Oh wait, I see now that Nathan said that he saw the error from input wells through FxParameterRetrievalAPI. And that the "main input" works okay (which I now realize is the input image passed to effects by FCP when it calls –renderOutput)
So I need to be clear here: I equate "main input" with the whole clip and not just the input image that I passed to me when FCP calls renderOutput. And I'm trying to get images at other places in time, so when I say I see this bug when retrieving frames on the "main input", I mean "main input CLIP using FxTemporalImageAPI::getInputBitmap to get other images in time .
It appears that the same bug that Nathan sees in FxParameterRetrievalAPI::getBitmap also appears when I try to get other frames in time from the clip the effect is applied to (what I call the main input), when I use FxTemporalImageAPI::getInputBitmap.
Darrin: I presume this is why are not seeing the resolution bug, because I didn't make it clear that I'm seeing this when I call FxTemporalImageAPI::getInputBitmap (and like Nathan, when I call FxParameterRetrievalAPI::getBitmap )?
We absolutely need to get other frames in time without resizing, because we are writing a retimed and do not want to (a) do extra processing and (b) we don't know what kind of scale up you are doing (I asked that before,… Darrin, can you answer?) in order to reduce back down. Even if we did know the scale up method, I'm not sure how we can use that info, because we can't tell when the scale reported to us is due to a template change in size or a viewer change in size.
So any bigger-sized all encompassing template we create will only allow us to get proper images at the template-sized images. Otherwise we either have to live with softened images (which is not an option), or force users to use templates that are resolution-specific (a tech support nightmare, let alone requiring us to create templates for every resolution we might see, which is also highly improbable that we'll get them all)
So this is a real problem for us.
Pete
From: Nathan Weston < email@hidden> Date: Fri, 29 Jul 2011 15:09:42 -0400 To: Peter Litwinowicz < email@hidden> Subject: Re: Resolution
On 07/29/2011 02:11 PM, Peter Litwinowicz wrote: seeing a similar issue here, but it seems to only affect images
retrieved from input wells through the FxParameterRetrievalAPI -- the main input is coming out unscaled.
Argh. That is not what we are seeing. Are you using GUP render? Perhaps the bug I am seeing only happens in software rendering? Pete
We're using software rendering as well.
|