soneric

I'm trying to write a patch that will capture direct to a hard drive from a DV camera, via firewire. I'll be recording people sitting and talking, so not too much motion. I'd like to get the highest picture quality possible, for later editing in FCP. I'm sort of piecing it together with the jit.qt.grab object, still there are questions.

1)Anybody have any suggestions of in terms of which codec to use? [The Jitter manual suggests motion-jpeg for playback within Jitter. Would this be enough, though for FCP?]

2)And how to maximize fps that Jitter will record with?

3)should the image size be 680 x 480 or 720 x 480?

In princible I understand this is quite simple, it's the details of course. any suggestions would be greatly appreciated.

I'm trying to write a patch that will capture direct to a hard drive from a DV camera, via firewire.  I'll be recording people sitting and talking, so not too much motion.  I'd like to get  the highest picture quality possible, for later editing in FCP.  I'm sort of piecing it together with the jit.qt.grab object, still there are questions.

1)Anybody have any suggestions of in terms of which codec to use?  [The Jitter manual suggests motion-jpeg for playback within Jitter.  Would this be enough, though for FCP?]

3)should the image size be 680 x 480 or 720 x 480?  

In princible I understand this is quite simple, it's the details of course.  any suggestions would be greatly appreciated.

direct-recording-to-disk

> I'm trying to write a patch that will capture direct to a hard 

> drive from a DV camera, via firewire. I'll be recording people 

> sitting and talking, so not too much motion. I'd like to get the 

> highest picture quality possible, for later editing in FCP. I'm 

> sort of piecing it together with the jit.qt.grab object, still 

> 1)Anybody have any suggestions of in terms of which codec to use? 

> [The Jitter manual suggests motion-jpeg for playback within 

> Jitter. Would this be enough, though for FCP?]

> 2)And how to maximize fps that Jitter will record with?

> 3)should the image size be 680 x 480 or 720 x 480?

As is always the case, you will never get any better stream than the 

source itself. Hence, I would recommend trying to keep as native as 

possible during capture. NTSC DV is 720x480. Use exactly DV NTSC 

witrh a resolution of 720x480 with this input to maximize fps and 

minimize any other potential compression artifacts. After the fact 

you can change as you see fit (like converting to 640x480 to 

compensate for non square pixels, or using the PhotoJPEG codec for 

> I'm trying to write a patch that will capture direct to a hard  
> drive from a DV camera, via firewire.  I'll be recording people  
> sitting and talking, so not too much motion.  I'd like to get  the  
> highest picture quality possible, for later editing in FCP.  I'm  
> sort of piecing it together with the jit.qt.grab object, still  
> there are questions.
>
> 1)Anybody have any suggestions of in terms of which codec to use?   
> [The Jitter manual suggests motion-jpeg for playback within  
> Jitter.  Would this be enough, though for FCP?]
>
> 2)And how to maximize fps that Jitter will record with?
>
> 3)should the image size be 680 x 480 or 720 x 480?

As is always the case, you will never get any better stream than the  
source itself. Hence, I would recommend trying to keep as native as  
possible during capture. NTSC DV is 720x480. Use exactly DV NTSC  
witrh a resolution of 720x480 with this input to maximize fps and  
minimize any other potential compression artifacts. After the fact  
you can change as you see fit (like converting to 640x480 to  
compensate for non square pixels, or using the PhotoJPEG codec for  
faster decompression).

> As is always the case, you will never get any better stream than the

> source itself. Hence, I would recommend trying to keep as native as

> possible during capture. NTSC DV is 720x480. Use exactly DV NTSC

> witrh a resolution of 720x480 with this input to maximize fps and

> minimize any other potential compression artifacts. After the fact

> you can change as you see fit (like converting to 640x480 to

> compensate for non square pixels, or using the PhotoJPEG codec for

by the way, I m having some troubles at the moment with setting the codec of

I m using a DVpal cam, and the codec seems to be locked on DVPAL which

> As is always the case, you will never get any better stream than the
> source itself. Hence, I would recommend trying to keep as native as
> possible during capture. NTSC DV is 720x480. Use exactly DV NTSC
> witrh a resolution of 720x480 with this input to maximize fps and
> minimize any other potential compression artifacts. After the fact
> you can change as you see fit (like converting to 640x480 to
> compensate for non square pixels, or using the PhotoJPEG codec for
> faster decompression).
>
> -Joshua

hi all,
by the way, I m having some troubles at the moment with setting the codec of
the jit.qt.grab grab_to_disk ability.
I m using a DVpal cam, and the codec seems to be locked on DVPAL which
takes too much room on the HDs.
have a look below.
I m running jit 1.52 max 4.57 & qt 7.01

direct recording to disk