Using Proxy Files In Previous Versions

The Shadowman

Senior Contributor Location: UK Joined: Dec 15, 2014 13:06 Messages: 1831 Offline

Solved by

Feb 02, 2017 10:20

I think it's pretty much accepted that PD15 has cracked the problem of slow to generate shadow files see http://forum.cyberlink.com/forum/posts/list/51146.page. On reading the link thread you'll note that not only PD15 has a very good score here, but PD14 is not too far behind. However, what about those of us who want to edit 4K and similar, but still have earlier versions of PD?

We can endure the long wait for shadow files to generate, or we can download something like Magic+PD. With that in mind I thought I would have a look at both methods to see the advantages or otherwise of shadow files and Magic+PD.

For the experiment I used the same 4 clips I used in the post I reported in the above link. I thought the results would interest you.

PD13 - intel HD4600 i7 (no dedicated card) 16GB RAM

Shadow Files:

4 clips equalling a total of 2 min 21 secs. 3840x2160 25.

Time to generate shadow files 20 min 14 secs

Insert three transitions - pass

Add 4 effects - Neutral density, Mosaic, Mirror, Moonlight. (one to each clip) - pass

Add to clip 1 - video in reverse - pass

add to clip 2 - Rotate 180 - pass

add to clip 3 - increase speed x4 - pass

add to clip 4 - video in reverse - Fail (system memory warning from PD)

After the failure I reduced the preview down to Non Real Time and attempted to continue. However, timeline was so jerky it was impossible to edit further.

Magic+PD

Time to generate proxy files 1 min 42 secs

Again, same 4 clips equalling a total of 2 min 21 secs. 3840x2160 25.

Add 4 effects - Neutral density, Mosaic, Mirror, Moonlight. (one to each clip) - pass

Add to clip 1 - video in reverse - pass

add to clip 2 - Rotate 180 - pass

add to clip 3 - increase speed x4 - pass

add to clip 4 - video in reverse Fail (system memory warning from PD)

Once again, after the failure I changed the preview to Non Real Time, and here's where the difference really shows.

I was able to add and edit the following without a single jerk on the timeline.

With splits I ended up with 15 separate clips edited as folows:

Clip 1 - Reverse video

Clip 2 - Reverse video

Clip 3 - Reverse + speed minus .4 +Rotate 180

Clip 4 - Speed x 5

Clip 5 - Rotate 180

Clip 6 - Reverse video+ Crop + rotate 180

Clip 7 - Speed minus .9 + Rotate 180

Clip 8 - Reverse video + speed minus .9 + Rotate 180

Clip 9 - Speed x 4.10

clip 10 - Reverse video

Clip 11 - Reverse video

Clip 12 - Reverse video + Speed minus .26

Clip 13 - Reverse video + speed minus .26

Clip 14 - Reverse video

Clip 15 - Speed x9

Each of the 15 clips was separated with a transition and had an effect included. With increases/decreases in speed the final project was 1min 41 secs long. At this point I stopped the test and Magic still hadn't failed, I wonder how much further I could have gone.

Conclusion

Clearly in Normal preview mode, Shadow files and Magic+PD are evenly matched except in terms of generation time. But as you will note from the above test, Magic is way ahead when the preview mode is changed to non real time. With splits, the original 4 clips were turned into 15, and all were given a heavy load of Power Tool assignment and in most case more than one. There was still no jerking on the time line, this surpised me greatly.

Given the above findings, and if you don't mind editing with no audio (non real time) then Magic is going to be better for you.

I hope that you find this result interesting

Robert

This message was edited 5 times. Last update was at Feb 02. 2017 10:29

Panny TM10, GH2, GH4,

JL_JL

Senior Contributor Location: Arizona, USA Joined: Oct 01, 2006 20:01 Messages: 6091 Offline

Feb 02, 2017 10:35

If you could, post your 4 clips and the 15 clip pds file to a shared area.

Jeff

The Shadowman

Senior Contributor Location: UK Joined: Dec 15, 2014 13:06 Messages: 1831 Offline

Feb 02, 2017 11:07

Quote If you could, post your 4 clips and the 15 clip pds file to a shared area.

Jeff

I have uploaded the PDS file here https://www.dropbox.com/s/6682dcyd4c00irz/shadowtest2m%2B.pds?dl=0 but if you want the originals it will take ages as the AVI files are very large

This message was edited 1 time. Last update was at Feb 02. 2017 11:11

Panny TM10, GH2, GH4,

JL_JL

Senior Contributor Location: Arizona, USA Joined: Oct 01, 2006 20:01 Messages: 6091 Offline

Feb 02, 2017 11:20

Quote

Quote If you could, post your 4 clips and the 15 clip pds file to a shared area.

Jeff

I have uploaded the PDS file here https://www.dropbox.com/s/6682dcyd4c00irz/shadowtest2m%2B.pds?dl=0 but if you want the originals it will take ages as the AVI files are very large

Your original 4 clips shouldn't be too bad unless you are bandwidth constrained, not the 10x increased in size Magic AVI's, I've got no interest in those. I don't believe I saw any place you posted any details of those 4 clips, just 4k so unsure how big they might be. Even if 50Mbps under 1GB in total size so that's not too bad.

Jeff

The Shadowman

Senior Contributor Location: UK Joined: Dec 15, 2014 13:06 Messages: 1831 Offline

Feb 02, 2017 11:31

Quote

Quote If you could, post your 4 clips and the 15 clip pds file to a shared area.

Jeff

I have uploaded the PDS file here https://www.dropbox.com/s/6682dcyd4c00irz/shadowtest2m%2B.pds?dl=0 but if you want the originals it will take ages as the AVI files are very large

Your original 4 clips shouldn't be too bad unless you are bandwidth constrained, not the 10x increased in size Magic AVI's, I've got no interest in those. I don't believe I saw any place you posted any details of those 4 clips, just 4k so unsure how big they might be. Even if 50Mbps under 1GB in total size so that's not too bad.

Jeff

OK, I thought you wanted the big ones Panny TM10, GH2, GH4,

The Shadowman

Senior Contributor Location: UK Joined: Dec 15, 2014 13:06 Messages: 1831 Offline

Feb 02, 2017 11:41

Quote

Quote If you could, post your 4 clips and the 15 clip pds file to a shared area.

Jeff

I have uploaded the PDS file here https://www.dropbox.com/s/6682dcyd4c00irz/shadowtest2m%2B.pds?dl=0 but if you want the originals it will take ages as the AVI files are very large

Your original 4 clips shouldn't be too bad unless you are bandwidth constrained, not the 10x increased in size Magic AVI's, I've got no interest in those. I don't believe I saw any place you posted any details of those 4 clips, just 4k so unsure how big they might be. Even if 50Mbps under 1GB in total size so that's not too bad.

Jeff

OK, I thought you wanted the big ones

Jeff

Am I doing something wrong?

The 4 clips are estimated at in excess of 6 hours, so I zipped them for you, and it's still 4 hours. Is that right?

Robert Panny TM10, GH2, GH4,

JL_JL

Senior Contributor Location: Arizona, USA Joined: Oct 01, 2006 20:01 Messages: 6091 Offline

Feb 02, 2017 18:49

Your upload speed simply depends on your bandwidth and the bandwidth capability of the upload store. For me ~1GB would take a few minutes but obviously will very for everyone based on there particular bandwidth.

The video in reverse comment or massive speedup playback benefit makes total sense because your AVI file is entirely "B frames", that's why using a intermediate codec can be beneficial. Think of the video as a deck of cards. Every card is a complete image on it's own, i.e. a B frame. Now take any compressed video (MPEG, H.264, ...)deck of cards, it has IBP frames, so now I don't know what the 10th frame (card) looks like as it's not just the 10th card, I need to uncompress the prior frames to figure it out, they are dependent on the IBP structure. So when one speeds things up it becomes very intensive to do real-time with compressed video. With uncompressed video, it's simply a data access rate issue, not a real challenge for any current HD, SSD, M.2. The decode computational effort becomes much easier with MPEG2, and even much easier with hardware (GPU) decode when supported, most likely why PD shadow files are always MPEG2, not simply lower quality H.264 for instance.

Jeff