Thank you and welcome to the channel! Please let me know anything you need :) I’m happy to hear about the audio also. A mic has been on my list, so it was time.
Hey, Tony. Discovered your channel recently and it's been really helpful. But since we are talking about noise, I'd like to offer one bit of feedback to you - it might be time to invest in a quality microphone. There was low hum (noise) in this video that made it very difficult to watch. Most TH-cam creators agree - it isn't the quality of the video picture that turns people away - it's the sound. I'm learning that lesson as well...👍
Thank you so much for the feedback, truly! I just ordered a mic and it should be ready later today. I just went back and rewatched the video as I will check during editing and also after upload. I didn’t notice that noise before, however, now that you point it out, it’s definitely there! I am so sorry about that. With that said, if the mic is ready and time allows, I am going to redo the audio for tomorrow’s release once I am off work. Again, thank you so much and I hope you are enjoying the channel otherwise. Please let me know if you need any help, have questions, or other feedback :)
So, mathematically, the shot noise and read noise isn’t being calculated when you take your images. It occurs, with a certain degree of randomness per pixel, and then gets mixed into the signal. Tools like NoiseXterminator will attempt to determine how much noise there is in the image on a per pixel basis, and this noise will come from multiple sources - including read noise and shot noise. Then NoiseXterminator will subtract what it thinks is the best estimate for noise from each pixel, thus hopefully clearing up the image. But even then, NoiseXterminator can’t really calculate the read noise, it can only estimate it based on the training that has been done on the AI system.
Correct! The three main sources of noise (read noise, shot noise and dark current) are all random. The camera quality will determine how “severe” the noise is and every camera will suffer from it. As you will see in next week’s video, dark current is reduced by dark frames just like bias frames reduce read noise and image stacking reduces shot noise. Modern CMOS cameras have minimal read noise, so you don’t need to do bias frames so a really good way to help reduce this is not take an extreme amount of short exposures. You want well exposed sub exposures which will reduce the read noise (and the same is true for any camera really). Your dark and bias calibration frames are a stack of each calibration frame type which gives an average of each noise type which is then used to subtract out the noise from your final image. NoiseXterminator analyzes the remaining noise and reduces it based off of its training by analyzing the remaining average just like you mentioned. Awesome job! Do you find your images have large amounts of noise in them?
What about Dark Flats instead of Bias frames? I use NINA Flat Wizard to do Flats and Dark Flats. Should I skip Dark Flats and just do bias. What is the exposure for bias on a ASI071MC Pro., 0.001?
Great question! Dark flats are essentially flats. You’ll still want to do dark flats to calibrate your flats. Bias frames are for read noise and you can use them to optimize your darks or dark flats. Keep in mind though, your ASI071MC is a modern CMOS so it already has a low read noise so bias aren’t necessary. If you are capturing bias, it is the fastest shutter speed you can do which I believe on yours is .001 or possibly .0001.
Hey Tony, recently subscribed also, I like how you explained this, keep up the good work. I had no problems with the sound by the way. 👌👌👍👍
Thank you and welcome to the channel! Please let me know anything you need :) I’m happy to hear about the audio also. A mic has been on my list, so it was time.
Hey, Tony. Discovered your channel recently and it's been really helpful. But since we are talking about noise, I'd like to offer one bit of feedback to you - it might be time to invest in a quality microphone. There was low hum (noise) in this video that made it very difficult to watch. Most TH-cam creators agree - it isn't the quality of the video picture that turns people away - it's the sound. I'm learning that lesson as well...👍
Thank you so much for the feedback, truly! I just ordered a mic and it should be ready later today. I just went back and rewatched the video as I will check during editing and also after upload. I didn’t notice that noise before, however, now that you point it out, it’s definitely there! I am so sorry about that. With that said, if the mic is ready and time allows, I am going to redo the audio for tomorrow’s release once I am off work. Again, thank you so much and I hope you are enjoying the channel otherwise. Please let me know if you need any help, have questions, or other feedback :)
So, mathematically, the shot noise and read noise isn’t being calculated when you take your images. It occurs, with a certain degree of randomness per pixel, and then gets mixed into the signal.
Tools like NoiseXterminator will attempt to determine how much noise there is in the image on a per pixel basis, and this noise will come from multiple sources - including read noise and shot noise. Then NoiseXterminator will subtract what it thinks is the best estimate for noise from each pixel, thus hopefully clearing up the image. But even then, NoiseXterminator can’t really calculate the read noise, it can only estimate it based on the training that has been done on the AI system.
Correct! The three main sources of noise (read noise, shot noise and dark current) are all random. The camera quality will determine how “severe” the noise is and every camera will suffer from it. As you will see in next week’s video, dark current is reduced by dark frames just like bias frames reduce read noise and image stacking reduces shot noise. Modern CMOS cameras have minimal read noise, so you don’t need to do bias frames so a really good way to help reduce this is not take an extreme amount of short exposures. You want well exposed sub exposures which will reduce the read noise (and the same is true for any camera really). Your dark and bias calibration frames are a stack of each calibration frame type which gives an average of each noise type which is then used to subtract out the noise from your final image. NoiseXterminator analyzes the remaining noise and reduces it based off of its training by analyzing the remaining average just like you mentioned. Awesome job! Do you find your images have large amounts of noise in them?
What about Dark Flats instead of Bias frames? I use NINA Flat Wizard to do Flats and Dark Flats. Should I skip Dark Flats and just do bias. What is the exposure for bias on a ASI071MC Pro., 0.001?
Great question! Dark flats are essentially flats. You’ll still want to do dark flats to calibrate your flats. Bias frames are for read noise and you can use them to optimize your darks or dark flats. Keep in mind though, your ASI071MC is a modern CMOS so it already has a low read noise so bias aren’t necessary. If you are capturing bias, it is the fastest shutter speed you can do which I believe on yours is .001 or possibly .0001.