This is the sample requested by my friend. :-)
In both the low stress scene and the high stress scene, the codec is working correctly. It has adjusted so that the level line of high qp may not appear as much as possible.
*According to StreamEye, there is no level line of high qp macroblocks. :-)
Thanks for feedback, mate! :-)
Although it becomes a repetition, in this setting, it has designed 80% by the premise used in 24L. However, as a result of adjusting to character gentler than BETA2, a possibility that 24H can be used may be increasing.
It seems that the codec is not wavering in the bright scene if a result is seen.
Although the explanation about this topic was added to top of the thread, does the meaning lead in this English? :-)
[The common feature]
For example, if some setting tries to use low qp with the frame size limited like Intra, many level lines of the macroblock of high qp will also appear in each frame pictures. Although it is unclear for appearance, it means that the codec is not committing this correctly. I am making the settings which can use the lowest possible qp appropriately, avoiding this problem by using the biggest possible I-frames. On the high definition scene, if the minimum qp is too low, the range of qp will spread and, as a result, many of this level lines of the macroblock of high qp will appear. Therefore, in my settings, on the high definition scene, it has adjusted so that the minimum qp may become high appropriately. On the other hand, it has adjusted so that the lowest possible qp can be positively used under the lowlight. I think that 3GOP can realize this all-round performance with the motion. Therefore, when using my settings as only for lowlight, you may feel thing insufficient only a few, although sufficient dark place performance is still given. However, also in a bright scene, the codec hardly wavers with my settings. This is the point that especially I am scrupulous. This feature is common in almost all my settings.
Tested GOLGOP3-13_SednaAQ1C_V2.1_BETA4 at 24H/80% on blowing leaves in full Sun (Panasonic 14/140 ISO 160, SS30 Vibrant -2-1-1-2) and had no write errors (Transend 32GB class 10) with 1 minute (1GB) clips. Again I was surprised because Beta2 would usually fail within 20 seconds under the same conditions. Not saying it's stable just that it held up without errors for 3 similar test shots.
I am tweaking with "natural" after a long time now.
However, if new matrix is got from @driftwood, of course, I will finish it previously. :-)
Hi, friend! Many thanks always. :-)
"vibrant"
Although I have not used until now, when seeing various persons' footage, I feel it very beautiful.
Lovely, lovely footage ishvar! And congrats bkmcwd, my good friend from Japan. Your settings get better and better.
I'm now going to spend a little more time with vibrant. Looks splendid in this vid.
"Isn't this because the focus possibly is out a little?" Quite possibly - such is the hazards of handheld panning grinz. A much better test would be static rather than panning using manual focus. In any-case I don't personally see a distinct advantage 24H over 24L when using 80% for the type of video I intend to shoot (street shots under less than ideal lighting).
Thanks always, Bob! :-)
As written also on explanation, in GOLGOP3-13_SednaAQ1C, the 80% mode can be used only in 24L. Although it may be able to use under the lowlight like your sample, on a high definition scene, I think that the bit rate exceeds tolerance level.
"The wood grain on the back of the stair risers is sharper in 24H. After repeatedly studying them I saw that the shadows in 24H were smoother in transition."
Isn't this because the focus possibly is out a little?
I compared the same image areas in the 24L and 24H at 80% clips (raw from camera). Captured a pure screen grab from each (attached) in .png format. Looped them in the image viewer for 5 seconds each. Looked at specifics in each image as they looped and halted the loop on the one I felt was best without knowing which was being displayed at that time. It was actually more difficult than I had figured to identify obvious differences in image quality. The wood grain on the back of the stair risers is sharper in 24H. After repeatedly studying them I saw that the shadows in 24H were smoother in transition. Not as large a difference as I had anticipated. 24L is first and then 24H.
@bkmcwd Update from NTSC land :-) I ran GOLGOP3-13_SednaAQ1C_V2.1_BETA4 using 24H@80% and a Transend 32GB class 10 card through the stairwell torture test and much to my surprise it hung in there with no write errors. I then repeatedly ran through the 1 minute sequence until it stopped recording at 3.99GB. This test entails panning over a lot of finely textured surfaces in poor lighting at ISO 3200. I'll test it out tomorrow in daylight on a close-up of blowing leaves... The advantage of this setting is that it provides nearly as much detail resolution of 24p Cinema with 'smooth' motion. The downside of course is no audio.
I am looking forward to the result. :-)
@bkmwcwd Great, I will download that one now. :)
Very interesting and excellent comparisons, mate! :-)
"Which setting would you prefer I use in the next low light test?"
Well, if it will be the test in the dark so far, I am interested in what happens to the result of the latest GOLGOP3-13_SednaAQ1C_V2.1_BETA4.
Thanks a lot!
@bkmcwd Here is GOLGOP3-13v4.13 compared to a few other settings in a dark room. Which setting would you prefer I use in the next low light test?
I used GOLGOP3-13v4.13 this time because I already had it ready to use.
Here are the shooting settings I used. http://www.personal-view.com/talks/discussion/comment/59978#Comment_59978
These images all had the gamma and contrast pushed in post.
Hi Nick, Of course, I would like to try it! Please send me!!! Many thanks!:-)
It looks like you're new here. If you want to get involved, click one of these buttons!