720/60p or 1080/60i??
- Thread starter iFilm7
- Start date
Bassman2003
Veteran
Well in the case of Premiere Pro CS4, it definitely uses a different algorithm for display in program monitor/source monitor vs. what it uses for sequence export . It even says so in the manual, it uses a bilinear resize for the preview, but a cubic resize for the export. So what you see is NOT what you get, even if you set preview quality to maximum instead of draft.
I'm not too familiar with Edius , or what algorithm it uses, but I would double check the exported footage (e.g. do a test with uncompressed AVI) to see that what you export is the same as what you see in the preview monitor. It's different in some NLE's for example, like Vegas and Premiere (see above), what you see isn't necessarily what you get in the actual exported file.
But to answer your question, yes there are other software to access different resize methods. An example would be avisynth, or virtualdub. Premiere has an .avs import plugin, so you can use avisynth directly in Premiere, for example. Not sure if Edius has something similar
Does the resize algorithm make a difference in practice? You bet it does. I've included below a zip file of a bilinear resize vs. a spline36 resize
http://www.mediafire.com/?mjjjncgmkdm
Bassman2003
Veteran
I don't know about the algorithm, but I have a Canopus video output card which is supposed to be frame accurate broadcast ouput. So I trust the monitor.
But your point about the dedicated software resize is true as many people in the Edius forum use Virtual dub for downsampling.
Why the NLE's don't offer a high quality downsample is beyond me. How does a free software beat an NLE? Crazy if you ask me.
But your point about the dedicated software resize is true as many people in the Edius forum use Virtual dub for downsampling.
Why the NLE's don't offer a high quality downsample is beyond me. How does a free software beat an NLE? Crazy if you ask me.
That may be the case, but note most display software uses bilinear resizing for display because of low overhead. (the other algorithms are slower to process). This is true of virtually all software players. So the question is, what is doing the actual resize in Edius? Is it Edius itself or the hardware card? If it's Edius, chances are you are using it's internal resizing, not the hardware card to resize.
I still suggest you compare actual exported footage as suggested above. It might be easier to run some test charts which would be easier to see the differences. It certainly doesn't hurt to learn the truth or any potential quirks about your workflow
I find that all software have little quirks or bugs - free , opensource, or paid. I think it's important to find out what these are for your particular workflow , so that you can work around them.
Bassman2003
Veteran
Thanks for the download. There is a detail difference.
Which program was used for this downsample?
Is there one you think has the best output & ease of use?
Which program was used for this downsample?
Is there one you think has the best output & ease of use?
I used avisynth to resize for that example. (Note the footage wasn't from HMC150, but from an Arriflex, but it was just to demonstrate that there are visible differences in resize algorithms)
If you say some people in the Edius forum use virtualdub, then that maybe the way to go. Internally, it has Lanczos Resize available, which is very similar to Spline36. So you might use vdub to resize to 720p and encode to a cineform .avi intermediate, then import into Edius. (or maybe canopus' intermediate format, i forget what it's called)
Sorry, I can't suggest anything more specific, because I'm not familiar with Edius' workflow
Bassman2003
Veteran
Thanks for your help. I will check them both out.