Use 720p or 1080p source for DVD?

[am_dew]

I have two sources available to me...one 720p and one 1080p. Assuming that the same encoding settings would be used for both, is there any benefit (video quality wise) to using the 1080p source over the 720p source when encoding to a DVD compliant MPEG-2 since the 1080p source would need to be reduced to 720 anyway? I will be using TMPGenc Xpress 4 to do the encoding. Here is MediaInfo for the two sources (I took out the Audio sections).

Thanks!

720p
Format : MPEG-4
Format profile : Base Media / Version 2
Codec ID : mp42
File size : 1.37 GiB
Duration : 1h 25mn
Overall bit rate : 2 305 Kbps
Encoded date : UTC 2011-10-12 09:48:23
Tagged date : UTC 2011-10-12 09:48:23
gsst : 8008
gstd : 5108908
gssd : BD075F383MH1321595706710323
gshh : tc.v11.cache7.c.youtube.com

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L3.1
Format settings, CABAC : Yes
Format settings, ReFrames : 1 frame
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Duration : 1h 25mn
Bit rate mode : Variable
Bit rate : 2 541 Kbps
Maximum bit rate : 5 301 Kbps
Width : 1 280 pixels
Height : 720 pixels
Display aspect ratio : 16:9
Frame rate mode : Variable
Frame rate : 29.970 fps
Minimum frame rate : 29.412 fps
Maximum frame rate : 30.303 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.092
Stream size : 1.51 GiB
Tagged date : UTC 2011-10-12 09:54:43

1080p
Format : MPEG-4
Format profile : Base Media / Version 2
Codec ID : mp42
File size : 3.24 GiB
Duration : 1h 25mn
Overall bit rate : 5 449 Kbps
Encoded date : UTC 2011-10-12 09:53:11
Tagged date : UTC 2011-10-12 09:53:11
gsst : 8008
gstd : 5108908
gssd : BD075F301HH1321595266094136
gshh : tc.v9.cache3.c.youtube.com

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.0
Format settings, CABAC : Yes
Format settings, ReFrames : 1 frame
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Duration : 1h 25mn
Bit rate mode : Variable
Bit rate : 5 291 Kbps
Maximum bit rate : 9 246 Kbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate mode : Variable
Frame rate : 29.970 fps
Minimum frame rate : 29.412 fps
Maximum frame rate : 30.303 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.085
Stream size : 3.14 GiB (97%)
Tagged date : UTC 2011-10-12 10:03:49

[poisondeathray]

Actually it's reduced to 480 pixel height for NTSC DVD-video, not 720

Probably not a lot of difference in the end result from these youtube sources, assuming they were made the same way (impossible to judge from numbers alone).

Encode a short sample and find out

[fritzi93]

Neither one has, hmm, an over-generous bitrate for the resolution. I would guess the 720p source might be somewhat better, but it's only a guess.

[am_dew]

Thanks. Yea, I know these YouTube "HD" sources are not exactly top-notch quality but they do turn into pretty nice looking DVDs believe it or not. I'll do some experimenting.

[jagabo]

Artifacts from over compression will be the same size (8x8 blocks, DCT ringing). Reducing the 1920x1080 video down to 720x480 will reduce those more than reducing the 1280x720 video (assuming they have similar amounts of artifacts). Another issue is which has the better quality to start with. Guessing that the 720p video was made from the 1080p video it has already been downscaled once, losing some quality and creating some scaling artifacts. A second downscaling to 720x480 will introduce more scaling artifacts. A single downcale of the original 1920x1080 source to 720x480 will introduce fewer scaling artifacts (depending to some extent on the algorithms used). Then you have the bitrate issue. The 1080p source has twice the bitrate of the 720p source. That's probably sufficient to keep the 1080p from having more blocking and ringing artifacts -- assuming they were both made from a high quality source.

Basically, after downscaling, the 1080p source will probably look a bit better than the 720p source. But the only way to say for sure is to try them both.

[smitbret]

Well, I would assume that the 1080p was the original source and the 720p was a re-encode of that source. If that's the case, you'd have better quality (measureable or unmeasureable) by using the original 1080p source. Everytime a stream is encoded, you lose information, so the jump from 1080p->480p will be better than 1080p->720p->480p.

[am_dew]

Well, I would assume that the 1080p was the original source and the 720p was a re-encode of that source. If that's the case, you'd have better quality (measureable or unmeasureable) by using the original 1080p source. Everytime a stream is encoded, you lose information, so the jump from 1080p->480p will be better than 1080p->720p->480p.

This makes a lot of sense. Thanks to everyone for your feedback!

[jagabo]

Everything downloaded from Youtube has been reencoded by Youtube. So even the 1080p download is not the original file that was uploaded.

[fritzi93]

Everything downloaded from Youtube has been reencoded by Youtube. So even the 1080p download is not the original file that was uploaded.

That was my assumption, that both were re-encoded from the same source file. And while the bitrates of both are inadequate, 720p at a bitrate of 2,541 is a little less inadequate than 1080 at 5,291.

Or am I leaving something out of account?

[jagabo]

while the bitrates of both are inadequate, 720p at a bitrate of 2,541 is a little less inadequate than 1080 at 5,291.

I think it's pretty even. 44 percent as many pixels and 48 percent as much bitrate. I pointed out some other issues to consider earlier.

[am_dew]

while the bitrates of both are inadequate, 720p at a bitrate of 2,541 is a little less inadequate than 1080 at 5,291.

I think it's pretty even. 44 percent as many pixels and 48 percent as much bitrate. I pointed out some other issues to consider earlier.

Thanks. Now I'm just waiting for my new system to arrive so I can do the encoding on it instead of my old AMD Athlon XP 2400 based system. Looking forward to large encodes like this taking much less time than on my current system!

[fritzi93]

So what are you getting? When I built my 6-core machine, it was like night and day coming from an oldish dual-core computer. I could wish to have done it sooner, but I built it fairly cheaply.

Good luck.

[am_dew]

I'm getting an Intel i5-2500K installed on an ASUS P8Z68-V Pro board with 8GB DDR3-1600 RAM in an Antec 300 case.

[fritzi93]

I'm getting an Intel i5-2500K installed on an ASUS P8Z68-V Pro board with 8GB DDR3-1600 RAM in an Antec 300 case.

You're gonna love it.

That was the other processor I considered for my new computer. Since AM3 mobos are less expensive (and the processor as well), I went with the AMD build. From what I could gather, both processors (2500k and 1090t) are pretty comparable for H264 encoding, which was my main requirement. The 2500k beats the 1090t in a lot of the other benchmarks, though, IIRC.