Acquisition rates with different binning and exposure times in the continuous mode (many repeats with only one ratio in the inner loop).
Q: How fast can I get? What is the maximum frames per second I can record?
A: The maximum frame rate depends on several factors like resolution, binning and the camera readout speed and mode. Please read on to find out more. You will find out more as we give you an insight on the technical background in image aquisition.
After the acquisition of an image the camera sends it at a rate
of 25MB/second to the frame grabber board. That corresponds to a
rate of 12.5MHz for the pixel transfer rate (16 bit information).
The PCI frame grabber has two memory banks, each equipped with 8MB
RAM. From the frame grabber the image data is transferred to the
computer RAM via the PCI bus.
Since the frame grabber has two banks of memory it is possible
that the camera writes data into the second bank while simultaneously
the computer reads the images from the first bank. It is obvious
that if you acquire more than 8 MB of data, the system has to switch
between banks during the measurement.
The TILL Protocol Editor uses the hardware in the following way:
All images within one repeat loop are grabbed into one bank. Since
the images of the next repeat loop are grabbed into the other bank,
the system has to switch between the banks. This switching and its
implications need a few milliseconds, depending on the chosen image
size.
As a result of this, the highest acquisition rate is achieved if
you grab as much data as possible into one bank to avoid bank-switching.
We call this "bursting acquisition" in contrast to the
continuous acquisition mode where you do not care about the bank switching.
|
Continuous Full Frame Ratios Acquisition rates with different binning and exposure times in the continuous mode (many repeats with only one ratio in the inner loop). |
You can speed up the acquisition rate by reducing the number of readout lines (reducing the chip window size). But keep in mind that the acquisition rate is mainly limited by the exposure time. In the following table we start with an exposure time of 30 ms and a 1x1 binning. With higher binning you can reduce the exposure time (keeping the same or getting an even better signal to noise ratio!) since you get all the electrons from the binned pixels.
Exposure time |
30 ms | 30 ms | 2 ms | 1 ms |
| binning | 1x1 | 2x2 | 4x4 | 8x8 |
| # chip lines | ratios/s | ratios/s | ratios/s | ratios/s |
| 10 | 14.1 | 35.7 (8) | 62.5 (4) | 66.9 (1) |
| 30 | 13.9 | 34.5 (15) | 55.6 (8) | 66.7 (4) |
| 60 | 13.7 | 33.3 (30) | 50.0 (15) | 62.5 (8) |
| 120 | 13.0 | 32.5 (60) | 47.0 (30) | 58.8 (15) |
| 240 | 10.3 | 23.3 (240) | 27.0 (120) | 43.5 (60) |
| 480 | 10.3 | 23.3 (240) | 27.0 (120) | 43.5 (60) |
The numbers of the used CDD-lines are indicated in brackets ().
 |
Bursting Full Frame Ratios 10 times 6 ratios. The overlapping exposure (exposure of an image while the image acquired before is transferred to the framegrabber) is used during the whole loop. In the table below are listed the inner burst acquisition rates measured as the difference between the first and the last acquisition of an image of the same wavelength divided by the number of image pairs (ratios) in one loop. |
| Exposure time | 30 ms | 8 ms | 2 ms | 1 ms |
| binning | 1x1 | 2x2 | 4x4 | 8x8 |
| # chip lines | ratios/s | ratios/s | ratios/s | ratios/s |
| 10 | 18.75* | 60.0 (8) | 120.0 (4) | 120.0 (1) |
| 30 | 18.75* | 60.0 (15) | 100.0 (8) | 120.0 (4) |
| 60 | 18.75* | 60.0 (30) | 85.7 (15) | 100.0 (8) |
| 120 | 18.75* | 60.0 (60) | 75.0 (30) | 100.0 (15) |
| 240 | 18.75* | 54.5 (120) | 54.5 (60) | 85.7 (30) |
| 480 | 18.75* | 37.5 (240) | 37.5 (120) | 75.0 (60) |
* limited by the exposure time; overall acquisition rate is 14.1 ratios/s. The numbers of the used CDD-lines are indicated in brackets ().