DaViD Boulet
Senior HTF Member
- Joined
- Feb 24, 1999
- Messages
- 8,826
It's not a compression artifact exactly...Mosquito noise is a little more grainy.
I do know what you're talking about and I've seen it myself.
Long story short...it's most likely a by-product of some form of vertical-domain edge-enhancement. EE becomes more pronouces as the pixel-transitions in the image rise higher and higher in frequency: The most dramatic transition you can get is going from all dark to all light pixel...and that's exactly type of thing that's happening in the 16x9 frame when the black letterboxing bars of the 2.35:1 image suddenly transition into the picture area of the active 2.35:1 iamge.
Many folks here and at AVS (avscience.com) have theorized about these mystery "lines" you see running parallel with the top (bottom?) of the letterboxing bar in the 2.35:1 image and that's the best that anyone has come up with. It's possible that many MPEG2 encoders in existence impart some level of HF emphasis...however mild...and so just because the technician swears *he* didn't apply any EE doesn't mean it wasn't applied at some point during the signal chain. Signal processing...even that which happens inside a particular MPEG encoder...should be confused with "compresion artifacts" which describe artifacting from the data reduction/decompression process of the MPEG algorithm.
Keep in mind folks that a transfer could have very little "vertical domain" EE applied (this type of EE shows up along *horizontal* lines) and see very little in the actual picture...but *still* get a visible artifact along that masking bcs of the extremely HF transition at that point in the 16x9 image.
I'd love for Bjoern to chime in this thread and contribute his thoughts!
-dave
I do know what you're talking about and I've seen it myself.
Long story short...it's most likely a by-product of some form of vertical-domain edge-enhancement. EE becomes more pronouces as the pixel-transitions in the image rise higher and higher in frequency: The most dramatic transition you can get is going from all dark to all light pixel...and that's exactly type of thing that's happening in the 16x9 frame when the black letterboxing bars of the 2.35:1 image suddenly transition into the picture area of the active 2.35:1 iamge.
Many folks here and at AVS (avscience.com) have theorized about these mystery "lines" you see running parallel with the top (bottom?) of the letterboxing bar in the 2.35:1 image and that's the best that anyone has come up with. It's possible that many MPEG2 encoders in existence impart some level of HF emphasis...however mild...and so just because the technician swears *he* didn't apply any EE doesn't mean it wasn't applied at some point during the signal chain. Signal processing...even that which happens inside a particular MPEG encoder...should be confused with "compresion artifacts" which describe artifacting from the data reduction/decompression process of the MPEG algorithm.
Keep in mind folks that a transfer could have very little "vertical domain" EE applied (this type of EE shows up along *horizontal* lines) and see very little in the actual picture...but *still* get a visible artifact along that masking bcs of the extremely HF transition at that point in the 16x9 image.
I'd love for Bjoern to chime in this thread and contribute his thoughts!
-dave