[App] AFT 1.3.2 - Ambarella firmware editor (development paused)

If you increase the bitrate on dual channel models expect problems

I suspect that's because we are just changing values with no real idea what it represents and for which cam, also just because a sd card has the ability to write 400Mbps that doesn't mean it can simultaneously write two stream at 200 Mbps, not sure how the controller deals with that and what its capable of tho.
 
The same chipset that can do 40mbit in single channel mode will struggle to do 2 x 10mbit in dual channel mode
 
Which makes perfect sense, but as you have mentioned before, manufactures don't release detailed specs so its tricky to find out what its physically capable of.

My thoughts were not to crank it up to 18 or 20Mbps or anything that high, but to incrementally increase it in small numbers to see where both the physical and thermal limit is, all fun and games when it comes to learning something new.

Just want to poke around and see whats what, there must be a way to work out which Bitrate applies to which resolution,
 
In dual channel mode it starts falling over above 20mbit combined, its well known that this is the current limitation
 
Falsifactor: How does the firmware you linked differ from the original dji firmware? What are The differences? Bitrate?
 
that's the difference in bit rates
As can be seen 12 - better than - 25
In a static picture the difference in the value of the bit rate is not very noticeable.
But the move shows the advantage of a high bit rate ...
TrendVision TV-Q5NV GPS - 8Mbit vs 22Mbit
 
Last edited:
TrendVision TV-Q5NV GPS - 8Mbit vs 22Mbit
the scene is different and the speed different.
At this bit rate has almost no effect, it affects the complex scenes - quicker change of shadow - the sun(12 original and 5 ~ YouTube).
Example
 
Last edited:
Valery, I think you understand me.
Our debate can be infinitely long, but we do not need.
The man wanted a high bit rate, he received...
 
In different versions of the Ambarella, different maximum video quality...
The optimal values of the bitrate for the:
A2 - 16-20Mbit
A5 - 18-24Mbit
A7 - 24-32Mbit
IMHO...
 
The optimal bitrate for h.264 encoded files running at 1920 x 1080, 30fps on fast moving pictures is actually around 17 to 18 Mbps.

Height x Width x FPS x compresson ratio x Motion Modifier (1 is slow, 5 very fast pace)
1920 x 1080 x 30 x 0.07 x 4 17,418,240 or in other words 17.4 Mbps

Even if you use a 5 times modifier (5 times the sensitivity of a normal moving picture) its only 21Mbps

Take a blueray movie, its approx. 40Mbps, I can half that with a good codec and you wont notice a thing had changed (although black can be a tricky aspect to deal with but that doesn't bother us) Now bare in mind that the camera in these cams are insignificant to what movies are shot with and you will soon see that bitrate above a certain level will not make any difference at all.

Course if they are using a crap codec or a poorly setup codec then things get more complicated, but when you start to hit these 17 or more figures you may find other things are holding you back.

I guess what I'm saying is, don't assume bumping up bitrates to silly levels will make the blindest bit of difference
 
The optimal bitrate for h.264 encoded files running at 1920 x 1080, 30fps on fast moving pictures is actually around 17 to 18 Mbps...
It's a theory.
I am guided by personal experience, and it showed the following results...
A2 - 16-20Mbit
A5 - 18-24Mbit
A7 - 24-32Mbit
As they say in Russia:
"The experience - the son of errors difficult ..."
or
"Theory without practice is dead, practice without theory is infertile..."
 
its a widely accepted theory, and if we want to get down too it Thermodynamics is a theory too, doesn't make it any less true.

The problem is, each human eye is different, even your own left and right eyes are different. which makes any conversation of what is actually better or not completely subjective, what one person sees as good, another sees as bad.
This is why theories help define a base line.

Generally speaking, most folk wont notice any difference above a certain level.

Its quite an interesting subject really, Its like our eyes cant see above 30FPS, but equally they can and even then some, applied motion blur selective dropping of frames, colour pallet, . our eyes and brains are extremely cleaver, they are also completely unique
 
The optimal bitrate for h.264 encoded files running at 1920 x 1080, 30fps on fast moving pictures is actually around 17 to 18 Mbps.

Height x Width x FPS x compresson ratio x Motion Modifier (1 is slow, 5 very fast pace)
1920 x 1080 x 30 x 0.07 x 4 17,418,240 or in other words 17.4 Mbps

Even if you use a 5 times modifier (5 times the sensitivity of a normal moving picture) its only 21Mbps

Take a blueray movie, its approx. 40Mbps, I can half that with a good codec and you wont notice a thing had changed (although black can be a tricky aspect to deal with but that doesn't bother us) Now bare in mind that the camera in these cams are insignificant to what movies are shot with and you will soon see that bitrate above a certain level will not make any difference at all.

Course if they are using a crap codec or a poorly setup codec then things get more complicated, but when you start to hit these 17 or more figures you may find other things are holding you back.

I guess what I'm saying is, don't assume bumping up bitrates to silly levels will make the blindest bit of difference

So, why does GoPro produce much better video compared to my Phantom's camera? Do you think that it has nothing to do that max video bitrate of GoPro 3 at 1080p 30fps is 20Mbps and Phantom 2 Vision+ camera has 12Mbps? The video that Phantom's camera produces is full of compression artifacts and jagged edges. I still haven't tested the firmware that Falsifactor made but I highly believe that cranking the bitrate up will improve the video quality.
 
Back
Top