DDogg
2nd February 2002, 17:13
Report to the group:
I was feeling a little crazy and wanted to play around with the new matrix support in 1.06 Build 3. I just wanted to see what the result would be if a full 2 hours was targeted for 1 800 cdr. The results were much better than I anticipated. More later on that.
Method used:
1>Used the latest tsunami patcher (0.2.1) to apply the "'Angel SVCD 1CD 2.35 avg800 max1500'" matrix to CCE 2.50SP
2> CCE was set up for 5 pass VBR, bilinear filter and Temporal Smoother 2,1, with I.Q.Priority 17, Anti-noise 6, Linear Q.S.=off, zigzag=on, progressive=on, Upper=on
3> Bitrate was set as average-800, max 1500 as the filter title suggested.
4> Audio was set for 96 Stereo with no downsampling.
Results: Quite good! I must say I was pleasantly surprised. My strictly subjective judgment using an older JVC 32 inch TV was the quality was equal or slightly better than a commercially rented VHS tape.
IMO, it was certainly better than a standard VCD by a good margin, but, as I said, this is just my perception and a very subjective one. For perspective, I would personally would rate a decent VCD encode about the quality of extended, long play VHS. Again, this encode was certainly equal to standard VHS if not a little better.
Note: Source was "Everafter" and was clean. Aspect was 4:3.
Bilinear filtering and Temporal smoother were used to increase the compressibility index as measured by GKnot. Index increased from .45 using SimpleResize to .62 using this combination.
Video was softer than original but not readily apparent when viewed on standard TV. No particular large blockiness or artifacts were observed. My 2 kids, 9 and 10, could not tell the difference between the standard SVCD encode and this one except to say "the standard was brighter".
I mention this because I would like to encode some movies for them for long trips in the van and would prefer one disk encodes for convenience. Their eyes are not trained, but they are keen. I wanted their input as part of a quick evaluation. The fact that they didn't moan, bitch and gripe told me a lot :).
They also made no comment on the audio. Frankly, I was very surprised with the audio quality given the extremely low bitrate of 96 Stereo. Great job DSPGURU!
All and all the result was very "watchable". I would have enjoyed the movie and not been too distracted by the quality at all. The fact that it was 4:3 and the quality was decent further surprised me.
Oh, to state the obvious, I am sure this would not look very good on HQ sets so PLEASE don't even .... :) This example is only given as part of our knowledgebase. I am not advocating people do this or use this for anything except an unusual situation where someone needs to tradeoff quality for compactness. This is just presented to point out extended encodes are possible, and can yield reasonably decent results, when needed.
Script:
LoadPlugin("C:\PROGRA~1\DVD2SVCD\MPEG2Dec\mpeg2dec.dll")
mpeg2source("D:\TEMP_T~4\DVD2AV~1.D2V")
TemporalSmoother(2,1)
BilinearResize(480,480)
I was feeling a little crazy and wanted to play around with the new matrix support in 1.06 Build 3. I just wanted to see what the result would be if a full 2 hours was targeted for 1 800 cdr. The results were much better than I anticipated. More later on that.
Method used:
1>Used the latest tsunami patcher (0.2.1) to apply the "'Angel SVCD 1CD 2.35 avg800 max1500'" matrix to CCE 2.50SP
2> CCE was set up for 5 pass VBR, bilinear filter and Temporal Smoother 2,1, with I.Q.Priority 17, Anti-noise 6, Linear Q.S.=off, zigzag=on, progressive=on, Upper=on
3> Bitrate was set as average-800, max 1500 as the filter title suggested.
4> Audio was set for 96 Stereo with no downsampling.
Results: Quite good! I must say I was pleasantly surprised. My strictly subjective judgment using an older JVC 32 inch TV was the quality was equal or slightly better than a commercially rented VHS tape.
IMO, it was certainly better than a standard VCD by a good margin, but, as I said, this is just my perception and a very subjective one. For perspective, I would personally would rate a decent VCD encode about the quality of extended, long play VHS. Again, this encode was certainly equal to standard VHS if not a little better.
Note: Source was "Everafter" and was clean. Aspect was 4:3.
Bilinear filtering and Temporal smoother were used to increase the compressibility index as measured by GKnot. Index increased from .45 using SimpleResize to .62 using this combination.
Video was softer than original but not readily apparent when viewed on standard TV. No particular large blockiness or artifacts were observed. My 2 kids, 9 and 10, could not tell the difference between the standard SVCD encode and this one except to say "the standard was brighter".
I mention this because I would like to encode some movies for them for long trips in the van and would prefer one disk encodes for convenience. Their eyes are not trained, but they are keen. I wanted their input as part of a quick evaluation. The fact that they didn't moan, bitch and gripe told me a lot :).
They also made no comment on the audio. Frankly, I was very surprised with the audio quality given the extremely low bitrate of 96 Stereo. Great job DSPGURU!
All and all the result was very "watchable". I would have enjoyed the movie and not been too distracted by the quality at all. The fact that it was 4:3 and the quality was decent further surprised me.
Oh, to state the obvious, I am sure this would not look very good on HQ sets so PLEASE don't even .... :) This example is only given as part of our knowledgebase. I am not advocating people do this or use this for anything except an unusual situation where someone needs to tradeoff quality for compactness. This is just presented to point out extended encodes are possible, and can yield reasonably decent results, when needed.
Script:
LoadPlugin("C:\PROGRA~1\DVD2SVCD\MPEG2Dec\mpeg2dec.dll")
mpeg2source("D:\TEMP_T~4\DVD2AV~1.D2V")
TemporalSmoother(2,1)
BilinearResize(480,480)