Live Video Input Compression Settings



  • Hi, apologies if this has been answered elsewhere --

    Are there best practices for speed when selecting the Compression codec for live video input?   Since Isadora is optimized for Photo-JPEG manipulation, does it make sense to set the live capture compression to Photo-JPEG?
    Is the Y'CbCr 4:2:2 yuyv setting the default for a reason (as in, is it the best choice for some specific reason)?
    I assume there are more performance hits for some of the compression settings, but how do they relate to benefits when manipulating that video stream within a scene?
    thanks for your wisdom, forum!
    David


  • Dear David,

    You should never modify the default compression setting for the Live Video Input on Mac OS. The default is always the best; choosing a different codec will only impair performance.

    Best Wishes,
    Mark



  • Thanks Mark,

    That's good to know!
    -David


  • THis is somehow something tricky for me. When I used pal or ntsc camera through a firewire digitizer (Canopus ADVC 110). I immediately go to -> Live capture settings -> video input Settings and an Apple window opens (I believe it is not Isadora but Quicktime am I right?)

    The canopus choses DV PAL but the image is interlaced.
    Always in the same window i can choose progressive. But nothing happens. And if I change the aspect ration too. Nothing changes in the image.
    So I change to codec to uncompressed 8 bit 4:2:2 and the image has no more interlacing.... Strange.And pannings are smoother.
    But a stranger thing happens with the Blackmagic Intensity shuttle thunderbolt I bought. if in the same live input settings window I chose the original Blackmagic RGB 10 bits codec the image is not very good, cycles are  moving (don't know why) between 70 and 150, 30 fps, and a VPO is 15\. f I use the uncompressed ! bits I have VPO 22, 50 fps and cycles moving (don't know why) between 150 and 240.
    So in my experience it is better to override choices automatically made by live video input settings.... Am I right?


  • I'm having a bit of trouble with what looks like jagged interlacing issues coming from a Sony HXR-NX5U set to capture at 480i into a blackmagic intensity pro via composite cable going into Isadora and out to a Panasonic PT-DW6300. I've got signal, but again it's giving interlaced images when outputting - fine when the subject's still, but jagged when it's in motion.

    We've tried to keep the image in NTSC, 29.97 along the chain. The computer's a Mac Pro - 2.8 Ghz Quad-core Intel Xenon; 16 GB 1066 MHz DDR3 ECC, ATI Radeon HD 5770 1024 MB. 
    Any thoughts would be mentioned. 
    Thanks, Alex ![](http://troikatronix.com/troikatronixforum/uploads/FileUpload/d9/e41ab3f7670f9bd5a3c90370c7e1e0.png)


  • Hi all,
    The only way I found to get deinterlaced signal from Blackmagic intensity shuttle USB3.0 (on mac), is to enable input processing (last box in the blackmagic settings window) to "SD to 16:9 / 50fps". You get an 720p upscaled progressive video. I know it's more data for Isadora to process, but that's the fastest and smoothest results I have archieved.

    Best
    Philippe



  • Dear Alex,

    I actually don't know the Black Magic as well I probably should, but wouldn't choosing NTSC mean that for sure you're going to have interlacing, no matter what you do? NTSC is an interlaced standard -- thus the 'i' in 480i. There wouldn't seem to be a way around that, would there? Or am I missing something?
    Best Wishes,
    Mark


  • Hi Mark,

    480i is indeed interlaced - we had to keep in on interlaced because the camera is outputting SD via the composite cable, and we don't have a digital way in to the blackmagic because of the cable length. So we figured we'd keep the entire chain on 480i to keep the signal as clean as possible. 
    We're still troubleshooting, but one question - in the preferences menue>Video Image Processing>Default Resolution, we've set the H to be 1280 and the vertical to be 720, which is the native resolution of the projector. If we do this, the processing is done at 720i or 720P, or is there any interlacing which may be influencing the artifacting?
    Thanks,
    Alex


  • Dear Alex,

    I updated the section on the "Default Resolution" in the manual to make its function more clear. Here's what it now says.
    The Default Resolution determines the resolution of a video stream when no other reference is available. There are two general situations when the Default Resolution setting comes into play.
    First, when an actor generates video and has no video inputs to determine what the resolution should be, e.g., the Video Noise actor. For these actors, the output resolution will always be forced to the Default Resolution setting.
    Second, when actors with one or more video inputs have one of those inputs disconnected, e.g., the Text Draw actor or Video Mixer actor. For these actors, a black frame will be generated at the Default Resolution to provide a “signal” for the missing inputs.
    To make this more clear, we’ve provide two examples of the second situation mentioned above. For purposes of these examples, we'll assume the Default Resolution is set to 320x240.
    First, let’s say you have a Text Draw actor. Its 'video' input (which allows the text to be drawn on top of another video stream) is disconnected. In this case, the output resolution of the Text Draw actor would be set to the Default Resolution of 320x240, because the black frame replacing the missing video input will have a resolution of 320x240.
    Second, let’s consider a Video Mixer actor. The first video input is connected to a video stream with a resolution of 1920x1080\. The second is not connected. Since there is no input to the second video input, a black frame with is generated at the Default Resolution of 320x240\. Now the Video Mixer must scale one of the two videos (the 1920x1080 input and the 320x240 black frame) so that their resolutions match.
    How this scaling is performed in this situation is controlled by the When Combining Video popup menu described in the next section. If this parameter is set to:
    • Scale To Smallest (Faster) then the 1920x1080 video will be scaled down to 320x240 and the output will be 320x240.
    • Scale To Largest (Slow) then the 320x240 video will be scaled up to 1920x1080 and the output will be 1920x1080.
    • Scale to Default Resolution then both inputs are scaled to the Default Resolution of 320x240. The black frame is already 320x240, so it won’t be scaled. But the 1920x1080 video will be scaled down to 320x240 and the output will also be 320x240.
    If, on the other hand, you had both video inputs of the Video Mixer actor connected – and if their resolutions were different from each other – then only the "When Combining Video" setting is used when deciding how the videos are scaled. In this case, the Default Resolution would only come into play is if When Combining Video popup menu is set to Scale to Default Resolution.
    Does this explain things for you?
    Best Wishes,
    Mark

Log in to reply
 

Looks like your connection to TroikaTronix Community Forum was lost, please wait while we try to reconnect.