all HD stations since all of them will be 16:9.

[edDV]

Hi all,

Q2: If I capture SD Video from Satellite box (using ADVC), is Half-D1 the right resolution to put on DVD? or I need to maintain full resolution?

Q3: If I capture HD Video from Satellite box as SD (through analog composite, S-Video or component connection), then would I benefit from full resolution (since input is HD) or I can still go with Half-D1?

My output format is DVD-Video.

Thanks.

North American satellite providers typically downlink between 480x480 (old DISH standard) to 704x480 with ~528x480 being typical for SD channels. HD channels are downlinked to the set top box between 960x720 to full 1920x1080. The S-Video down conversion would be to 704x480 before D/A and NTSC encoding.

Under no circumstances is 352x480 sufficient to adequately sample satellite video. The only time you would consider 352x480 is for extreme compression where the trade off is resolution vs compression artifacts.

Also, for internet uplink, the safe default is to square pixel 640x480 or 320x240 for 4:3 SD.

My output format is DVD-Video.

In that case, you must choose either 720x480 (oversampled) or 352x480 (undersampled).

[lordsmurf]

Under no circumstances is 352x480 sufficient to adequately sample satellite video. The only time you would consider 352x480 is for extreme compression where the trade off is resolution vs compression artifacts.

But that's only accurate on paper, from theoretical stance.

Resolution is just the container. You can have lower detail video (closer to 352x480) regardless of broadcast container resolution. Much of the detail is thrown out at the time of encode for uplink. It really depends on which channel, which broadcaster, etc. It's not as easy as saying "352 isn't enough" because that's just not accurate. 352 usually is enough, especially if you're wanting to maximize bitrate without sacrificing disc length (i.e., more than 1 hour per disc).

[edDV]

Under no circumstances is 352x480 sufficient to adequately sample satellite video. The only time you would consider 352x480 is for extreme compression where the trade off is resolution vs compression artifacts.

But that's only accurate on paper, from theoretical stance.

Resolution is just the container. You can have lower detail video (closer to 352x480) regardless of broadcast container resolution. Much of the detail is thrown out at the time of encode for uplink. It really depends on which channel, which broadcaster, etc. It's not as easy as saying "352 isn't enough" because that's just not accurate. 352 usually is enough, especially if you're wanting to maximize bitrate without sacrificing disc length (i.e., more than 1 hour per disc).

But in most cases, minimum input to the satellite encoder/transponder these days is 704x480. Next you have to look at the source video resolution ingested to the playout server. Typical tapes will be

Betacam SP ~ 400x480 (analog 4.5 MHz)
Analog 1" type C ~640x480 (analog 5.5 MHz)
DV/DVCAM/DVCPro digital 720x480
DigiBeta 704x480
D2/D3 digital 768x480

Typcally Betacam SP was restricted to news and local programming so most sat network tape library will be 640x480 or above effective resolution.

[lordsmurf]

Yes, that's all true --- but the effective resolution is lowered significantly as a byproduct of the low bitrate high MPEG-2 compression now in use with most SD digital transmission streams. At 2.5Mbps average datarate, even with custom tweaks and long GOP, you're not going to fit that much resolution (and allocated pixel/bit) into the stream. The MPEG encode process will have a softening effect, limiting actual detail to that of much less than the container. It almost always falls in the 352x480 to 480x480 range, even if found within 640x480 space.

I've been complaining about this for years. Image quality was sacrificed in the interest of stuffing the bandwidth with more channels. And many times, those are worthless channels forced by must-carry riders on contracts to carry the desired network.

[edDV]

That can be true but DBS and cable networks have added statistical multiplexing allowing a transponder's 36 Mb/s to be allocated across ten or more SD channels. First each channel is biased for average bit rate. For example HBO is given more average bit rate vs the weather channel. Second motion bit rate demands are shared across the group of channels. This can be used to increase quality over all channels, or to squeeze in more channels depending on how it is programed.

[lordsmurf]

or to squeeze in more channels depending on how it is programed.

Sadly, this is the common method of using open allocation. A few select channels definitely get a preferential treatment (HBO is always mentioned in these conversations), but most don't. The same is true of HD. Cartoon Network is a second-class citizen, with noticeable HD artifacting, while HBO gets max bandwidth even when the show is semi-static (Real Time With Bill Maher). Even PPV is no longer safe, aside from the big-bucks shows like WWE and UFC.

[nharikrishna]

Thanks both for the responses. I have also noticed some channels are 'okay' while other channels are crappy, full of compression artifacts.