Feeding at the Bacteria Swarm Ring

Images made through a microscope. All subject types.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

Thomas Ashcraft
Posts: 133
Joined: Mon Aug 07, 2006 5:56 am
Contact:

Feeding at the Bacteria Swarm Ring

Post by Thomas Ashcraft »

Image


The above image is not a photograph but rather an extracted image from an .avi movie file.

Here's an update to a previous forum post I made about a "possible bacteria swarm ring". Since that post I have been observing other bacteria rings under the coverslip and videoing behaviour patterns. Charlie K. remarked that he had seen some protozoans feeding at the rings and since then I have seen this as well. Here is a 13 second movie of a couple of Paramecia feeding.

http://www.heliotown.com/Paramecia_Feed ... _Ring.html

I am still experimenting with movie formats. Previously I wrote that I thought the Quicktime .mp4 format might be superior to DivX. Since writing that I am feeling that DivX might have the edge. The .mp4 movies seem in general to be softer and less crisp and contrasty. In any case, here are two choices of the same movie clip for anyone who wants to compare. ( For me, the .mp4 file on my Windows screen suffers greatly but seems acceptable on a Mac screen. )

Canon S3 IS still camera in movie mode
Meiji phase contrast scope and objectives

Tom

Charles Krebs
Posts: 5865
Joined: Tue Aug 01, 2006 8:02 pm
Location: Issaquah, WA USA
Contact:

Post by Charles Krebs »

Hi Tom.... nice videos! I had no problem playing either one. The DivX sure looked the better of the two on my screen.

rjlittlefield
Site Admin
Posts: 23543
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Tom,

I agree with Charles, that the DivX looks better on my screen.

But when I study the images in detail, I see a couple of funny things going on.

What I did, to study the images in detail, was to launch both movies, position them to the same frame roughly in the middle of the sequence, grab screen images, and pull the screen images into Photoshop so that I could easily look at histograms, overlay the two images to click between them, etc.

I see two funny things.

1. The images show the same area, but do not have the same size or even the same aspect ratio. The Quicktime movie is displayed at 640x480, while Windows Media Player displays the DivX at 640x432. I'm guessing that your camera originally produced 640x480, in which case DivX is squashing the images a bit.

2. The Quicktime movie is distinctly brighter and the histogram shows gaps typical of level adjustment.

If I resize and level-adjust the Quicktime to match the DivX version, then at first glance the images are quite similar.

However, when I stack the two images in Photoshop and click between them, there are striking differences in detail. The DivX frame is just loaded with fine features that might be real image content or might be compression-induced artifacts. I cannot tell which is which, but you probably can, by referring back to your original huge movie files.

Can you do something like these tests on your system, and let us know what you find? It could be that DivX and Quicktime are each doing bad things to your images, but different bad things!

If you would like (and will grant permission), I can post a couple of images showing what I'm talking about, but when I did what I describe above, the differences were pretty obvious.

If you were to build a Photoshop stack containing the same frame from your original avi and as reconstructed from both of the compressed versions, I suspect you could get a good evaluation pretty quickly.

--Rik

Thomas Ashcraft
Posts: 133
Joined: Mon Aug 07, 2006 5:56 am
Contact:

Post by Thomas Ashcraft »

Rik wrote:

"If you would like (and will grant permission), I can post a couple of images showing what I'm talking about, ....... "


Hi Rik,

As ever, thanks for your analysis and for sure, please do post your findings. This is all a steep learning curve for me without many places to get information and the opportunity for improvement is of great value.

Yay for this forum!

Tom

rjlittlefield
Site Admin
Posts: 23543
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Thomas Ashcraft wrote:...steep learning curve...without many places to get information and the opportunity for improvement is of great value.

Yay for this forum!
Amen...

Here are the recovered DivX and QuickTime MP4 frames:

DivX
Image

QuickTime
Image

And here's a cropped & animated gif that flashes between the two.

Image

I think I might have been off by one frame when I selected them, but you'll get the idea. Look at the OOF background areas -- it's almost like the QuickTime file was run through noise-reduction software :? :-k

--Rik

Thomas Ashcraft
Posts: 133
Joined: Mon Aug 07, 2006 5:56 am
Contact:

Post by Thomas Ashcraft »

Rik wrote:

I agree with Charles, that the DivX looks better on my screen.
. I'm guessing that your camera originally produced 640x480, in which case DivX is squashing the images a bit."


Hi Rik, Charlie, everyone,

Interesting to compare your frames, Rik. Thanks! I was unaware of the different size discrepancies. Yes, I am shooting at 640x480 at 30frames per second.

A couple of general responses off the top of my head:

I think the reduced .mp4 noise also reduces detail, which may not serve the microscopist. For example, I believe the "noise" in the DivX version is actually bacteria and particles catching the phase contrasted light, though indistinct. ( Phase contrast light is a difficult video format light.)

I guess that .mp4 and DivX are general compression formats trying to hit some average medium for all sorts of users. Maybe someone could eventually design a codec specifically for videomicroscopy.....or a plug-in or something.

I have not found a way in Quicktime Pro 7, which is the .mp4 codec, to adjust any parameters. Maybe I need to look inside deeper.

Also, DivX seems to look better on both pc and Mac whereas an .mp4 looks appreciably worse on a pc screen than a mac sreen. I tried adjusting my PC monitor for gamma and such but could not improve it enough to make the .mp4 look better.

I will study the the process better later and report further.

Tom

rjlittlefield
Site Admin
Posts: 23543
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Tom,

I'm eagerly awaiting further tests -- your investigation is very helpful to me too.

About my "noise-reduction" comment... I'm easily convinced to agree with your interpretation that what's being lost in the MP4 really is fine detail caused by OOF objects viewed with phase contrast. I'm certainly not a fan of noise reduction software in general. For me, it always seem to make something or other go "plastic", wiping out texture or detail that I know is real. The individual frames that I pulled from your video seem to illustrate exactly that. If you could pull a matching frame from your original huge avi file, and see that it has the same fine detail/texture as the DivX frame, that would clinch the case.

I don't have a Mac to test on, but I wonder if the difference in appearance of the mp4 between PC and Mac is due to QuickTime Player trying to correct for the usual difference in gamma between PC and Mac video systems. If it's taking the Mac as standard, and brightening the image for the PC, that would explain the histogram effects that I observed in screen captures.

It's disturbing that the DivX file is apparently changing the aspect ratio (squashing in height). From a scientific standpoint, I'd really like to know that I can measure the same object at any angle and get the same number, and the usual assumption is that pixels are square. Given limited time to investigate, I think solving that issue would be my highest priority. Other than that, your DivX stuff looks great!

--Rik

Thomas Ashcraft
Posts: 133
Joined: Mon Aug 07, 2006 5:56 am
Contact:

Post by Thomas Ashcraft »

Rik wrote: "I'm eagerly awaiting further tests -- your investigation is very helpful to me >.."too.

Image


Rik,

here is a frame I pulled from the original .avi file. To get this frame I loaded the original .avi file in my Mac's I-Movie program which can pull out a 72dpi frame and make a .jpg. I'm not sure how to grab an individual frame from a movie other than this program.

Anyway, it's pretty close to the DivX. Maybe I am one frame off from the samples....not sure. 1/30 of a second frame to next frame does make a noticeable difference in activity.

And....I checked my DivX settings.....dang..........I wasn't set to 640x480 but rather 680 x 432. Thanks for catching that. ( Human error. ):roll:

Tom

rjlittlefield
Site Admin
Posts: 23543
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Tom,

Thanks for the followup.

Yep, the DivX is far superior in this example. That's kind of depressing, 'cuz that H.264 codec used in your .MP4 is supposed to be the greatest thing since sliced bread. Apparently it's not as good as it's supposed to be, or perhaps Apple's implementation is not the best. (A lot of compression algorithms allow for varying tradeoffs during compression, although all of the resulting files are compatible with the same decompressors. The same is true even for still image encoders. It turns out for example that the JPEG's produced directly by my Canon 300D are far inferior to JPEG's of the same size produced by Photoshop from the Canon's raw files.)

Interesting... :-k

--Rik

Post Reply Previous topicNext topic