SONY codecs XAVCS and AVCHD

Not an A7S expert, but I'd guess from past camera experience that you could record 1080p using XAVC S mode if you wanted to, but when choosing the higher frame rates (for slow mo) it forces the camera down to 720p mode.

In Premiere Preferences, you can set Default Scale to Frame and then if you import 720p clips and drop them into a 1080p sequence, they just automatically fit (changing this setting does not affect clips that were already imported previous to changing the setting). Nothing to scale or render, works automatically behind the scenes. 720p clips can look fine in a 1080p sequence, but you might consider the delivery options and decide that 720p delivery would work in which case the sequence could be 720p, then you are not "blowing up" any clips. 1080p of course scales down to 720p beautifully.

I don't believe you are correct about AVCHD being much slower to edit than .mp4 though.

IMPORTING - copy entire contents of SD card to a NEW folder on your hard drive. Do not cherry-pick the video clips - copy ALL folders directly over from SD card, keep it all intact as it was on the card. Then in Premiere, use Media Browser to import clips, rather than File > Import. This solves a lot of issues.

Pixel Aspect Ratio (PAR) -  the shape of the pixels (dots) making up your video image. Pixels are typically square (1.0 PAR), especially with still images. Some video formats use non-square pixels. In NTSC-land, DV video (720x480) used 0.9 PAR for 4:3 video and 1.2 for 16:9 wide video. Note that the DV resolution was always the same, being 720x480 pixels, but by making the pixels wider (rectangular) it produced widescreen video as a result!

Non-square pixels are also used for "thin raster" HD video formats. When HDV came out many years back, they needed to somehow fit HD video onto the same miniDV cassettes meant to hold SD video. In order to do this, they compromised the resolution of the 1080i video, making it just 1440x1080 rather than the full 1920x1080 pixels. They made this work by using a 1.333 PAR. Use less pixels to save room in the compressed file, and just make the pixels wider to make up for it! Simple math shows how this works. 1440 x 1.333 = 1920 on playback.

Modern cameras shooting 720p or 1080p all use square pixels usually, although some may still offer an optional lower quality 1440x1080 recording option to get longer record times on the SD card.

Progressive vs. Interlaced - back in the days of SD video, NTSC and PAL (what you watched on TV broadcasts) were Interlaced video signals. Old TVs with CRT technology scanned the image lines onto the screen one row at a time, top to bottom, causing phosphor dots to glow briefly. However, in the time it took to draw all the lines top to bottom to make a complete image (frame of video), the dots at the top of the screen had already started to lose brightness. This would cause a flicker when new (bright) lines were drawn over the fading dots of the previous frame.

In NTSC (US), there are 29.97 frames of video each second. The 1/30th second "refresh" between drawing frames of video was too long. In Europe, they use PAL which is 25fps.

The fix was interlaced video. They broke a single frame of video into two fields. A field is half a frame, made up of the even or odd scan lines only (also called upper or lower fields). A field of video could be drawn in just 1/60th of a second (using 59.94fps to be precise) which is half the time of a frame, and before the dots of the first field could lose their glow, the second field was drawn in to "fill in the blanks" and complete a frame. This eliminated the image flicker, however...when a camera is shooting video using fields, objects in the frame that are moving will change position between one field and the next, so the even and odd fields may not completely match thus a strange motion artifacting is introduced.

Think of a stationary camera, aimed at a standing person. If the person is waving his/her arms, the arms will be in different positions on each field (second field grabbed 1/60 of a second later than first), so they get duplicated when the two fields are blended into a frame. It does happen so quickly though that the eye perceives smooth motion when watching the video, but if we pause the video when editing, or add slow-motion, then we see more of the issue.

When HD video first came out, 1080i was the standard (as HDV), which was 1920x1080 interlaced. There was also 720p and later 1080p.

So what is Progressive video? Progressive means that every frame is complete - there are no fields. 720p30 means video is recorded as 29.97 complete frames every second. Same for 1080p30. 24p video grabs 23.976 frames per second. I believe you use PAL, so that is 25fps. The camera shutter opens and the entire scene - the frame of video - is saved at once. The downside is that at low frame rates, motion can look choppy, for instance a fast camera pan, or someone running through the scene, as the subject moves a lot from one frame to the next. The scene changes in 30 steps per second. While NTSC or 1080i also used the same 30 frames per second, those 30 frames were broken into 60 fields to provide 60 steps per second. Much less change happening between fields = smoother perceived motion.

What this means for the videographer shooting 24p, 25p, or 30p is that they have to be conscious to not move the camera too fast for instance, to keep motion smooth. I should mention that when CRTs were replaced by LCD and LED flat-panels, we no longer had to worry about phosphor dots losing their glow - progressive video looks fine and does not flicker. Many editors see 1080i as a curse and wish interlaced video would just die already! It causes lots of issues in many ways and most editors can live without it. Interlaced video made sense to work with old CRT TV sets, but there is really no need/no use for it with modern HD displays, which are by their very nature and design, progressive.

720p60 has been around longer than 1080p60, for the simple reason that older cameras could not handle processing and storing 60 frames of HD video every second at 1920x1080 - that's a lot of data - basically twice as much as 1080i which was storing just 30 frames per second. The 720p frame size of 1280x720 was easier to accommodate at 60p. Most new cameras now offer 1080p60, or beyond. However, moving up to 4k, we get back to the same situation that most cameras cannot handle 4kp60 and are limited to 30p recording, but in time that will also change.

Just did a Google search for "interlaced video" and got a million hits for descriptions and sample images, try it and it will illustrate the above ramblings.

Back to Sequence Settings, 1080i is the only HD format that is interlaced, and it is always Upper Field. In the US, with the 29.97fps standard, 1080i goes by different names but all representing ONE format. Might be 1080i29.97, 1080i30, 1080i59.94, or 1080i60. ALL MEAN THE SAME. In PAL countries with 25fps standard, you have 1080i25 or may be called 1080i50. The 60 or 50 referring to fields.

Thanks

Jeff Pulera

Safe Harbor Computers