[Dan Sakols]

Hi, Im having the same problem and about to give up. I have a 30 min sequence I want to apply a slight bit of stabilization to. after analyzing, the crop (or scale) gets very big due to a few sections. is there a way that WARP can limit its stabilization to prevent this radical zoom in (130%)? this is a pain. ive set scale >> stabilize to 5%.

[Dan Sakols]

QT Pro has some nice capabilities, its convenient, but from what I see, doesn’t support High Profile, just Main, so many of the advanced AVC tools like CABAC aren’t supported, and therefore the compression efficiency is less. Id be happy to learn I am wrong. check it out.

d

[Dan Sakols]

Ryan, is there an adaptive mode where the encoder will deviate from the one keyframe/sec rule for lots of random motion or a talking head? Else 1 keyframe/sec sounds reasonable. I always multipass encode.

[Dan Sakols]

Im very interested to hear how this plays out. Im just moving off FCP 7 because of frusteration with smoothcam, and plug ins like Mercalli and Lock and Load. Preliminary tests with PP and Warp look really promising…id freak out too if stuff starts jumping around or clips get radically zoomed in even for a light touch of stabilization.

dan

[Dan Sakols]

Ryan,

Im looking at the dialog box for the sequence settings, where it specifies how previews are rendered as MPEG I frames. You answered my question tho, that PP will decode the B and P frames depending on where my play head is, and i will see the same quality as the AVC in the sequence.

After looking at Adobe Media Encoder, it looks pretty good from settings and control. Much better than Compressor 3 which I abandoned for Handbrake (I never moved off FCP 7) I still need to do some tests, and will compare perceptual quality and bitrate with Handbrake.

Im shooting at 30fps. for encodes I force an I frame every 24 frames as a rule of thumb. does that sound about right?

dan

[Dan Sakols]

Hey Ryan,
I still haven’t got my head around the preview settings under “segment settings” If I want a high quality preivew to evaluate adjustments, like color or an effect, should I set the preview to the same settings as the segment? Im not sure how MPEG I frame, (the setting I see now) is usefull, as those should only occur about once a second. BTW, I took a look at your site- some very impressive work!

Thanks

dan

[Dan Sakols]

The reason Im thinking about this (perhaps wrongly) is that I want to avoid quantization in intermediate decode/encode cycles in rendering. From this thread I have learned that cuts and transitions will not matter (transitions don’t need to be rendered?) but composting and color correction will. if that’s the case- I frequently tweak the lighting in certain clips, it sounds like it would make sense to convert my camera clips to ProRes before going to work on it.

Im just an enthusiast making fun videos, but I know alot about compression technology, and want to apply some best practices. It certainly would be more convenient just to keep it in H.264 and then do a final High Profile, exhaustive motion vector search export. Am I on the right track here?
dan

[Dan Sakols]

Im not actually a video production professional, but do work with the underlying technologies for media streaming. Im very interested in your inputs.

What about sequence formats. Im not sure if I should actually edit in the same format as the AVC source from my DSLR or in a less compressed format.

dan

[Dan Sakols]

Ryan,
don’t knock Handbrake. Its a very sophisticated encoder that exposes alot of control Ive been able to get far better compression for a give file size than FC Compressor and QT Pro when you tune the encode settings. I have not looked at Adobe stuff in any detail yet.

In the event I were to use PP for my final encode, would it be good practice to edit in the source clip format or is it best edit in a less lossy format and then transcode back to AVC?

thanks
dan