a6300 news

Overheat already isn't a problem if you record externally. For short clip acquisition, I use it naked. But for longer recordings I want to get around the 30 min recording limit anyway, so heat is magically fixed in the same solution. :P

Atomos Ninja 2's are in the sub $300 range these days, I love using mine with Sony apsc cams. But they need to fix the waxy skin bug once and for all.
 
Oh, no, I hope it doesn't have any overheating in 1080p, my RX100 IV is smaller and it doesn't even get warm in 1080p. I was talking about 4K (for which it will probably overheat when recording internally, and may not when recording externally... to a more expensive recorder than that Ninja 2).

BTW, now that I mention my RX100... with a smaller sensor it's easier to do higher fps, if there's no IBIS then clean 1080p 120fps should be there, right?

And of course no waxy skin bug please...
 
I'm not completely sure but that's how it seems to be: by looking at catalogs from sensor makers, designing a 12 Mpix 1/2.3" sensor with full readout at 60 fps seems to be a lot easier than making a 12 Mpix FF sensor with full readout at 24 fps.

But really, I have no idea, I'm just observing a trend. Might be wrong...
 
Samuel H said:
I'm not completely sure but that's how it seems to be: by looking at catalogs from sensor makers, designing a 12 Mpix 1/2.3" sensor with full readout at 60 fps seems to be a lot easier than making a 12 Mpix FF sensor with full readout at 24 fps.

But really, I have no idea, I'm just observing a trend. Might be wrong...
Click to expand...

I'd imagine the same way they make microprocessors faster by reducing the size of components printed on it (why CPUs are have nanometer printing size specs in their listed features). Smaller circuit paths means less energy and more efficiency, chips can be clocked higher, etc.

BTW- I'm making some progress on the waxy skin bug with the help of some very fine people over at Cinema5D.
 
M. Gilden said:
I'd imagine the same way they make microprocessors faster by reducing the size of components printed on it (why CPUs are have nanometer printing size specs in their listed features). Smaller circuit paths means less energy and more efficiency, chips can be clocked higher, etc.

BTW- I'm making some progress on the waxy skin bug with the help of some very fine people over at Cinema5D.
Click to expand...

Progress as in... firmware?
 
M. Gilden said:
I'd imagine the same way they make microprocessors faster by reducing the size of components printed on it (why CPUs are have nanometer printing size specs in their listed features). Smaller circuit paths means less energy and more efficiency, chips can be clocked higher, etc...
Click to expand...
I would say it's because the larger pixels gather more light, which translates to the higher analog values.

I any case, the overheating should be solved with the 14 nM processors that have began to come out en masse in the middle of 2015 (though, several large fabs are still working toward this). The previous cycle cameras were on the 28 nM chips, which definitely produced more heat.
 
This is kind of a side note but Sony's market share in Japan's mirrorless segment took a major dive (Olympus is way up ... which is still part owned by Sony and still uses their sensors). And that's what you get when you update your high end line - in this case, A7R/S series - and ignore your low end. Sony's execs aren't likely to be complaining, however, as profit margins are much greater on the higher end models.
 
But it would be nice if they paid a bit more attention this time and the a6100 was super nice and the a5200 amost-as-nice and cheap enough that people will buy it.
 
Samuel H said:
But it would be nice if they paid a bit more attention this time and the a6100 was super nice and the a5200 amost-as-nice and cheap enough that people will buy it.
Click to expand...
When there's a RX MKII for $1,300, a A6100 would probably be retailing at $800 or so. That represent two problems for Sony - A7SII/RII are priced at ~ four times higher and likely contribute ~ six times profits per unit sold; and the fact that an ILC with 4K would cannibalize the sales of the higher priced models with the commensurate loss of the revenue.

And so they decided to ride the more popular and, at the same time, the more profitable models. From the business perspective, it's hard to blame them.
 
DLD said:
M. Gilden said:
I'd imagine the same way they make microprocessors faster by reducing the size of components printed on it (why CPUs are have nanometer printing size specs in their listed features). Smaller circuit paths means less energy and more efficiency, chips can be clocked higher, etc.
Click to expand...
I would say it's because the larger pixels gather more light, which translates to the higher analog values.
Click to expand...
I'm no electrical engineer (although I may be related to some), but I'm not so sure this would be the case. I'd argue the opposite, much like gain applied to an audio source; a louder source signal would require less gain applied to the signal and therefore have less power drawn overall. Less gain = less power = less heat.

One could argue that this may be why the A7s is the only Sony compact camera that isn't known to overheat in 1080p recording, but that's likely more to do with the less dense sensor equaling less wiring and also less processing power to re-sample.

I any case, the overheating should be solved with the 14 nM processors that have began to come out en masse in the middle of 2015 (though, several large fabs are still working toward this). The previous cycle cameras were on the 28 nM chips, which definitely produced more heat.
Click to expand...
I would guess so, as well as welcome battery usage improvements. However, do you have any reason to believe a smaller transistor size will make it into a camera in the next generation? The chips that get used in most electronics don't keep up with the times as often as products like computers and mobile phones. Canon, for example, keeps the same Digic image processor for years across generations of products before releasing an updated one.
 
M. Gilden said:
I'm no electrical engineer (although I may be related to some),
Click to expand...
I was related to a mechanical engineer ... was ... (sad face)


but I'm not so sure this would be the case. I'd argue the opposite, much like gain applied to an audio source; a louder source signal would require less gain applied to the signal and therefore have less power drawn overall. Less gain = less power = less heat.
Click to expand...
I took physics in high school ... which was long ago ... I mean, looooong ago ... but I did sell audio back in the day and ... uh ... well, the high end amps have much beefier power supplies and a much higher current ... how that relates to sensors, I have no idea ... but I'd think that a greater light gathering sensor needs to have a much higher corresponding analog value ... sort of like ... if your amp goes up to 10 and mine to 11, the extra 1 would require double the power (if in 3 db increments) ... in that same way, a sensor that can deliver 409,000 ISO must have the corresponding analog current value from 100, 200, 400 and up, whereas a different sensor only has to go up to 12,800 or something ... no idea if this is correct but it kind of makes sense to me ... if I am wrong, my high physics teacher would probably say, "I told you so" ...

One could argue that this may be why the A7s is the only Sony compact camera that isn't known to overheat in 1080p recording, but that's likely more to do with the less dense sensor equaling less wiring and also less processing power to re-sample...
Click to expand...
I would think the heat generated during in a video recording has a lot less/almost nothing to do with the ISO portion and almost everything to do with the oversampling and whatever other quirks the engineers installed into these units. I remember having a PC that I bought in 2006 and its CPU and memory usage would go way up during any video streaming. And that's just the "easier" decoding part.

I would guess so, as well as welcome battery usage improvements. However, do you have any reason to believe a smaller transistor size will make it into a camera in the next generation? The chips that get used in most electronics don't keep up with the times as often as products like computers and mobile phones. Canon, for example, keeps the same Digic image processor for years across generations of products before releasing an updated one...
Click to expand...
Can't speak for these OEM's but, in the smartphone world, heat and power management is very important. Samsung - or was it Qualcomm? - even lost a huge order because its chips ran hotter than promised. And, if these humongous fabs move onto the next stage - it's 14-16 nM now, then 10 in a couple of years, then IBM promised 7 by 2018-20 - there simply won't be any 28 nM chips made.

The hard part to foresee if the savings will be passed onto the consumers or if the OEM's will just absorb them as higher profits. The problem with the Japanese photo-video cartel is that they're not really competing - or think they are not competing - with the smartphone industry, which is essentially a free-for-all. The next iPhone will have two front facing cameras with different focal lengths, moving smartphones closer to the DSLR quality. In other words, the smartphone companies have to innovate to survive. The photo-video folks can afford to be more incremental.
 
DLD said:
I was related to a mechanical engineer ... was ... (sad face)
Click to expand...

Sorry. :(

I took physics in high school ... which was long ago ... I mean, looooong ago ... but I did sell audio back in the day and ... uh ... well, the high end amps have much beefier power supplies and a much higher current ... how that relates to sensors, I have no idea ... but I'd think that a greater light gathering sensor needs to have a much higher corresponding analog value ... sort of like ... if your amp goes up to 10 and mine to 11, the extra 1 would require double the power (if in 3 db increments) ... in that same way, a sensor that can deliver 409,000 ISO must have the corresponding analog current value from 100, 200, 400 and up, whereas a different sensor only has to go up to 12,800 or something ... no idea if this is correct but it kind of makes sense to me ... if I am wrong, my high physics teacher would probably say, "I told you so" ...
Click to expand...
Ah, see, you're looking at it backwards, I think. In the case of the bigger amp (setting aside the obvious Spinal Tap joke about going to 11), you're talking about taking the same input signal and applying more gain to it. Of course that will require more power. But that's not how the A7S achieves its super high ISO.
Without getting too technical, ISO values are not the same as gain values. In fact, I like to say that ISOs are fake on digital cameras. With film, an ASA rating on a roll determined the light sensitivity of that roll. With digital, every sensor has a certain light gathering ability which we like to call its "native" ISO. Then, in order to hit other standard values, it must be derived from that by applying gain up or down to hit standard ISO values. A FF camera with larger photosites can collect more light in each pixel natively (oversimplification, but you get the idea) and therefore wouldn't need to apply as much gain to reach, lets say 3200 ISO, whereas a tiny little micro4/3 sensor with teeny tiny photosites would have to apply a great deal of gain to reach that sensitivity.
The A7s is a low light monster because the sensor is not only big, but the 12 megapixel density means each photosite is unusually large. Since each pixel is derived from collecting more photons at a time, the native ISO is much higher without applying any gain, and it requires LESS gain to reach higher values than other cameras.
In other words, it has to work less hard to reach a high ISO value than a less sensitive camera would.

Not sure if I explained it properly, trying to rush through this. Does that make sense?

I would think the heat generated during in a video recording has a lot less/almost nothing to do with the ISO portion and almost everything to do with the oversampling and whatever other quirks the engineers installed into these units. I remember having a PC that I bought in 2006 and its CPU and memory usage would go way up during any video streaming. And that's just the "easier" decoding part.
Click to expand...

True. Proof is in the pudding- the A5100 is a tiny cam that overheats every 15 min of video recording. But just running the camera out the HDMI to an external recorder like an Atomos Ninja, the thing can run for hours on the same battery and never overheat. It seems that the heat and battery problems are mostly from the internal encoding!

Can't speak for these OEM's but, in the smartphone world, heat and power management is very important. Samsung - or was it Qualcomm? - even lost a huge order because its chips ran hotter than promised. And, if these humongous fabs move onto the next stage - it's 14-16 nM now, then 10 in a couple of years, then IBM promised 7 by 2018-20 - there simply won't be any 28 nM chips made.

The hard part to foresee if the savings will be passed onto the consumers or if the OEM's will just absorb them as higher profits. The problem with the Japanese photo-video cartel is that they're not really competing - or think they are not competing - with the smartphone industry, which is essentially a free-for-all. The next iPhone will have two front facing cameras with different focal lengths, moving smartphones closer to the DSLR quality. In other words, the smartphone companies have to innovate to survive. The photo-video folks can afford to be more incremental.
Click to expand...

Exactly. Smartphones are an artificially accelerated consumer market that moves a lot faster than other markets. You can't really look at them as a good example of other electronics. Look at how often consumers buy a new smartphone vs how often they buy a camera. I still see people shooting with 5 year old DSLRs, but people would be embarrassed to pull out a phone that old!
 
A) I understand what you're saying on gain for the signal coming off the sensors .. so, I'll do a run-around ... let's look at cars (currently in the analog world) ... one has a 4-cyl, 2L engine, the other an 12-cyl 6L engine (we all want an AMG) ... in an semi- idle moment, the 6L will still eat up more gas and run hotter than a 2L ... so, to transfer this to the photo world, would an idle photo sensor with a super high ISO capability consume more juice than a smartphone size one, that poops out six stops earlier ...

B) as to the smartphones vs. other electronics ... I have a 25-year old (not presently working) Class-A amp under my bed ... it'd probably sound great, if it were ever fixed ...

C) forget smartphones ... someone just shot a "film" using the rear camera from Toyota Prius ...
 
DLD said:
A) I understand what you're saying on gain for the signal coming off the sensors .. so, I'll do a run-around ... let's look at cars (currently in the analog world) ... one has a 4-cyl, 2L engine, the other an 12-cyl 6L engine (we all want an AMG) ... in an semi- idle moment, the 6L will still eat up more gas and run hotter than a 2L ... so, to transfer this to the photo world, would an idle photo sensor with a super high ISO capability consume more juice than a smartphone size one, that poops out six stops earlier ...
Click to expand...

No because you are consuming more fuel to run that larger engine to move a stationary object. The photo sites on the camera are being hit by an amount of photons whether the sensor is on or off. Each one are like cups collecting rain water, bigger they are the more they collect. This doesn't require more energy, it simply happens. So if you want to compare to cars, imagine a regular camera is a car on a level street, whereas the a7s is a car already heading downhill at 50mph. You want each vehicle to move at 65mph. The a7s downward facing vehicle already moving had a natural head start and requires less energy to hit and maintain that .
Incidentally, if the goal is to go 35mph, the first car would need to use a certain amount of gas but the a7s would actually need the brakes applied!

That's what is happening here. Imagine iso values are mph. Some cameras need to apply gain to see as much as a a7s does TURNED DOWN from it's neutral native state.

I can't comment on B or C. :)
 
Last edited:
You must log in or register to reply here.
Back
Top