IMPORTANT NOTE FOR PEOPLE USING SIMUL-FOCAL MICROSCOPES: DO NOT include your eye pieces (typically 10x magnification) Only use: (the HEAD'S magnification range) x (your barlow lens) x (camera's zoom) The camera's image IS NOT AFFECTED BY YOUR EYE PIECE'S MAGNIFICATION (often they're 10x). unless you're connecting the camera TO the eyepiece. For simul-focal, where it uses a 3rd port -- only the BARLOW lens + camera lens influences the camera's zoom factor for this equation. Thank you! :-D. Your video alerted me to a THINKO I'd made! I'd completely failed to consider prior to seeing your computation. (also, probably bc I've never used // looked though a microscope / camera, but as I make buying-decisions, this will help!)
I’ve been thinking about this and your explanation is perfect. From your advice 2 MP - 5 MP is sufficient in most situations. One of the guys in the comments helpfully points out that fps video capture is reduced at higher CCD capabilities. This is an important consideration as well, in addition to signal to noise and data rate requirements. Many thanks for this great video. -Very helpful.
Microbehunter: I think you are off by a factor of 4 Your calculation are good for a monochrome sensor; but we use monochrome sensor cover with a RGB "Bayer filter". A 8MP camera have 3x8=24 Mega smaller photosites or subpixels. In addition, it's not 1/3 of each color. The Bayer filter use 1/2 of green filter, 1/4 of blue and 1/4 of red. (Click to get more info) For a 4K or roughly 8MP camera (3x8=24 individual Mega subpixels): 🟩🟩12Mega are green 🟦6Mega are blue 🟥6Mega are red BAYER FILTER🟦🟩🟥🟩🟦🟩🟥🟩🟦🟩🟥🟩🟦🟩🟥🟩: Each of theses 24Mega individual subpixels only cover 1/4 of the space of a "regular pixel" size. Therefore, such individual subpixels don't get any information related to his respective color for 3/4 of the size of the "regular pixel" (for red and blue at least). For blue, it's like having a monochrome sensor with 4 times more pixel (compare to a same size regular sensor) and with 75% of dead pixels ⬛⬛🟦⬛⬛ and only 25% of working pixels spread apart by empty space. Each of theses bleu photosites would also be 4 times smaller compare to a 'regular pixel' ◾vs ⬛. Dam, it's hard to explain this properly! All wavelength are not created equals🏳🌈: The calculations of pixel size should use the worst case scenario. Instead of green 550nm light, we should use shorter wavelength such as blue 450nm or even purple 400nm light. The Nyquist Criterion〰: Half the wavelength is the absolute minimum for a perfect sensor and a perfect lens. If you get more data than that, it will gives you more info (not more resolution) that can be use to compensate for some of the noise. What does the electronic hardware and AI software is doing with this info? From there... the math goes sideway in mystery... My guesstimate: Consider all the above, your number seems to be off by a factor of at least 4x (just for the Bayer filter alone, not consider the wavelength and magical noise stuff). So, in worst scenario (mainly violet sample and 4x objective), a camera sensor should have about 19MP Since 4K is 8.2MP, it should be good enough for most video application in visible light microscopy, but 6K would be optimal at 19MP for the lowest magnification. This is a very big difference with the numbers you calculated in the video (monochrome sensor). Maybe you should consider doing an update (unless I made mistake???). A curious mind from Montréal I've spend more than an hour to write this down.. I hope somebody will read it!
Just a comment of the definition of megapixels. Generally, the megapixels are all the actual photos sites. The individual pixels are not made up by color subpixels. To get RGB values from the Bayer filter, the missing color information is in top plated from neighboring pixels (check Wikipedia). Bottom line: an 8MP Sensor consists of 4mp Green dots, 2MP red and 2MP blue. And the rest is interpolated
@@TalkingWayTooMuch This is also what I've understand from the Bayer filter: 8MP = 4MP green, 2MP blue 2MP red. But if you compare to a real black n white 8MP, you have 4 times more info (for blue and red, equivalent of gray shade) and 2 times more info (for green, equivalent of gray shade). If you are looking to a blue object with a 8MP color camera, you will get 4 times less information compare to a 8MP black n white camera that has no Bayer filter (bare sensor). With the color camera, each pixel, for the blue, is 25% of the whole pixel. For a true B/W senor, each pixel is 100% of the pixel size. So, the color camera will have less definition and will need to interpolate the missing 75% of the image. When the image is not 100% blue, mainly white light, the interpolation is way more precise. But if you take the worst case scenario, a 100% blue (or red) object will get a 4x less information on a color pixel At least, this is what I've understand. We can argue that most microscopic pictures/video use white light and the interpolation is not as critical.
This is a great explanation but still leaves me wondering why the Microscope industry seems oblivious to the fact that most modern microscopes are also used for taking pictures at some stage, by whatever means. Users/customers are ill advised on what they really need. 'Dedicated' microscope cameras are not really dedicated at all and available under numerous different brand names. You ask to see images taken by the camera you are thinking of buying and are shown misleading images supposedly taken with the camera model concerned, but are exactly the same stock images used and sent out by many different retailers. When the camera you buy is one specifically recommended for your make and model of microscope, why is that same camera far from perfect when used on that specific model? I use a DSLR for imaging and it generally gives very good results although needed some tweaking with tubes to get the same image size as the microscope. It still isn't par focal with the microscope and a lot of chromatic aberration is generated by the supposedly achromatic objective lenses which I can't seem to get rid it. I would love to be able to buy a smaller dedicated microscope camera that just worked and produced reasonable results without having to mess about without image and focus matching issues. It doesn't seem to be a lot to ask given modern technology and the volume of the market. Once again we all seem to be at the mercy of retailers who only have one objective (which is also not achromatic) - making huge amount of money. Unfortunately we can't all afford to go for a tailored service with Olympus or other high-end manufacturer. There should be more thought given to providing a better service to customers who still spend a lot of money on their equipment and want reasonable results. Finally, it is really great that there are people like Oliver around to help us all! :)
A microscope camera with a good implementation of the Sony IMX 290 sensor is a very good, tried and battle hardened compromise for budget to midrange microscopy. It is therefore quite common, so you'll find a very competitive market among them with pricing accordingly. The 290 had excellent low light performance, and if you use the camera for direct output, it can also do 1080p at 120 fps with a monitor supporting high framerates over HDMI. Most microscope cameras are not primarily designed for still photos or occational video recording. They will certainly do that, but most modern microscope cameras are engineered for being a primary output source. To output a high resolution, very low latency video feed to a monitor instead of (only) observing through the eyepieces on the microscope. This changes many of the priorities and focus (pun intended) areas when designing them. You want a sensor package that can do crisp bright image at high framerates with very low latency, you need a graphical user interface overlay and means of controlling that (because the video has to go directly to a monitor's input, over USB requires passing through multiple interfaces, compression and decompression, which introduces way too much latency to be useful as a substitute for direct optical port view). Mpst of them also operate with *very* large field of view compared to what the type of objectives in this video enable, and are often used at lower magnification. Your challenges with a lack of clear product information, the stock images thing etc - I believe that is not really a mark of the cameras, but rather the marketplace you're seeking these out in. If you need solid information on these cameras, seek that on forums, or from vendors that actually specialize in this equipment. General electronics sellers on AliExpress, Ebay, Banggood rarely know snything more about it than is written in the sales ad, and often they don't know what even 10% of that text actually means. Shopping this kind of stuff from them, is like shopping microscopes, barlow lenses and specializes cameras at Walmart or Lidl. Don't expect the clerks to even know which way is up on the damn things - and the manager only knows there's a demand for a thing, so they wanna capitalize. There might be a great deal _to be had_ there, because the price can be very low, but you're on your own in researching that before buying. If you want proper knowledge and understanding of the parts from a vendor, you need to seek out vendors specializing in that equipment. You're gonna be paying more, because now there's very specific competence requirements for their employees, and speciality skill workers cost a lot more. But you also get curated product offerings. Those sellers won't offer for sale, the myriad of bad, low quality and garbage equipment out there - they have the skill and knowledge to only stock products that suit their customers, and only stock products they trust will perform well and last. Because the customer base is small and have many common work field, industry and interest points, a bad customer experience reflects back and hurts a specialized vendor far more than it hurts a general, everything-goes vendor. Because with a general electrics seller, maybe 300 people hearing of this bad experience buying that microscope camera. Of them, maybe 1-2 _might_ at some point consider buying such a camera from them. With a specialized vendor however, maybe only 60 hear of this bad experience. But of those 60, 40-50 (!) of them *will* be looking to buy a microscope camera within the next 6-24 months. So you see, if you're willing to research some models on your own and with help from forums, reddit, youtube, industry/university contracts etc - then you might get more bang for buck with buying from a general electronics vendor from China. But if you want professional guidance, information, low risk of buying something bad - not to even mention if you want aftersales support - then you need to seek out a speciality vendor. And expect to pay a premium for that skill and support, because their exoenses ar higher, and they sell much fewer goods in total per month than general electronics vendord, so they need a bigger profit margin per sold unit. If you want such speciality, highly competent vendors available for information and goods in the future - you have to support them by buying those products from them. If you only use them for free information and then do the actual shopping from chinese general "stuff" vendors instead, then all the sales revenue go to the big fsceless corporations like those "general stuff sellers" in China, banggood, Amazon etc. Then those specialized vendors that offer deep knowledge and understanding, professional service, skill and technical support - they go out of business from lack of sales. And once the big corporations eat up the sales in that market, it is no longer possible for such speciality vendors and shops to open new specialized shops - so then those speciality vendors and shops are gone for good, along with all the product and customer need expertise they offered. The same is true for almost all other speciality vendors and shops, from hobby stores, knitting and sowing stores, art and crsfts supply stores and camera gear stores, to specialized car parts stores and so many more. If they don't get the sales, no shop can survive. Once the big online stuff sales corporations eat up what little was left of that market, then the speciality stores are gone for good.
That's the most important thing I required ..thanks again.. I was just thinking to buy a microscope camera n here once again u helped me to save a lot..
I wonder how much money this video has saved after people watched and shared it! Great explanation and breakdown Sir! In another video you mentioned USB2.0 would be too slow and suggested getting a USB3.0 camera. Is the slow speed of a USB2.0 connection related only to live streaming to a computer screen or is it extended to the recording also? In other words would the recorded video be smoother?
I think that the USB affects latency, or the delay between the image shown on the computer and what's actually happening on stage. USB 2.0 should be able to transfer almost half a Gb per second, it should be 10 fold more than enough even for (compressed) 4K 60fps video in theory. I'm guessing..
why do you think that scientific professional microscope cameras have resolution of minimum 50 megapixels. Do you think that all of them they are fools...
@@lotharmayring6063 Do you have a link to them? I've only seen ones with smaller resolution so far. Maybe those microscopes have a very large field of view? Maybe they've got such resolution to be able to digitally zoom in in a picture taken at lower magnification?
@@lotharmayring6063 I expect those professional microscope cameras either to have a very large sensor (way beyond the 1/2.3 / 1/2.8 1/3 inch you usually find) so the pixels are also way bigger and thus far more susceptible to light, increasing the SNR significantly (and thus also reducing noise) or that they actually feature that many pixels on a small size, but are actively cooled by peltier-elements to reduce the noise (as sensors at that size with that many pixels have extremely poor sensitivity). Either that or there's some very exquisite processing done to the image from the sensor, just like the supercrap-sensor in the average phone only produces a sensible image due to the very powerfull processing being done to the sensors' output, as the size is so small, the sensitivity is just ridiculous (and the amount of noise therefore is as well.
did you forgot the fact the screen/monitor on wich you project/watch your video recorded by your camera induce some magnification? with a 1.7Mp, you record a video of 1622x1084 px. if you watch it on 1920x1080 resolution, you 'll need to magnifiy the horizontal res. or you'll get a pixelized image it can become worst if you have a monitor with 4k resolution : 3840x2160. if i understand , you 'll need a 8mp to record in 4k. what gain can i have with a 13mp then? more bright image?
Congratulations for the video, very interesting. These calculations would be for a monochrome sensor, right? For a color sensor with a Bayer type matrix, what would then be the ideal pixel size and the resolution in Mpx?
That’s going to be very similar because AFAIK the luminance Values are taken directly from the pixels, and the color information is in interpolated, So the resolution of a RGB sensor ( with Bayer filter) is not that much worse compared to monochrome sensor
I was very happy to happen on this video for two unrelated reasons: 1. For no practical reason I built an adapter to connect an old time lens to my Canon t7i DSLR (22mm sensor) . It worked but the pictures are fairly blurry. The lens was designed to work with 10 x 10 inch film which means that there is a huge crop factor with this lens (11.3 by my calculation). So are the images blurry because the limit set by physics or are the images blurry because 140 year old lenses weren't that great? I think there might be enough in this video to get a rough idea of the answer. At least I'm going to take a stab at it. 2. The Webb telescope uses detectors with only 40 mega pixels. Is that enough to achieve the maximum resolution given the mirror size. I tried to calculate this but I wasn't sure I was doing it correctly. It surprised me that it was enough to get the maximum resolution given that modern cameras, which often have more than 40 megapixels, have much smaller lenses than the Webb primary mirror. The Webb operates primarily in the infrared so for a given mirror size it has lower resolution. Maybe that's enough so that a higher resolution wouldn't be enough? There was also a related reason. I had been curious about this for awhile. Obviously there was a limit where a higher resolution sensor wasn't going to provide a higher resolution image. I was happy to find out what that limit was.
Back of envelope calculation: a 1/2.5" to 1/3" sensor size (about 4mm and change x 5mm and change) at 5 megapixels meets the requirements, assuming a 0.5 reduction lens for the eyepiece camera.
for video all is limited by transfering time to the memory and you can forget about pixels and lenses, And only for this reason i think 2 megapixel is sufficent
on my opinion the resolution of an microscop is only given by the light wavelength and NA of the ojektiv. So it is not afected by ocular or adaper lenses. The size of real image produced by the objektiv is nearly the same for al microscops and this image has to fit on the sensor. Concerning this you need about 100 Megapixels for 1 my resolution getting the full image. But if you blow up your picture with optics after the objektiv i am shure that 1 megapixel camera could be sufficent. In this case you waste NA of your objectiv, for video then you can moove around the object so scanning it and that works with a 2 pixel camera
I am trying to do this calculation for my circuit board soldering microscope with objective AN 0.065 @ 4.5X & 0.02 @ 0.7X. I have an 8.3pm camera with a Sony imx485 sensor 1/1.2 in. The optical diagonal is 12.86mm. The calculations I am coming up with say this camera doesn't have near enough resolution. Can anyone confirm this calculation?
I think MicrobeHunter did forgot a very important thing: he made the calculation for a black and white sensor. So, we need to multiply by four the number of MP (due to Bayer filter). In your case, 8.3MP ok 4K is a great resolution, but it's not maxed out yet at this magnification. So, you have enough resolution to get a great image quality, but you can get even more if you wanna use a digital zooming without losing resolution. So... keep what you have and be happy!
This is very interesting and also a little depressing! Without doing all the calculations out of complete lazyness on my part. I made two big assumptions. First being you would be able to use a small camera with low number of mega pixels just as well as a heavey dslr. This seems true because my phone held to the high piece seemed to do surprisingly better than my dslr attached to microscope. Secoundly Iv recently got into astrophotography and they use cmos cameras witch have quite big pixel sizes and small numbers for megapixels presumably for exactly this reason. So if you used an astrophotography camera on a microscope it would do a pretty decent job? Of course they are astronomically expensive but maybe if you get sick of waiting for clear skies or during the summer time when its dark very late, you could get some extra value out of such a camera?
Hey Oliver, I recently bought a microscope and I am having some troubles with the noise on my videos. I shoot with a Nikon D3100 which shoots at 1080p 30fps @ 14MP. I have a basic trinocular port adapter directly to the camera (not a photo tube.) I mostly notice the noise when trying to take video in darkfield (which makes sense because of lower light.) Is this just pure fault of the camera sensor, or is there somthing I can do to reduce the noise in my videos? Thanks for any insight.
I will like more details on the last part of the calculation , the Camera Check, I have a 2.1 Mp (HD) camera no adapters, with 3 um image sensors, and 5.86x3.28 cm sensor. For me the required pixel sensor size was 5.5 um , and the resolution for 10x is 1.2 mp .. and gets lower with magnification. How is the 13.75 um pixel size not enough for your 3 mp camera? what strange sensor size does it have ? or adapter. Also why is the sensor size relevant here ? Do we want to project to the whole sensor the resulted image ? it will be great to give the details the the 3mp camera also, not only an example of one sample. I also think that the adapter only changes the image plan location relative to the sensor plane, so you see less of the image, so it should not change the "required" resolution for the camera. also one important question, If i change my "adapter" value from 1( none) to 0.5~) I get that I need 2.1 mp . does that mean that I need a 0.5 adapter to make the most of my microscope/camera setup ? as now I see only half of fov ? Sorry for my back english or my beginner questions but I started not long ago and there are a lot of things to understand :)
Short answer (for now, longer answer later): the 13,75 required pixel size was for an adapter with a 2.5x magnification and the USB microscope camera has a 0.5x sensor.
Die Aufloesung liegt im reellen Zwischenbild des Objektives multipliziert mit der Vergroesserung, ich bin mir ziemlich sicher, dass bezogen auf die Aufloesung des Mikroskops die Vergroesserung des Okulars oder irgendwelcher Adapteroptik keine Rolle bei der Berechnung spielen sollte, da diese Linsen das Bild nur aufblasen ohne die Aufloesung zu verbessern, so dass es ratsam waere diese Optiken wegzulassen und das reelle Zwischenbild direkt auf den Sensor der Kameras zu projezieren, alles andere ist nur eine Frage der Darstellung und haengt vor allen von der Aufloesung unseres Auges ab.
Hab ich ausprobiert, ist nicht ganz so simpel wie es hier klingt, die okulare scheinen als korrekturoptik zu wirken. Jedenfalls was die Bildqualität deutlich besser wenn ich mikroskop-objektiv + okular+ kameraobjektiv verwendet habe, als direkt nur mikroskop-objektiv auf kamera-sensor
Yes there is a difference. The total magnification is the same but the image using 20x Obj and 20x Eyepeice will be more blurry. The reason is becasue the 20x Obj can not capture enough detail (its numerical aperture is lower compared to the 40x). I explained this here: ua-cam.com/video/HIDjksiwn6c/v-deo.html
Great stuff. I've worked my way through this and my only query is about the Nyquist Criterion in this context. Is that factor of 1/2 included because pixels must be smaller than resolution limit suggests, but also not so small that there are too many pixels and they introduce unnessesary noise into the system?
Criterion states that capturing device must have at least twice the resolution. So more is better. Any less will result in loss of information. If there are more pixels, then you capture the full information, but the image becomes too large. Nyquist does not include noise, image size etc.
Excellent video. I took notes, did all of the calculations then looked at the magnification on the Amscope Sony Exmor MU530BI with the fixed FMA050 adaptor... WHAT !!!!!!!????????????? the specs say that it is 0.50 X That's HALF REDUCTION! No wonder the images at higher magnification look like garbage. Did they just sell junk equipment? Or can I put on a magnifying adaptor? or should I just scrap the camera and get a real one? thanks
the adapter magnification depends on the size of the sensor. for small sensors, they need to make the image smaller to fit it on the sensor. for larger sensors (eg DSLR) the magnification of the adapter is larger than one, again to fit it on the sensor.
So, if the sensor size reduces, my required MP goes down as well? Doesn't that mean an 18MP Phone Camera will have lower MP requirement than that of an 18MP Canon DSLR ?
Hello Oliver, I was looking at the 5 mp cameras. However, they all seem to have extremely low fps. I wanted to record in 1080p with a frame rate fast enough for the microorganisms not to look choppy when moving around. What fps is needed for smooth video? The only cameras with 24 fps or higher are much higher, +14 mp cameras. How can I get smooth 1080p video?
Excellent question. I was wondering about the exact same thing. Everything I look at is ridiculously low like 3 fps at VGA quality, or 10 fps at HD. That is unacceptable. Certainly a USB 3.0 type connection is the bare minimum since USB3 is at least 5 Gigabits/sec and HD content is around that level. Beyond that ? I think the sensor must be able to transmit the data at a higher rate but where can we find such a sensor and data transmission rate ? I don't know. Let me know if you find one. Cheers.
So if I understood correctly to increase the resolution of the image obtained the thing that affects the most is how big you can get your sensor to be and how low you can get the magnification of the adaptor to be ( Ma).
versteh ich nicht.....bei einer Aufloesung des Mikroskopobjektivs von 1 my sollten die Pixel der Kamera auch nicht viel groesser sein und da sind wir schon bei ca 50 MP........sag was dazu wenn du kannst
@@TalkingWayTooMuch nein das Mikroskop vergroessert das Objekt. Bei Verwendung einer Kamera ist nur das Objektiv des Mikroskopes noetig. Wichtig dabei ist ein moeglichst gosser Sensor der Kamera
I'm blown away with my new switft ss110. I just witnessed a copepod hatching from an egg live. I just watched a mother copepod with eggs digest phytoplankton....😯 I'm getting a camera now.
Hello, the basic rule of thumb is, that you should not change the video resolution and framerate when you render the video. Use the setting that your camera recorded. In my case, for example, the resolution of my camera is FullHD (1920x1080) at 30fps. After video editing, the final video should have the same. If you change these values, then you lower the image quality. You should also not increase the fps from eg. 25 (recording) to 30 (when exported. This will increase computation time and introduce extra frames that lower the quality. If your camera allowes for full HD, then I would use that, as this is more than enough for microscopic images. This resolution even allows you to digitally zoom in without much quality loss in editing (export resolution should still be full HD). If you have a 4K camera, then I would still record microscopic images at full HD (which is lower) because otherwise the video is too large. If you record at 4K, then I would reduce resolution to full HD in editing, but I would not change the framerate. All this is assuming you have a righ resolution camera connected. If you have a USB microscope camera, then I would take static images and add zooming in the video editing process to add some movement. I see that there is so much more to say here. So I will make a video on this! :-)
I micro solder. Amscope simul no adapter on camera. Barlow .5 on scope. Flir Blackfly GigE BFLY-PGE-23S6C-C 2.3mp 41fps 5.86um pixel 1/1.2" IMX249 59.7qe 67.37dr. Has HDR feature.
The default appears to be a WF10x I believe ... the scope I'm looking at is: 10x eyepieces, 0.7x - 4.5x (the actual head) and the barlow lens ... which seems like sticking to the 1x barlow is wise otherwise they get long and cut in to my working area ... if I want to adjust the zoomable-range ... I'd rather do it with the eyepieces.
So I bought the Swift 380T with the phototube as recommended by you and many others. This microscope does not or nor am I using any kind of camera adapter. I am basically just putting the camera over the phototube. So I do not have a figure for MA to plug into the formula. So how would that calculate? And yes, I can achieve focus with all my cameras (ZWO ASI224mc, Celestron NexImage5 and a Nikon D7000 and D7200).
DSLRs are not a problem. Their sensor is very large and they have a good signal to noise ratio. The problem is more with USB microscope cameras that sometimes pack in a lot of megapixels on a small sensor size. They might have a problem with with lower light performance (and fast shutter speeds).
I have a Canon 7D and I chose a dedicated 2mpx microscope camera. It was cheaper than the DSLR adaptor, I don't have to worry about camera shake, and it means I can get away with a monocular head. Pls don't hate me.
Yes, this is the reason why these microscope cameras exist. They allow also for easier time-lapse capturing because they can take automatic pictures and save them directly to a computer.
Yes, there are certain advantages to DSLRs, but in some cases, smaller microscope cameras can be better (eg. when you have problems with shutter shock (camera vibrations)
Can you explain vignetting that may occur when using the photo tube and certain eyepieces. I have the option to pick from a 2.5x to 4x projection eyepiece. Will both produce vignetting or will one have a larger viewing angle and not have any vignetting? I would think image sensor size plays a role and the distance from the lens... Any help would be great, kind regards!
GREAT question. There's been a trend of longer tubes (optics) for cameras and I think this is an under discussed topic that makes it very difficult to make accurate consumer-decisions ... and forces us to become product experts to avoid pitfalls.
How can you measure resolution of a microscope that is purely analog optical device? My Canon lenses will serve equally well 15 years old Rebel as they will latest full frame 5D. Nobody can really measure resolution of such a lens. I guess at one point your image will lose sharpness, it will fail to deliver more detail, but how do you really measure that? Your argument stands that 15MP camera is unnecessary for sole reason that you do not need such large files because of storage and processing requirements, or because of noise ratio of the relatively cheap sensor.
I have both the 3 megapixel dedicated microscope camera (the top quality Optixcam unit) and the same DSLR 18mpxl camera (with a microscope adapter.) In spite of your math, the DSLR CLEARLY gives a sharper image on all the magnifications. Also, using the canon software that magnifies the image by pixel spreading I get double the magnification with a screen resolution comparable to the 3 megapixel camera at the lower magnification. So clearly if you have the 18mpxl Canon, buying a $100 adapter for it is MUCH better than paying the $250 for the Optixcam.
@Gazza Boo well, the camera I bought is 38 Mb and came with an adaptor to fit into the eyepiece. The problem that I am encountering is that neither Picolay or ToupView recognise it, which is a pain.
I disagree with this calculation. You are assuming that the resolution of the camera is the size of the pixel which it is not. There are several other factors which reduce the resolution significantly, including the anti-aliasing filter and the bayer interpolation. I would argue that your calculation is ball-park for a monochrome camera and underestimates the number of pixels required for a colour camera. Being able to spread the main lobe of the PSF over 5-7 pixels is a nice case for a monochrome camera IMHO
Yes, I think josh is correct, and I do not agree with other parts of the explanation either. First of all, the Nyquist limit assumes no noise; a better criterion (used often in the microscopy community) is to have at least three pixels per feature, not two. Doing this would more than double his estimated Mp size. And the Bayer filter means that for blue or red light, only 1/4 of the pixels are actually capturing any data at all; for every blue-measuring pixel, there are three pixels which are interpolated rather than measured. In practice, for many subjects, this can be compensated by increasing pixel numbers by a factor of at leas 2. So the right number is at least 4x his estimated MP number. In addition, have my doubts that a high-pixel-count sensor is more noisy than a lower-pixel-count camera, per unit area on the sensor. Yes a small pixel is noisier than a large one, but in the final image, a given feature will have more pixels in the former case, and the noise will average out. The dynamic range issue, which IIRC was not mentioned in the video, might be the more serious drawback for a camera with small pixels.
Hey Oliver great videos. I belong to a meteorite group that could use your expertise. Great people we have zoom calls every wed. It’s a bit of a time difference but if you are ever interested I could send you the info. Take care.
IMO FullHD is good enough for most applications. There are other parameters which are more important, like framerate and latency if you wanna chase parameciums around, dynamic range if you're interested in higher magnifications or darkfield, interlacing of the video frames (to prevent wobbling) if you wanna make videos of your observations, etc. ;)
I use a FullHD ocular adapter. In a binocular microscope, you can put it in one tube and use the other with a normal 10x or 12,5x ocular to watch as you're chasing ciliates or other microorganisms without the delay and the wobbling of the camera, capturing video at the same time.
I want to see different megapixel cameras used as an example to see the actual real world results
IMPORTANT NOTE FOR PEOPLE USING SIMUL-FOCAL MICROSCOPES:
DO NOT include your eye pieces (typically 10x magnification) Only use:
(the HEAD'S magnification range) x (your barlow lens) x (camera's zoom)
The camera's image IS NOT AFFECTED BY YOUR EYE PIECE'S MAGNIFICATION (often they're 10x).
unless you're connecting the camera TO the eyepiece.
For simul-focal, where it uses a 3rd port -- only the BARLOW lens + camera lens influences the camera's zoom factor for this equation.
Thank you! :-D.
Your video alerted me to a THINKO I'd made!
I'd completely failed to consider prior to seeing your computation.
(also, probably bc I've never used // looked though a microscope / camera, but as I make buying-decisions, this will help!)
It's not only a nominal resolution of the camera, but the dynamic range of the camera is also important, especially for darkfield.
I’ve been thinking about this and your explanation is perfect. From your advice 2 MP - 5 MP is sufficient in most situations. One of the guys in the comments helpfully points out that fps video capture is reduced at higher CCD capabilities. This is an important consideration as well, in addition to signal to noise and data rate requirements. Many thanks for this great video. -Very helpful.
Microbehunter: I think you are off by a factor of 4
Your calculation are good for a monochrome sensor; but we use monochrome sensor cover with a RGB "Bayer filter".
A 8MP camera have 3x8=24 Mega smaller photosites or subpixels. In addition, it's not 1/3 of each color. The Bayer filter use 1/2 of green filter, 1/4 of blue and 1/4 of red.
(Click to get more info)
For a 4K or roughly 8MP camera (3x8=24 individual Mega subpixels):
🟩🟩12Mega are green
🟦6Mega are blue
🟥6Mega are red
BAYER FILTER🟦🟩🟥🟩🟦🟩🟥🟩🟦🟩🟥🟩🟦🟩🟥🟩:
Each of theses 24Mega individual subpixels only cover 1/4 of the space of a "regular pixel" size. Therefore, such individual subpixels don't get any information related to his respective color for 3/4 of the size of the "regular pixel" (for red and blue at least). For blue, it's like having a monochrome sensor with 4 times more pixel (compare to a same size regular sensor) and with 75% of dead pixels ⬛⬛🟦⬛⬛ and only 25% of working pixels spread apart by empty space. Each of theses bleu photosites would also be 4 times smaller compare to a 'regular pixel' ◾vs ⬛. Dam, it's hard to explain this properly!
All wavelength are not created equals🏳🌈:
The calculations of pixel size should use the worst case scenario. Instead of green 550nm light, we should use shorter wavelength such as blue 450nm or even purple 400nm light.
The Nyquist Criterion〰:
Half the wavelength is the absolute minimum for a perfect sensor and a perfect lens. If you get more data than that, it will gives you more info (not more resolution) that can be use to compensate for some of the noise. What does the electronic hardware and AI software is doing with this info? From there... the math goes sideway in mystery...
My guesstimate:
Consider all the above, your number seems to be off by a factor of at least 4x (just for the Bayer filter alone, not consider the wavelength and magical noise stuff).
So, in worst scenario (mainly violet sample and 4x objective), a camera sensor should have about 19MP
Since 4K is 8.2MP, it should be good enough for most video application in visible light microscopy, but 6K would be optimal at 19MP for the lowest magnification.
This is a very big difference with the numbers you calculated in the video (monochrome sensor). Maybe you should consider doing an update (unless I made mistake???).
A curious mind from Montréal
I've spend more than an hour to write this down.. I hope somebody will read it!
Just a comment of the definition of megapixels. Generally, the megapixels are all the actual photos sites. The individual pixels are not made up by color subpixels. To get RGB values from the Bayer filter, the missing color information is in top plated from neighboring pixels (check Wikipedia).
Bottom line: an 8MP Sensor consists of 4mp Green dots, 2MP red and 2MP blue. And the rest is interpolated
@@TalkingWayTooMuch This is also what I've understand from the Bayer filter: 8MP = 4MP green, 2MP blue 2MP red. But if you compare to a real black n white 8MP, you have 4 times more info (for blue and red, equivalent of gray shade) and 2 times more info (for green, equivalent of gray shade).
If you are looking to a blue object with a 8MP color camera, you will get 4 times less information compare to a 8MP black n white camera that has no Bayer filter (bare sensor).
With the color camera, each pixel, for the blue, is 25% of the whole pixel. For a true B/W senor, each pixel is 100% of the pixel size. So, the color camera will have less definition and will need to interpolate the missing 75% of the image.
When the image is not 100% blue, mainly white light, the interpolation is way more precise. But if you take the worst case scenario, a 100% blue (or red) object will get a 4x less information on a color pixel
At least, this is what I've understand. We can argue that most microscopic pictures/video use white light and the interpolation is not as critical.
This is a great explanation but still leaves me wondering why the Microscope industry seems oblivious to the fact that most modern microscopes are also used for taking pictures at some stage, by whatever means.
Users/customers are ill advised on what they really need.
'Dedicated' microscope cameras are not really dedicated at all and available under numerous different brand names.
You ask to see images taken by the camera you are thinking of buying and are shown misleading images supposedly taken with the camera model concerned, but are exactly the same stock images used and sent out by many different retailers.
When the camera you buy is one specifically recommended for your make and model of microscope, why is that same camera far from perfect when used on that specific model?
I use a DSLR for imaging and it generally gives very good results although needed some tweaking with tubes to get the same image size as the microscope. It still isn't par focal with the microscope and a lot of chromatic aberration is generated by the supposedly achromatic objective lenses which I can't seem to get rid it.
I would love to be able to buy a smaller dedicated microscope camera that just worked and produced reasonable results without having to mess about without image and focus matching issues.
It doesn't seem to be a lot to ask given modern technology and the volume of the market.
Once again we all seem to be at the mercy of retailers who only have one objective (which is also not achromatic) - making huge amount of money.
Unfortunately we can't all afford to go for a tailored service with Olympus or other high-end manufacturer.
There should be more thought given to providing a better service to customers who still spend a lot of money on their equipment and want reasonable results.
Finally, it is really great that there are people like Oliver around to help us all! :)
A microscope camera with a good implementation of the Sony IMX 290 sensor is a very good, tried and battle hardened compromise for budget to midrange microscopy. It is therefore quite common, so you'll find a very competitive market among them with pricing accordingly. The 290 had excellent low light performance, and if you use the camera for direct output, it can also do 1080p at 120 fps with a monitor supporting high framerates over HDMI.
Most microscope cameras are not primarily designed for still photos or occational video recording. They will certainly do that, but most modern microscope cameras are engineered for being a primary output source. To output a high resolution, very low latency video feed to a monitor instead of (only) observing through the eyepieces on the microscope. This changes many of the priorities and focus (pun intended) areas when designing them. You want a sensor package that can do crisp bright image at high framerates with very low latency, you need a graphical user interface overlay and means of controlling that (because the video has to go directly to a monitor's input, over USB requires passing through multiple interfaces, compression and decompression, which introduces way too much latency to be useful as a substitute for direct optical port view). Mpst of them also operate with *very* large field of view compared to what the type of objectives in this video enable, and are often used at lower magnification.
Your challenges with a lack of clear product information, the stock images thing etc - I believe that is not really a mark of the cameras, but rather the marketplace you're seeking these out in. If you need solid information on these cameras, seek that on forums, or from vendors that actually specialize in this equipment. General electronics sellers on AliExpress, Ebay, Banggood rarely know snything more about it than is written in the sales ad, and often they don't know what even 10% of that text actually means. Shopping this kind of stuff from them, is like shopping microscopes, barlow lenses and specializes cameras at Walmart or Lidl. Don't expect the clerks to even know which way is up on the damn things - and the manager only knows there's a demand for a thing, so they wanna capitalize. There might be a great deal _to be had_ there, because the price can be very low, but you're on your own in researching that before buying.
If you want proper knowledge and understanding of the parts from a vendor, you need to seek out vendors specializing in that equipment. You're gonna be paying more, because now there's very specific competence requirements for their employees, and speciality skill workers cost a lot more. But you also get curated product offerings. Those sellers won't offer for sale, the myriad of bad, low quality and garbage equipment out there - they have the skill and knowledge to only stock products that suit their customers, and only stock products they trust will perform well and last. Because the customer base is small and have many common work field, industry and interest points, a bad customer experience reflects back and hurts a specialized vendor far more than it hurts a general, everything-goes vendor. Because with a general electrics seller, maybe 300 people hearing of this bad experience buying that microscope camera. Of them, maybe 1-2 _might_ at some point consider buying such a camera from them. With a specialized vendor however, maybe only 60 hear of this bad experience. But of those 60, 40-50 (!) of them *will* be looking to buy a microscope camera within the next 6-24 months.
So you see, if you're willing to research some models on your own and with help from forums, reddit, youtube, industry/university contracts etc - then you might get more bang for buck with buying from a general electronics vendor from China. But if you want professional guidance, information, low risk of buying something bad - not to even mention if you want aftersales support - then you need to seek out a speciality vendor. And expect to pay a premium for that skill and support, because their exoenses ar higher, and they sell much fewer goods in total per month than general electronics vendord, so they need a bigger profit margin per sold unit.
If you want such speciality, highly competent vendors available for information and goods in the future - you have to support them by buying those products from them. If you only use them for free information and then do the actual shopping from chinese general "stuff" vendors instead, then all the sales revenue go to the big fsceless corporations like those "general stuff sellers" in China, banggood, Amazon etc. Then those specialized vendors that offer deep knowledge and understanding, professional service, skill and technical support - they go out of business from lack of sales. And once the big corporations eat up the sales in that market, it is no longer possible for such speciality vendors and shops to open new specialized shops - so then those speciality vendors and shops are gone for good, along with all the product and customer need expertise they offered.
The same is true for almost all other speciality vendors and shops, from hobby stores, knitting and sowing stores, art and crsfts supply stores and camera gear stores, to specialized car parts stores and so many more. If they don't get the sales, no shop can survive. Once the big online stuff sales corporations eat up what little was left of that market, then the speciality stores are gone for good.
That's the most important thing I required ..thanks again.. I was just thinking to buy a microscope camera n here once again u helped me to save a lot..
I wonder how much money this video has saved after people watched and shared it! Great explanation and breakdown Sir!
In another video you mentioned USB2.0 would be too slow and suggested getting a USB3.0 camera.
Is the slow speed of a USB2.0 connection related only to live streaming to a computer screen or is it extended to the recording also?
In other words would the recorded video be smoother?
I think that the USB affects latency, or the delay between the image shown on the computer and what's actually happening on stage. USB 2.0 should be able to transfer almost half a Gb per second, it should be 10 fold more than enough even for (compressed) 4K 60fps video in theory. I'm guessing..
why do you think that scientific professional microscope cameras have resolution of minimum 50 megapixels. Do you think that all of them they are fools...
@@lotharmayring6063 Do you have a link to them? I've only seen ones with smaller resolution so far.
Maybe those microscopes have a very large field of view?
Maybe they've got such resolution to be able to digitally zoom in in a picture taken at lower magnification?
@@ZioStalin scientist use optics for abeout 100 000 Euros
@@lotharmayring6063 I expect those professional microscope cameras either to have a very large sensor (way beyond the 1/2.3 / 1/2.8 1/3 inch you usually find) so the pixels are also way bigger and thus far more susceptible to light, increasing the SNR significantly (and thus also reducing noise) or that they actually feature that many pixels on a small size, but are actively cooled by peltier-elements to reduce the noise (as sensors at that size with that many pixels have extremely poor sensitivity). Either that or there's some very exquisite processing done to the image from the sensor, just like the supercrap-sensor in the average phone only produces a sensible image due to the very powerfull processing being done to the sensors' output, as the size is so small, the sensitivity is just ridiculous (and the amount of noise therefore is as well.
So this means that if i want to measure spore size at 100x objective, then a 1.3MP camera is more than enough?
did you forgot the fact the screen/monitor on wich you project/watch your video recorded by your camera induce some magnification?
with a 1.7Mp, you record a video of 1622x1084 px.
if you watch it on 1920x1080 resolution, you 'll need to magnifiy the horizontal res.
or you'll get a pixelized image
it can become worst if you have a monitor with 4k resolution : 3840x2160.
if i understand , you 'll need a 8mp to record in 4k.
what gain can i have with a 13mp then? more bright image?
It would be helpful to actually see images taken through the scope with both cameras to see the difference.!
Excellent presentation. As an EE, I was aware of Nyquist’s theorem, but had never thought of applying it to optics.
Congratulations for the video, very interesting.
These calculations would be for a monochrome sensor, right? For a color sensor with a Bayer type matrix, what would then be the ideal pixel size and the resolution in Mpx?
That’s going to be very similar because AFAIK the luminance Values are taken directly from the pixels, and the color information is in interpolated, So the resolution of a RGB sensor ( with Bayer filter) is not that much worse compared to monochrome sensor
I was very happy to happen on this video for two unrelated reasons:
1. For no practical reason I built an adapter to connect an old time lens to my Canon t7i DSLR (22mm sensor) . It worked but the pictures are fairly blurry. The lens was designed to work with 10 x 10 inch film which means that there is a huge crop factor with this lens (11.3 by my calculation). So are the images blurry because the limit set by physics or are the images blurry because 140 year old lenses weren't that great? I think there might be enough in this video to get a rough idea of the answer. At least I'm going to take a stab at it.
2. The Webb telescope uses detectors with only 40 mega pixels. Is that enough to achieve the maximum resolution given the mirror size. I tried to calculate this but I wasn't sure I was doing it correctly. It surprised me that it was enough to get the maximum resolution given that modern cameras, which often have more than 40 megapixels, have much smaller lenses than the Webb primary mirror. The Webb operates primarily in the infrared so for a given mirror size it has lower resolution. Maybe that's enough so that a higher resolution wouldn't be enough?
There was also a related reason. I had been curious about this for awhile. Obviously there was a limit where a higher resolution sensor wasn't going to provide a higher resolution image. I was happy to find out what that limit was.
Back of envelope calculation: a 1/2.5" to 1/3" sensor size (about 4mm and change x 5mm and change) at 5 megapixels meets the requirements, assuming a 0.5 reduction lens for the eyepiece camera.
for video all is limited by transfering time to the memory and you can forget about pixels and lenses, And only for this reason i think 2 megapixel is sufficent
on my opinion the resolution of an microscop is only given by the light wavelength and NA of the ojektiv. So it is not afected by ocular or adaper lenses. The size of real image produced by the objektiv is nearly the same for al microscops and this image has to fit on the sensor. Concerning this you need about 100 Megapixels for 1 my resolution getting the full image. But if you blow up your picture with optics after the objektiv i am shure that 1 megapixel camera could be sufficent. In this case you waste NA of your objectiv, for video then you can moove around the object so scanning it and that works with a 2 pixel camera
That's an excellent explanation (much clearer than that paper you referenced) and with a surprising result at the end!
I am trying to do this calculation for my circuit board soldering microscope with objective AN 0.065 @ 4.5X & 0.02 @ 0.7X. I have an 8.3pm camera with a Sony imx485 sensor 1/1.2 in. The optical diagonal is 12.86mm. The calculations I am coming up with say this camera doesn't have near enough resolution. Can anyone confirm this calculation?
I think MicrobeHunter did forgot a very important thing: he made the calculation for a black and white sensor. So, we need to multiply by four the number of MP (due to Bayer filter).
In your case, 8.3MP ok 4K is a great resolution, but it's not maxed out yet at this magnification. So, you have enough resolution to get a great image quality, but you can get even more if you wanna use a digital zooming without losing resolution. So... keep what you have and be happy!
This is very interesting and also a little depressing! Without doing all the calculations out of complete lazyness on my part. I made two big assumptions. First being you would be able to use a small camera with low number of mega pixels just as well as a heavey dslr. This seems true because my phone held to the high piece seemed to do surprisingly better than my dslr attached to microscope. Secoundly Iv recently got into astrophotography and they use cmos cameras witch have quite big pixel sizes and small numbers for megapixels presumably for exactly this reason. So if you used an astrophotography camera on a microscope it would do a pretty decent job? Of course they are astronomically expensive but maybe if you get sick of waiting for clear skies or during the summer time when its dark very late, you could get some extra value out of such a camera?
Hey Oliver, I recently bought a microscope and I am having some troubles with the noise on my videos. I shoot with a Nikon D3100 which shoots at 1080p 30fps @ 14MP. I have a basic trinocular port adapter directly to the camera (not a photo tube.) I mostly notice the noise when trying to take video in darkfield (which makes sense because of lower light.) Is this just pure fault of the camera sensor, or is there somthing I can do to reduce the noise in my videos? Thanks for any insight.
yet another information video....thank you so much oliver.....
Thank you for explaining this as throughly as you have here. Very helpful!
I will like more details on the last part of the calculation , the Camera Check, I have a 2.1 Mp (HD) camera no adapters, with 3 um image sensors, and 5.86x3.28 cm sensor. For me the required pixel sensor size was 5.5 um , and the resolution for 10x is 1.2 mp .. and gets lower with magnification. How is the 13.75 um pixel size not enough for your 3 mp camera? what strange sensor size does it have ? or adapter. Also why is the sensor size relevant here ? Do we want to project to the whole sensor the resulted image ? it will be great to give the details the the 3mp camera also, not only an example of one sample. I also think that the adapter only changes the image plan location relative to the sensor plane, so you see less of the image, so it should not change the "required" resolution for the camera. also one important question, If i change my "adapter" value from 1( none) to 0.5~) I get that I need 2.1 mp . does that mean that I need a 0.5 adapter to make the most of my microscope/camera setup ? as now I see only half of fov ? Sorry for my back english or my beginner questions but I started not long ago and there are a lot of things to understand :)
Short answer (for now, longer answer later): the 13,75 required pixel size was for an adapter with a 2.5x magnification and the USB microscope camera has a 0.5x sensor.
Thank you for producing this Video Oliver, really appreciate this. Kind regards Graham Furlonger.
i do not have any adapter , so what value should i take for 'Ma'?
Die Aufloesung liegt im reellen Zwischenbild des Objektives multipliziert mit der Vergroesserung, ich bin mir ziemlich sicher, dass bezogen auf die Aufloesung des Mikroskops die Vergroesserung des Okulars oder irgendwelcher Adapteroptik keine Rolle bei der Berechnung spielen sollte, da diese Linsen das Bild nur aufblasen ohne die Aufloesung zu verbessern, so dass es ratsam waere diese Optiken wegzulassen und das reelle Zwischenbild direkt auf den Sensor der Kameras zu projezieren, alles andere ist nur eine Frage der Darstellung und haengt vor allen von der Aufloesung unseres Auges ab.
Hab ich ausprobiert, ist nicht ganz so simpel wie es hier klingt, die okulare scheinen als korrekturoptik zu wirken. Jedenfalls was die Bildqualität deutlich besser wenn ich mikroskop-objektiv + okular+ kameraobjektiv verwendet habe, als direkt nur mikroskop-objektiv auf kamera-sensor
@@TalkingWayTooMuch die Aufloesung wird nur durch die NA des Objektivs festgelegt
How does this work for video, is it worth it to record in 4k or it wont make a difference because of the megapixel limitation of the microscope?
4k is about 8MP so it's perfect to do a little zoom and not overkill. Just make sure you get at least 24fps.
Is a 2O X objective with 2O X eye piece the Same as a 4O X objective with a 1O X eyepiece? Both would give a total magnification 4OO X .
Yes there is a difference. The total magnification is the same but the image using 20x Obj and 20x Eyepeice will be more blurry. The reason is becasue the 20x Obj can not capture enough detail (its numerical aperture is lower compared to the 40x). I explained this here: ua-cam.com/video/HIDjksiwn6c/v-deo.html
Great stuff. I've worked my way through this and my only query is about the Nyquist Criterion in this context. Is that factor of 1/2 included because pixels must be smaller than resolution limit suggests, but also not so small that there are too many pixels and they introduce unnessesary noise into the system?
Criterion states that capturing device must have at least twice the resolution. So more is better. Any less will result in loss of information. If there are more pixels, then you capture the full information, but the image becomes too large. Nyquist does not include noise, image size etc.
Excellent video. I took notes, did all of the calculations then looked at the magnification on the Amscope Sony Exmor MU530BI with the fixed FMA050 adaptor...
WHAT !!!!!!!?????????????
the specs say that it is 0.50 X
That's HALF REDUCTION!
No wonder the images at higher magnification look like garbage.
Did they just sell junk equipment? Or can I put on a magnifying adaptor? or should I just scrap the camera and get a real one?
thanks
the adapter magnification depends on the size of the sensor. for small sensors, they need to make the image smaller to fit it on the sensor. for larger sensors (eg DSLR) the magnification of the adapter is larger than one, again to fit it on the sensor.
So, if the sensor size reduces, my required MP goes down as well? Doesn't that mean an 18MP Phone Camera will have lower MP requirement than that of an 18MP Canon DSLR ?
Hello Oliver,
I was looking at the 5 mp cameras. However, they all seem to have extremely low fps. I wanted to record in 1080p with a frame rate fast enough for the microorganisms not to look choppy when moving around. What fps is needed for smooth video? The only cameras with 24 fps or higher are much higher, +14 mp cameras. How can I get smooth 1080p video?
Excellent question. I was wondering about the exact same thing. Everything I look at is ridiculously low like 3 fps at VGA quality, or 10 fps at HD. That is unacceptable.
Certainly a USB 3.0 type connection is the bare minimum since USB3 is at least 5 Gigabits/sec and HD content is around that level. Beyond that ? I think the sensor must be able to transmit the data at a higher rate but where can we find such a sensor and data transmission rate ? I don't know. Let me know if you find one.
Cheers.
So if I understood correctly to increase the resolution of the image obtained the thing that affects the most is how big you can get your sensor to be and how low you can get the magnification of the adaptor to be ( Ma).
I'm new to your site and had asked this question when you were discussing camera in another lesson. I got now, thanks.
versteh ich nicht.....bei einer Aufloesung des Mikroskopobjektivs von 1 my sollten die Pixel der Kamera auch nicht viel groesser sein und da sind wir schon bei ca 50 MP........sag was dazu wenn du kannst
Das Mikroskop vergrössert doch dein Objekt..
@@TalkingWayTooMuch nein das Mikroskop vergroessert das Objekt. Bei Verwendung einer Kamera ist nur das Objektiv des Mikroskopes noetig. Wichtig dabei ist ein moeglichst gosser Sensor der Kamera
Mit der gleichen Begründung könntest du anstatt einem Mikroskop auch einfach eine starke Lupe verwenden.
Ach wie ich internetdiskussionen liebe, lol
@@TalkingWayTooMuch ein Lichtmikroskop ist eine starke Lupe LOL
Thank you very much. What about the USB of the camera?
So it look that overall a 5 mp camera is perfect ?
I'm blown away with my new switft ss110. I just witnessed a copepod hatching from an egg live. I just watched a mother copepod with eggs digest phytoplankton....😯 I'm getting a camera now.
As usual professionally explained . but in case we need to upload video on youtube ( resolution and frame rate) what do you advise us
Hello, the basic rule of thumb is, that you should not change the video resolution and framerate when you render the video. Use the setting that your camera recorded. In my case, for example, the resolution of my camera is FullHD (1920x1080) at 30fps. After video editing, the final video should have the same. If you change these values, then you lower the image quality. You should also not increase the fps from eg. 25 (recording) to 30 (when exported. This will increase computation time and introduce extra frames that lower the quality. If your camera allowes for full HD, then I would use that, as this is more than enough for microscopic images. This resolution even allows you to digitally zoom in without much quality loss in editing (export resolution should still be full HD). If you have a 4K camera, then I would still record microscopic images at full HD (which is lower) because otherwise the video is too large. If you record at 4K, then I would reduce resolution to full HD in editing, but I would not change the framerate. All this is assuming you have a righ resolution camera connected. If you have a USB microscope camera, then I would take static images and add zooming in the video editing process to add some movement. I see that there is so much more to say here. So I will make a video on this! :-)
I micro solder. Amscope simul no adapter on camera. Barlow .5 on scope. Flir Blackfly GigE BFLY-PGE-23S6C-C 2.3mp 41fps 5.86um pixel 1/1.2" IMX249 59.7qe 67.37dr. Has HDR feature.
Please, may you be so kind to clarify if the Nyquist criterion applies to telescopes with astrophotography cameras?
And what about stereo microscopes? I don't have any information about the resolution of the optics on my stereo microscope, only magnification...
The default appears to be a WF10x I believe ... the scope I'm looking at is:
10x eyepieces, 0.7x - 4.5x (the actual head) and the barlow lens ... which seems like sticking to the 1x barlow is wise otherwise they get long and cut in to my working area ... if I want to adjust the zoomable-range ... I'd rather do it with the eyepieces.
I was told there would be no math.. 🙄
But like McDonalds, I'm Loving it
Thank You
So I bought the Swift 380T with the phototube as recommended by you and many others.
This microscope does not or nor am I using any kind of camera adapter. I am basically just putting the camera over the phototube. So I do not have a figure for MA to plug into the formula. So how would that calculate?
And yes, I can achieve focus with all my cameras (ZWO ASI224mc, Celestron NexImage5 and a Nikon D7000 and D7200).
Same question from me. I do not know where to see Microsco;pe Adaaptor magnficaiton>. I do not seem to have any
Use 1.0
Thanks, this was very helpful.
At what point does the resolution of a DSLR become too high? Is there a formula for calculating that?
DSLRs are not a problem. Their sensor is very large and they have a good signal to noise ratio. The problem is more with USB microscope cameras that sometimes pack in a lot of megapixels on a small sensor size. They might have a problem with with lower light performance (and fast shutter speeds).
thx! any recommendations for an EP camera 5 MP?
I have a Canon 7D and I chose a dedicated 2mpx microscope camera. It was cheaper than the DSLR adaptor, I don't have to worry about camera shake, and it means I can get away with a monocular head. Pls don't hate me.
Yes, this is the reason why these microscope cameras exist. They allow also for easier time-lapse capturing because they can take automatic pictures and save them directly to a computer.
@@MicrobehunterMicroscopy I must admit I like the look of a DSLR on top of a Microscope. It looks more like serious business.
Yes, there are certain advantages to DSLRs, but in some cases, smaller microscope cameras can be better (eg. when you have problems with shutter shock (camera vibrations)
@@MicrobehunterMicroscopy
I have a remote shutter for my SONY mirrorless A7R3. I never physically touch the camera hence ZERO camera shake.
Can you explain vignetting that may occur when using the photo tube and certain eyepieces. I have the option to pick from a 2.5x to 4x projection eyepiece. Will both produce vignetting or will one have a larger viewing angle and not have any vignetting? I would think image sensor size plays a role and the distance from the lens... Any help would be great, kind regards!
GREAT question. There's been a trend of longer tubes (optics) for cameras and I think this is an under discussed topic that makes it very difficult to make accurate consumer-decisions ... and forces us to become product experts to avoid pitfalls.
Super useful. Thanks. I'm learning a lot!
How can you measure resolution of a microscope that is purely analog optical device? My Canon lenses will serve equally well 15 years old Rebel as they will latest full frame 5D. Nobody can really measure resolution of such a lens. I guess at one point your image will lose sharpness, it will fail to deliver more detail, but how do you really measure that? Your argument stands that 15MP camera is unnecessary for sole reason that you do not need such large files because of storage and processing requirements, or because of noise ratio of the relatively cheap sensor.
I have both the 3 megapixel dedicated microscope camera (the top quality Optixcam unit) and the same DSLR 18mpxl camera (with a microscope adapter.) In spite of your math, the DSLR CLEARLY gives a sharper image on all the magnifications. Also, using the canon software that magnifies the image by pixel spreading I get double the magnification with a screen resolution comparable to the 3 megapixel camera at the lower magnification. So clearly if you have the 18mpxl Canon, buying a $100 adapter for it is MUCH better than paying the $250 for the Optixcam.
This could also be due to the better signal to noise ratio of the DSLR. And/or the adapter optics that might be better for one of them.
the first you should consider that the resolution of the human eye is about 4 Megapixels
Stupid question, but the answer is always MORE!!!
Lol, great info! I'm still using my Sony a7 mkiii though!
Thank you again. I wish I had watched this video before I bought my eyepiece camera!
@Gazza Boo well, the camera I bought is 38 Mb and came with an adaptor to fit into the eyepiece. The problem that I am encountering is that neither Picolay or ToupView recognise it, which is a pain.
Thank you! Your channel is a big help!
Very useful knowledge.
Thx
Thanks for another great video.
I disagree with this calculation. You are assuming that the resolution of the camera is the size of the pixel which it is not. There are several other factors which reduce the resolution significantly, including the anti-aliasing filter and the bayer interpolation. I would argue that your calculation is ball-park for a monochrome camera and underestimates the number of pixels required for a colour camera. Being able to spread the main lobe of the PSF over 5-7 pixels is a nice case for a monochrome camera IMHO
So what max MP camera would you suggest them for a colour camera?
Yes, I think josh is correct, and I do not agree with other parts of the explanation either. First of all, the Nyquist limit assumes no noise; a better criterion (used often in the microscopy community) is to have at least three pixels per feature, not two. Doing this would more than double his estimated Mp size. And the Bayer filter means that for blue or red light, only 1/4 of the pixels are actually capturing any data at all; for every blue-measuring pixel, there are three pixels which are interpolated rather than measured. In practice, for many subjects, this can be compensated by increasing pixel numbers by a factor of at leas 2. So the right number is at least 4x his estimated MP number. In addition, have my doubts that a high-pixel-count sensor is more noisy than a lower-pixel-count camera, per unit area on the sensor. Yes a small pixel is noisier than a large one, but in the final image, a given feature will have more pixels in the former case, and the noise will average out.
The dynamic range issue, which IIRC was not mentioned in the video, might be the more serious drawback for a camera with small pixels.
Like other's have said, another great video, thanks!
could you make video, Covid -19 cells are able to see and compare with what kind of microscopes Thanks for kindly videos Seyed marashi Dubai
Good one. Thanks.
Nyquist Criterion. Invented by Ryan Nyquist
Hey Oliver great videos. I belong to a meteorite group that could use your expertise. Great people we have zoom calls every wed. It’s a bit of a time difference but if you are ever interested I could send you the info. Take care.
IMO FullHD is good enough for most applications.
There are other parameters which are more important, like framerate and latency if you wanna chase parameciums around, dynamic range if you're interested in higher magnifications or darkfield, interlacing of the video frames (to prevent wobbling) if you wanna make videos of your observations, etc. ;)
I use a FullHD ocular adapter. In a binocular microscope, you can put it in one tube and use the other with a normal 10x or 12,5x ocular to watch as you're chasing ciliates or other microorganisms without the delay and the wobbling of the camera, capturing video at the same time.
Wonderful
Great,thanks👍
Ohhh now i know.....
KeeP ClaM AnD TrY astro camera
Way too expensive! DSLR is good enough and you can use it for various stuff... even astrophoto.
I bought a 20 mp microscope camera... I wish I had seen seen your video before!!!
Microbehunter: You don't need more than 1.7MP
Deconvolution: I am about to end this man's whole career