We heard they used ffmpeg on board, so they shot and stored the individual frames and then compressed them into video files before presumably sending the video files quickly to Earth after landing. Now we're waiting for the individual frames to come down, which will take a while since they sum to a much larger data volume than the compressed video.

Do you think we might get the actual videos? I would bet there are higher quality videos on the rover as they would have further compressed the camera output for rapid uplink after landing. Would there be raw video somewhere waiting to be uplinked? Or is it just a bunch of stills that have been fed into FFMpeg and we just have to wait for them to come down?

I'm not sure what you mean by "raw" video. I don't see the point of processing less-compressed video (than the initial downlinked video) on board and sending that down. We'll be able to make superb quality video from the raw frames once they're all down.

This is educated speculation from reading the EDLcam paper and from what was said at the press conferences. What was sent down initially was the ffmpeg output produced onboard in the EDLcam Linux computer from the raw frames, and they are in the process of downlinking those raw frames now. What's on the raw images website from EDLcam in cases where the raw frame hasn't come down are the frames that were extracted from the ffmpeg file, and then processed however that website processes images in general. So there aren't any "higher quality videos" on the rover.

Basically what fredk said.

There are video compression algorithms that typically achieve much better performance than you would get from a sequence of n compressed images, and these take advantage of the fact that typically there is high correlation from one frame to another. For example, with video of a newscaster in a studio, the pixels representing the background, the desk, etc., don't often change from one frame to the next, so a compression algorithm can achieve big wins by storing only those portions of each frame which changed significantly from the last frame. However, with something like Perseverance's rapid descent, the fluttering of its parachute, etc., you have a lot more rapid change than you would normally expect.

Occasionally, you see compressed video break down suddenly and momentarily in quality when the frame-to-frame change exceeds expectations. One example I'll always remember is the way the image of the crowd "pixelated" when I was watching a basketball game on TV when there was a fast break in the game and the camera suddenly panned, causing the entire image to fluctuate from one frame to the next.

All that said, I'm not sure if either of these things was ever at a premium with the descent videos: (1) Returning to Earth, eventually, a video with the absolute highest degree of detail. (2) Achieving optimal bandwidth in the transmission of an initial compressed video.

It seems that with these videos "pretty good" is all that was needed in both respects. Descent video isn't the high-value science data for which the mission was launched.

Thanks for replying and agreed.

But it is very cool and will be an artifact of historic significance being the first video ever captured of a spacecraft during descent as well as the image data having been captured by very good cameras. It is also capable of providing a good quantity of valuable engineering data from chute deploy to TD and the more frames and higher quality the better.

My initial point was, where is the actual video file they uplinked? It was pretty high resolution. Maybe they can release it in raw form (unedited I mean).

They have said they were going to do that, and there are several thousand RDC raw frames on the website already, so why don't you give them a chance? It is only sol 13.

Point taken.

Nice new image came down.

All of them came down - including a view of the Heatshield impact.

Here's a page of just the LCAM pics, from Ryan's awesome raw browsing interface
https://rkinnett.github.io/roverpics/?mars2...&filter=ELM


(Congrats Mike C - great work from the La Jolla team)

That heat shield impact sequence is on the very short list of the greatest "action" videos in the history of planetary exploration!

Here is a quick animation of all the LVS frames: https://bit.ly/PercyLCAM

It will take a while before the higher resolution versions become available (looks like they did).

WOW, iso it did catch the impact!!!!! I thought it had gone out of frame.

The heat shield took a quite the bounce when it impacted (cropped and zoomed):

Click to view attachment

If it bounced it would be visible separately from the impact ‘splat’ in the crime scene HiRISE image.....which it isn’t..

I think you’re seeing the dust cloud of impact and its shadow move with the wind. These are 1hz images, so it’s actually pretty slow.

Hi!

I tried to combine 23 of the LCAM images into a panorama. I combined that with the compressed RDCAM images, which I already had stitched. I filled the rest, which did not overlay, with cloned color. The colors for the atmosphere was taken from one of the PUCAM1 images. There is definetly no scientific value in this, as there was a lot of postprocessing involved. Stretching, contrast, brightness, sharpening, colorization, resolution upscaling etc. I tried to do it to my best understanding of what it would approximately look like from 11km above the Jezero crater. I will post a 360 version on youtube later. If someone is interested in the version without inpainted color, just say so

On Windows, I use FSPviewer to look at 360 images.

Youtube version:
https://www.youtube.com/watch?v=m1hwTNgZK7E

My take on processing the Descent Down-Look Camera raw images.

Left image colors are based on camera response and and illumination optimized from calibration target on top of rover, and includes chromatic adaption from modeled illuminant to D65 standard. In right image, only camera response is modeled and illumination is fixed at D65, so no chromatic adaption is applied.

Image processing done with custom software in Mathematica.
Mathematica notebook is open source and available at https://github.com/BrianSwift/image-processing
Notebook also contains links to various Mars 2020 and color processing papers and resources.

Video on YouTube

Click to view attachment

Just wondered if anyone knows if there's been a report shared yet which compares how the EDL went in relation to what was expected?
Or is it too soon for that maybe?
Particularly interested in the improved guidance/hazard avoidance in Perseverance vs Curiosity and how that panned out.

FWIW, you can read somewhat open discussion starting at 21:16 here:
https://www.rev.com/blog/transcripts/nasa-p...or-life-on-mars
... but I am not sure if not mentioning hazard avoidance is evidence of omission or omission of evidence, as they say.

Waking up an old thread to mention the impacts of the two cruise balance masses. They were ejected just prior to arrival to allow control over the flight and struck the surface about 100 km west of the landing site. For Curiosity the same applied and the impact sites were imaged by HiRISE. Now with little fanfare we have the same for Perseverance, it seems. See the InSight thread for a reference to this in connection with an attempt by InSight to pick up a seismic signal.

This map locates the impacts:

Click to view attachment

The background is MOLA elevation and THEMIS daytime IR imaging. The overlay is a CTX image. The grid spacing is 0.1 degree or 6 km.

Phil

This is part of HiRISE image ESP_069231_1990 showing two impacts with characteristics I would expect of spacecraft hardware (comparing especially with the Curiosity hardware impacts). They are about 3000 m apart in that image. The two balance masses for Curiosity were only 400 m apart, but two fragments of the cruise stage also struck nearby, a few km apart. For Perseverance, one of these might be the cruise stage, and a second balance mass impact has not yet been identified, or these could both be balance mass impacts and the cruise stage site has not been identified. Curiosity impac site ejecta was dark, here it is bright, probably just due to local conditions.

Phil

Click to view attachment

PS - another 6 small masses were ejected during flight. For Curiosity their impacts were a little downrange of the landing site. HiRISE has looked for their impacts but not seen them yet as far as I can tell from a cursory look.

I just noted a third impact in this area:

Click to view attachment

The impact 3 - impact 2 vector is the direction of travel.


Phil

I have now located the impact sites of the small balance masses west of the landing site. Quick look here, something more detailed later. The top image shows dark ejecta from two impacts. The bottom image shows three sites where new objects struck the surface - it's a before-after difference image of two CTX images. The Seitah-like structure south of the impacts can be found quite easily within Jezero if you are so inclined. As I say, a better representation of this will follow. Yesterday my LPSC poster said these had not been found yet. Today I finally noticed them.

The top image is the easternmost of the three arrows in the lower image. HiRISE image ID is in the file name if you download the image.

Phil

Click to view attachment

Click to view attachment