Rendering HD video Pixel Format 8-bit or 32-bit? Deinterlace Blend or None?

[c627627]

When rendering HD video using Sony Vegas, what are the implications of setting

• Pixel format: 8-bit VS. 32-bit floating point (full range)
• Deinterlace method: Blend Fields VS. None

CHECK VS. UNCHECK: Adjust source media to better match project or render settings


For 1080p, what do you usually set the Variable rate to?
Example: Variable bit rate: Maximum (bps): 20,000,000 bps and Average (bps) 15,000,000

How about 720p or 480p?

[budwzr]

8-bit vs. 32 has to do with quantization level. 32-bit has 4X the pixel depth. 32-bit supports an alpha channel (RGBA), whereas 8-bit is only RGB. RGBA is used to render WITH alpha intact, but only on codecs that support an alpha channel.

Deinterlace Mode instructs Vegas what to do if you put an interlaced video on the timeline. Blend blends the fields, Interpolate creates new fields based on a calculated half way point.

Adjust Source Media? I'm clueless on that one. Some "Sony Style" perhaps? Vegas is resolution independent, so you're always going to see your media resampled in the preview window anyway.

OK, I found this in the help. It's best to leave that OFF, and do the resizing and cropping yourself.



For VBR, there is no "usual" setting, but if you look at the templates you can get an idea. 10Mbps is probably good for sending to YT. Same thing for 720/480p.

[c627627]

Thank you for your reply. Does that mean you would recommend using 32-Bit for 720p/1080p and 8-Bit for Standard DVD?

I understand that Deinterlace Mode is only important IF there is interlaced video in the time line. If there is one, can you post which should be chosen and why (Blend Fields VS. None).


What would you set for Variable and Average rate equivalent to your 10Mbps constant suggestion?

[budwzr]

32-bit mode is for compositing and motion graphics. Has nothing to do with HD or SD.

Deinterlace mode is commonly set to blend, because in interlaced video, each frame(Field) has pixels on every other line. So blending puts the two fields together and halves the framerate. 60i becomes 30p. Interpolate is used if the file has field errors, but it's not a clean result.

I didn't recommend "Constant Bitrate Encoding", it's wasteful. Use VBR. With VBR, the bitrate can rise during fast motion, then drop lower. That's more efficient. With CBR, you have to watch the whole output to verify the result.

[c627627]

I am trying to see if I should use 32-bit mode or 8-Bit mode for editing family camcorder footage. Which should I use?


I am also trying to see what most people are using when rendering 1080p, or when rendering 720p of such footage.

Is this too high, what do people usually use?
Variable bit rate: Maximum (bps): 20,000,000 bps
Average (bps) 15,000,000

[smrpix]

What's your source -- raster size, frame rate, compression?

[c627627]

This is my source, 1920x1080 video shot by this:
http://www.usa.canon.com/cusa/consumer/products/cameras/digital_cameras/powershot_elph...Specifications

[c627627]

Here are my notes on how to output to 1080p 720p 480p video from its footage as I am trying to figure it out on Sony Vegas.

I need help to make sure I am selecting everything correctly, considering that little camera Bluray footage source.

HD Rendering [Draft 1 to be EDITED and corrected based on your input]
=========


• File > Properties > All the way to the right of Template, click on the icon: Match Media Video Settings > Browse to the video file being edited.

Right click on opened video file > Properties > Disable Resample (to prevent Ghosting) > OK



• Render as 1080p mp4/aac: File > Render As... > Drop-down next to Main Concept AVC/AAC (*.mp4;*.avc) > Internet HD 1080p > Customize Template... >

Variable bit rate: Maximium (bps): 20,000,000 bps and Average (bps) 15,000,000 >
Encode mode: render using OpenCL if available or Render using CUDA if available
CHECK: Enable progressive download > OK

Project TAB > Video rendering quality > Best

• or Render as 1080 wmv: Render As... > Drop-down next to Windows Media Video V11 (*.wmv) > 8 Mbps HD 1080-30p > Customize Template... >

Video TAB > Image Size: (Custom): Width: 1920 Height: 1080

Pixel aspect ratio: 1,000 (Square)
Video smoothness: 100

Bit rate TAB : 10 M

Project TAB > Video rendering quality > Best



• Or manually set 720p: File > Properties > Drop-down Menu next to Template: Internet 360-30p 16:9 (640x360, 29.970 fps) > Width: 854 Height: 480

Field Order: None (progressive scan)
Pixel aspect ratio: 1.0000 (Square)
Output rotation: 0° (original)

Frame rate: 29.970 (NTSC)
Stereoscopic 3D mode: Off

Frame Rate: 29,970 (NTSC)

Pixel format: 8-bit or 32-bit floating point (full range)

Compositing gamma: 1.000 (Linear)
View transform: Off

Full-resolution rendering quality: Best
Motion blur type: Gaussian
Deinterlace method: Blend Fields

UNCHECK: Adjust source media to better match project or render settings > Apply > OK



• Render as 480p: File > Render As... > Drop-down next to Main Concept AVC/AAC (*.mp4;*.avc) > Internet 480p Widescreen > Customize Template... >

UNCHECK: Allow source to adjust frame rate

Variable bit rate: Maximium (bps): 4,000,000 bps and Average (bps) 2,000,000 >
Encode mode: render using OpenCL if available or Render using CUDA if available
CHECK: Enable progressive download > OK

[_Al_]

If you just edit, meaning cut clips and shuffle them around, 8bit settings is what you use.

If you color correct, or use some effects of this sort you can go with 32bit settings, it is more accurate, but also more power demanding. And you check the differences, maybe you will not notice any.

Your camera shoots progressive footage. You do not need to disable resamle, you shoot 24p you export 24p, you shoot 30p, you export 30p, you leave Smart Resample On and it might be handy elsewhere , slow down for example, don't know. Beware for proper BD authoring you are good with 24p, full HD, not 30p720. Also, if you make mpeg2 for DVD it is going to be progressive also, like your source , if I remember, just wrapped up in interlace wrapper, so smart resample seems to not have any bad influence here. Interesting things might happen if you try to make DVD (30p)from 24p, never tried that. Perhaps you need smart resample in this matter.

Progressive download check is ok, it is meant to be there for internet streaming, if you ever intend to do this, without it checked video would need to be downloaded to PC first and only after that actual streaming would start. Notice, Sony avc encoder doesn't have that, so you'd need to run your render through 'mp4 fast start' software to get progressive download. If you upload for Youtube doesn't matter - it is going to be re-rendered anyway.

EDIT: I think encoder doesn't fix progressive download per se, it is muxer perhaps, so muxer that muxes video and audio doesn't place moov atom properly. For example mp4box,that most x264 encoders use, does that automatically now.