FindDQE

Hi Niko, I can run finddqe well for my counted mode images, but there is an error when I run finddqe for the super-resolution images. I attach the image and log file here. Thanks for your any help.

After carefully gain corrected, the DQE of our K2 summit measured again and compared them to the DQE curve of a K2 image provided by Niko on this Web page. The red curve is from a low dose rate (3e/pixel/second) movie by Thomas. It is significantly better than Niko's (blue) imaged at similar dose rate and total dose.

Fresh K2 summit be installed on our Titan last week. the DQE curve and the images of beam stop were post here, and the silhouettes were imaged at High Voltage: 300kV, dose: 50e/pixel, pixel size: 1.31A/pixel. I just used e2proc2d.py to average the DM4 stack images to a MRC file.

I have two questions here,

the new version can work well for my US4000 data, but compare to the NPS curve of Niko's US4000, my NPS too low, that's maybe abnormal. my pixel dose is 56e/pixel. 300kV FEG.

How can I improve it? lower dose and longer exposure time?

the new version runs well for the downloaded US4000 data. but for the downloaded K2 data, error again, see below:

***********************************************************************************
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0 0x7FFAA32457D7
#1 0x7FFAA3245DDE
#2 0x7FFAA2B97D3F
#3 0x404D7A in calcfx_
#4 0x4053BC in va04a_
#5 0x40B9B4 in fit_counting_nps_
#6 0x4193BA in MAIN__ at finddqe.f:?
./US4002.com: line 8: 705 Segmentation fault (core dumped) finddqe.exe > 50e.log <

I download the US4000.tar.gz data from the DQE Data Files page, I use one of them to test the program. but it cannot run. error like this:

  Input image with pointer silhouette
50e.mrc
   50e.mrc

 Opening MRC/CCP4 file for READ...
 File      : 50e.mrc
 NX, NY, NZ:   4096  4096     1
 MODE      : int*2
 Min, max  :     0.000000        1328.000    
 Mean, RMS :     453.6872        0.000000    
 TITLE    1: SerialEM: Cameras on TF20 0:Falcon 1:US4000 2:Orius     14-Nov-12  11:43:05     

When I run the software FindDQE.exe, but some error message like this:

*** Error in `finddqe.exe': free(): invalid next size (normal): 0x0000000002260d20 ***

Program received signal SIGABRT: Process abort signal.

Backtrace for this error:
#0 0x7EFC425827D7
#1 0x7EFC42582DDE
#2 0x7EFC41ED4D3F
#3 0x7EFC41ED4CC9
#4 0x7EFC41ED80D7
#5 0x7EFC41F11393
#6 0x7EFC41F1D66D
#7 0x414C61 in MAIN__ at finddqe.f:?
Aborted (core dumped)
************************************

somebody said it just because my small buffer in my terminal, is that right?

Hi, In your provided data for 200-kV DQE calculation, there is one file called K2super-200kv-50e-09eps.log. Were both 50e/pixel and 9e/pixel/s in this dataset (K2super-200kv-50e-09eps.log) still measured from counted mode (not super-resolution mode)?
Thanks in advance!
Dong-Hua

In case anyone else is having a problem with low-resolution spikes in their Noise Power Spectrum, attached is a version of FindDQE that flattens the first 1% of the NPS by replacing it with the average of the second 1%. Should quash those crazy artefactual spikes, resulting in better NPS scaling.

Those spikes might also be indicative of a bad gain reference, so check that as well...

Hi,
I have used findDQE software to test the DQE of three different cameras. I collected images on TF20 operating at 120 kV. Although the MTF and DQE curves seem very good, the NPS curves decrease sharpely from 2 to ~0, which are not reasonable.
I have added the image here. the three diffrence color curves indicate three kinds of cameras. In the above are MTF curves, middle DQE curves and bottom NPS curves. You can see the unnormal data.