Has anyone of you encoded video with different video and luminance resolution?
I have such task to do and first time heard about such thing.
Could you please help me
+ Reply to Thread
Results 1 to 15 of 15
Can you perhaps be more specific?
What exactly is the task you need to perform?
Encoding a video which will ultimately have a digital resolution different from that of the luminance resolution is normal. You simply need to calculate what digital resolution to set the encoder to, based off of the luminance resolution. (Or vice versa)
This link may help a little:
For example: A luminance resolution of 420, equates to 566x480 in today's digital terms.
To be more specific I have to create 2 videos one with luminance resolution 480x576, second luminance resolution 720x576.
Both movies must have SDTV video resolution (720x576) in h.264/mp4.
Could you please tell me what program I can use to generate such content
720x576 as used for DV and DVD uses non-square pixels. 4x3 and 16x9 aspect ratios are supported.
480x576 is non-standard.
Handbrake is one of many programs that allow customized luma resolutions + h.264 encode with mp4 container.
@edDV: I would be interested to know too how you set the luma resolution in Handbrake
With Avisynth, I would so something like this:
source = FFVideoSource("Path to input", threads=1) # loadsource luma = source.GreyScale() # chroma only low = luma.PointResize(480, 576).PointResize(720, 576) # resize to lower luma and resize again to get the same resolution in low&luma # StackHorizontal(luma, low) #to compare # return luma #to get luma only with full luma resolution # return low # to get luma with 480x576 luma resolution
Thank you Selur
I have made video as you said and generated some video in greyscale. From what I see here http://avisynth.org/mediawiki/GreyScale
I got luma in rec601 format.
Now can I use such luma to generate new video in color (YUV format) ?
How Can I apply modified luma to output video?
Assuming 13.5MHz sampling thus Nyquist criteria Y bandwidth is for:
Simplest way is down resize picture with high quality resizer from 720 to 480 (or required resolution) then up sample such video to 720.
If purpose of this is signal measurements then SinC filter is required (or rather mathematically optimal).
I don't know what is purpose of this question - this is comparison of the resizing algorithms and particular implementations or something else?
480 is comparable to typical BETACAM studio VCR and S-VHS luminance bandwidth