_haider
Well-known member
The sensor has 8.1M effective pixels. It is capable of recording 8.1M variations accross the scene. Like many believes the result of putting too many pixels is a cause of degraded details in an image.
As we have seen some H9 photos do have less details, the reason could be the sensor having too many pixels, or the processor thats doing the NR and compression, or a mix of both.
I have been trying to find out how detailed photo I can get from the H9 in various light conditions.
Overall, I agree with Alan in that the smearing NR artifacts show up in very low contrast gray, light brown and sometimes other colors.
Purposely underexposed indoor lowlight shots reveal the loss of detail along contrasted areas.
However, I did managed to get some shots that I believe is outstanding and proves H9's sensors ability. Here is a couple of photos from today showing quite good details:
Wide angle, indoor, typical dark environment of Red Lobster restaurant. Handheld at ISO 80, 1/13 shutter and F2.7:
It is slightly sharpened after resizing. You can see the barrel distortion. But look at the original and see there is very good detail for this lighting condition. The IS is great on H9. This photo was taken at 5MP, but I believe there won't be much difference at 8MP.
The next one taken early this morning, at the full 15x telephoto, where I wasn't quite sure if it was flowers or some red leaves, and where it actually focused.
It is taken at 8MP, handheld, ISO 80 and at the maximum F stop available at 15x:
It is cropped and resized. The focus seems to be on the buds in the middle, and everything before and after it is slightly out of focus.
Again, if you look at the originals, there is little artifacts indicating loss of detail.
The originals are here:
http://haiders.net/h9/detail/originals
Thanks for looking. The sensor, I believe has nothing to do with the problems that we are seeing sometimes. Although many are suggesting the new compression system, which appears to be dynamic, is not responsible for it either, I believe it is in someway contributing to it. But perhaps the main problem is the processor.
Sony has created an intelligent processor, and it is making decisions to various degree, depending on various factors in any given situation, mode, lighting and everything else that affects a photo.
In automatic/scene modes, it makes decisions on ISO, white balance as well as color balance. It doesn't even allow overriding these decisions. In other modes, these aspects are in our control, but the processing of an image is still fully automatic. It is possible that the processor is discarding minor contrast/detail before handing the image over to compression algorithm. This will result in much smaller file sizes even if the same level of compression is applied. The very minor details are sacrificed possibly because human eye can't detect that in normal sizes. Then of course there seems to be a dynamic compression algorithm further trying to act intelligent.
The problem is even an intelligent human makes mistakes. A new generation of atrificial intelligent processor is likely to make more mistakes.
Many of us would rather make the mistake ourselves rather than having the camera make a wrong choice.
Those who shoot fully automatic will probably be disappointed at the inconsistent behavior of the camera.
Sony should have given an option to put that system into sleep, if and when we want to.
The H9, although being a very powerful camera, is going to result in a very mixed response from users.
I am keeping mine. But Sony has to make some careful decisions here, and quick.
I think all the professional reviewers are being very sympathetic to the H series and Sony by not publishing what they think of H9 at this point. They are also waiting for Sony to answer their questions.
As we have seen some H9 photos do have less details, the reason could be the sensor having too many pixels, or the processor thats doing the NR and compression, or a mix of both.
I have been trying to find out how detailed photo I can get from the H9 in various light conditions.
Overall, I agree with Alan in that the smearing NR artifacts show up in very low contrast gray, light brown and sometimes other colors.
Purposely underexposed indoor lowlight shots reveal the loss of detail along contrasted areas.
However, I did managed to get some shots that I believe is outstanding and proves H9's sensors ability. Here is a couple of photos from today showing quite good details:
Wide angle, indoor, typical dark environment of Red Lobster restaurant. Handheld at ISO 80, 1/13 shutter and F2.7:
It is slightly sharpened after resizing. You can see the barrel distortion. But look at the original and see there is very good detail for this lighting condition. The IS is great on H9. This photo was taken at 5MP, but I believe there won't be much difference at 8MP.
The next one taken early this morning, at the full 15x telephoto, where I wasn't quite sure if it was flowers or some red leaves, and where it actually focused.
It is taken at 8MP, handheld, ISO 80 and at the maximum F stop available at 15x:
It is cropped and resized. The focus seems to be on the buds in the middle, and everything before and after it is slightly out of focus.
Again, if you look at the originals, there is little artifacts indicating loss of detail.
The originals are here:
http://haiders.net/h9/detail/originals
Thanks for looking. The sensor, I believe has nothing to do with the problems that we are seeing sometimes. Although many are suggesting the new compression system, which appears to be dynamic, is not responsible for it either, I believe it is in someway contributing to it. But perhaps the main problem is the processor.
Sony has created an intelligent processor, and it is making decisions to various degree, depending on various factors in any given situation, mode, lighting and everything else that affects a photo.
In automatic/scene modes, it makes decisions on ISO, white balance as well as color balance. It doesn't even allow overriding these decisions. In other modes, these aspects are in our control, but the processing of an image is still fully automatic. It is possible that the processor is discarding minor contrast/detail before handing the image over to compression algorithm. This will result in much smaller file sizes even if the same level of compression is applied. The very minor details are sacrificed possibly because human eye can't detect that in normal sizes. Then of course there seems to be a dynamic compression algorithm further trying to act intelligent.
The problem is even an intelligent human makes mistakes. A new generation of atrificial intelligent processor is likely to make more mistakes.
Many of us would rather make the mistake ourselves rather than having the camera make a wrong choice.
Those who shoot fully automatic will probably be disappointed at the inconsistent behavior of the camera.
Sony should have given an option to put that system into sleep, if and when we want to.
The H9, although being a very powerful camera, is going to result in a very mixed response from users.
I am keeping mine. But Sony has to make some careful decisions here, and quick.
I think all the professional reviewers are being very sympathetic to the H series and Sony by not publishing what they think of H9 at this point. They are also waiting for Sony to answer their questions.
