[Chris Wright]

Digieffects Fantazm-Bad Codec is exactly that.
download the demo here.
https://www.digieffects.com/products/fantazm.html

[Chris Wright]

is he using cineon, 32 bit log, an alpha channel, or color management perhaps?

[Chris Wright]

I think you’re looking for reelsmart’s Re-interlacer that fixes the problems you are discussing.

< Half Frame Rate: This option takes each 2 pairs frames from the source and makes 1 frame by taking the field scanlines specified in Fields Order from the first frame of the pair, and the field to be displayed second from the second of the pair. This option assumes that your source material is at one rate (say 59.95 fps) and the destination is at half the frame rate (in this case, 29.94 fps).This option only works in After Effects, Combustion 2.1 or later and Commotion 3.1 or later. • < Keep Frame Rate: This option takes each 2 pairs frames from the source and makes 1 frame by taking the field scanlines specified in Fields Order from the first frame of the pair, and the field to be displayed second from the second of the pair. However, this option makes the destination 1/2 the duration (so 4 seconds of input becomes 2 seconds of output). • = Field Blending : Field Blending is mixing two fields from pairs of images in a sequence. This option neither creates nor destroys frames, but rahter, blends fields of adjoining pairs of frames in a way that may be less-flickery on an interlaced monitor. This is a useful filter when your source material has small animated things (like stars) or horizontal lines that move about an image. Note this is a destructive process in the sense that you will not be able to deinterlace that render to recuperate the exact original source. • = Combine: Say you have two sequences of whole frames aligned in time. One sequence if of whole frames that should represent the first field of the output and the other, the second field. Apply Reinterlacer to the sequence representing the first fields, set the source to the other sequence that is to make up the second fields of the output. • = Create motion estimated fields: This option will convert progressive material to interlaced material using motion estimation to create new fields between each original frame. This mode looks at the current frame and the next frame and estimates a new field at the inbetween position. Looking at the current frame and the next frame can create some issues at a cut point in your sequence. In applications that supports animation of menu options, you might need set a keyframe at the beginning of a cut. This mode is helpful to get a consistent look when merging material without fields and material with fields (or simply generate a field based render in a progressive project…) .

[Chris Wright]

The PhotoJPEG codec can’t do lossless – even at 100% quality it is a lossy codec. At 100% quality only it is a 4:4:4 codec and at 99% and below it is a 4:2:2 codec.

PhotoJPEG at 100% is 4:4:4 lossy
PhotoJPEG at 99% and below is 4:2:2 lossy
Jpeg2000 at 100% is lossless (but uses lots of CPU power)
Animation at 100% is lossless 8 bit

PNG is lossless too btw, but in the quicktime container it handles only 8bpc too.
PNG at 16bpc can be good for images sequences.

I compared TGA, PNG, and TIff to PSD and only Tiff(lzw off) at 3200% was flawless in comparison in RGB color picker values.

TGA, PNG, and Tiff (only with LZW compression on) seem to have strange micro pixel color changes at 3200%. I’m guessing LZW has coding precision rounding errors in the area of a millionth of a rgb value.

PSD and Tiff(LZW off) are perfectly lossless but since Tiff is half the size, that’s what I use.

I would love to hear other people’s roundups of video codecs.

[Chris Wright]

What you’re looking for is mathematically lossless but smaller file size. This is where math compression comes in. 3 of the best are (8 bit run length encoding) animation, (8 bit wavelet) jpeg2000, and Tiff (lzw off). Now if you change histograms a lot, then use a 10 bit encoder to minimize the chance of posterations (gradients of missing color usually from heavy redundant color correction) And if you do have sky gradients of color, working in a 16 bpc workspace will reduce precision errors.

[Chris Wright]

please read the last 15 posts below so it is clear.
https://forums.creativecow.net/thread/2/919924

If you have not found a solution to the problem yet, here you go:

My guess is all of you have an AJA Io or Card, right…? If I did not open a project that rendered fine prior to getting an AJA Io HD I would never have deducted this. But my suite worked fine with round-trip DVCPRO HD & FCP before getting the Io. And checked the same clip on the other suite without Io – and that worked fine too.

Spoke to one of the very helpful guys at AJA and he suggested removing a codec that installs with the A Io control panel. root/ library/ Quicktime/ AJAUncompressed

I don’t need that codec. And uninstalled it. Voila! AE is working fine again with DVCPRO HD. I can live again.
—
the output from my Kona is displaying an incorrect gamma while using video preview from within AE. I looked at the control panel for the Kona and played around with the “Codec Options”. I change the RGB-YUV conversion GAMMA from AUTO to REC 601 (2.20) which did solve the output to the monitor, but this is not intuitive.

https://forums.creativecow.net/thread/2/919924

[Chris Wright]

there’s time-echo which creates frame echoes of previous movement then there’s reelsmart motion blur which creates natural looking camera slow shutter effects. Depends on what look you’re going for.

[Chris Wright]

Reelsmart motion blur has a remove blur vector setting that I have tested. It will remove a little bit of blur from slow shutters. Download the demo and use main blur ammount -2 (negative 2) and main background sensitivity 100. It won’t work wonders but will enhance a little. Think sharpen effect in photoshop but as an intelligent sharpen.

[Chris Wright]

quicktime’s jpeg2000 at 100% is also 8 bit lossless, alpha support, like animation but 1/2 the size.
Its standard is used in digital projection equipment.

There are also other kinds like jpeg2000 sequencing that can go 10 bit or higher, but many techies and I agree that Tiff is best for image sequencing. It’s thread, multi-proc safe, and with lzw compression off, is the smallest size for lossless.

[Chris Wright]

AE echo effect?