-
After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.
What if no 3D??
- Thread starter boriqua2000
- Start date
-
- Tags
- htc evo 3d
You may be right on the two sensors, I was repeating forum remarks on that one. Sorry.
novox77
Leeeroy Jennnkinnns!
You may have heard that CMOS is better than CCD or whatnot. Not necessarily. The quality of a given sensor is defined by its signal/noise ratio, and is directly related to cost. So, if you want a CMOS sensor that's sucky? No problem; your final image will still be sucky. You want a quality CMOS sensor? No problem; your final image will look better, but you will also pay for that quality with $$.
Canon and Nikon are the most respected camera manufacturers in the world. One uses CMOS sensor technology; the other uses CCD. If one tech is so much superior than the other, wouldn't both companies use the better one?
So... it's not that you want CMOS, but you want high-quality CMOS chips, which will make your pinhole 5mm cell phone camera look a little better. But are you willing to pay for the premium? Remember, the increase in price per pixel sensor is multiplied by the total number of pixels, so if you wish for more mp on that sensor, that becomes an even bigger multiplier for cost.
novox77
Leeeroy Jennnkinnns!
There is a sensor behind both cameras. Each is 5mp. The sensor is like the retina in our eyes. Each eye has a retina. Each retina sends a separate image to the brain for processing. The brain combines the two into one image with depth.
If there were only one sensor, why would there be two apertures?
If there were only one sensor, why would there be two apertures?
For paralax.
I wondered about that, if they'd had a MEMS structure or something acting as a shutter against a prism.
Sauske
Android Enthusiast
I wasn't disputing which was better out of CMOS or CCD, I was stating I'm glad that HTC added a sensor period, which the Evo 4G doesn't have. And I mention earlier depending on the cost you could add the 8 or 10mp
I know that Bluescreen has been less than reliable, ironically that's how I found out about this forum site and joined...and how BlackDynamite took it to heart
! I was referring to where he might of read that the two camera can combine and create a higher res image and would've been cool if was true
!
novox77
Leeeroy Jennnkinnns!
Maybe I'm just not on the same page as you and EarlyMon regarding the term "sensor." When I say "sensor," I'm referring to one of two things:
1) the individual pixel sensor
2) the entire matrix of pixel sensors that captures the overall image.
The pixel sensor (#1) technology is either CMOS or CCD. Every digital camera has a sensor chip(#2); that's the critical piece that converts analog light into a digital signal. Why would you have a camera lens without a sensor behind it? Likewise, I fail to see how the Evo doesn't have a sensor; it does; it's made with CCD technology and has 8 million pixels in the matrix.
While we're on terminology, let me clear up that there's no such thing as a CMOS lens. A lens is the glass in front of the sensor and has nothing to do with CMOS. The word CMOS only applies to the sensor behind the lens. When people say CMOS lens, they are referring to the entire camera which is powered by a CMOS sensor. Therefore a CMOS lens and CMOS sensor are not two distinct things. They both refer to the sensor technology.
novox77
Leeeroy Jennnkinnns!
Sorry for the repeated doubleposts
Ok, back to addressing the one-sensor theory. Let's say that the two lens apertures are bringing in light, and something inside the camera is using prisms/mirror to combine said light and shining it on one sensor chip. Now you've lost the ability to isolate left image from right. Light is analog; the chip doesn't know left from right. How are you going to then split these images out for the final display so you can isolate right image for left eye and vice versa?
This is like flattening two layers in photoshop. Once combined, you can't edit/view each layer independently.
The left and right channels need to be preserved so they can be shown independently at the time of display. That's why there MUST be two sensors; one directly behind each camera.
Ok, back to addressing the one-sensor theory. Let's say that the two lens apertures are bringing in light, and something inside the camera is using prisms/mirror to combine said light and shining it on one sensor chip. Now you've lost the ability to isolate left image from right. Light is analog; the chip doesn't know left from right. How are you going to then split these images out for the final display so you can isolate right image for left eye and vice versa?
This is like flattening two layers in photoshop. Once combined, you can't edit/view each layer independently.
The left and right channels need to be preserved so they can be shown independently at the time of display. That's why there MUST be two sensors; one directly behind each camera.
Sauske
Android Enthusiast
I understand the lens ans the sensor part, didn't know Evo 4G had CCD! I was once told by someone the reason the Evo 4G camera had issue's is because when you lay your phone down and the camera lens get scratched up and you take pics the camera is focusing on the scratches which in turn you get the bad pics...if this is true then great, but if not then CCD suck for the Evo 4G! You seem to know a lot about how the camera process operate, so Question, why when describing the camera on the Evo 4G is doesn't state the CCD?
HTC Mobile Phones - EVO Sprint - Overview
Or is it implied when you don't state CMOS that is CCD?
HTC Mobile Phones - EVO 3D(Sprint)
novox77
Leeeroy Jennnkinnns!
I did some digging. This is the camera that HTC used in the Evo:
OmniVision
It is CMOS technology. I had assumed it was CCD since you had wishlisted CMOS for the Evo 3D. For comparison, the same company provides Apple's iPhone 4 rear camera:
OmniVision
In terms of marketing to the mass public, people have been trained to identify "megapixel" with how decent the camera is. With that word so ingrained, it's not necessary to throw around distinctions like CMOS vs CCD, although both terms are also as useless as megapixel in determining camera quality.
So here's the breakdown on how to tell if a camera is good or not:
1) focal length. All cell phone cameras have a fundamental physical limitation: they have to be narrow. No one wants a thick phone. This really reduces the quality of light entering the lens.
2) aperture. This is how big the hole is behind the lens. Bigger aperture means more light gets in, resulting in better picture. Cell phone camera lenses are tiny, and therefore the aperture behind it is tiny/tinier. You pretty much have a pinhole camera, the kind kids used to make with oatmeal cans.
These two factors above all else really determine picture quality. Put a professional dSLR sensor behind a cell phone camera's optics, and you still have a crappy picture.
3) Pixel sensor quality. Each pixel in your image is rendered by a sensor. The speed in which the sensor can correctly capture the light's intensity and color determines its quality. With cheap sensors, the color/brightness reproduction is very bad especially in low light; that's why low light photos are always grainy. The grain is the inconsistency between individual pixel sensors.
4) pixel sensor density. This is the one not many people think about. If your sensor matrix is a fixed size (in the Evo, it's roughly a 1/3" x 1/3" square), and you cram 256 pixels on it, the resulting image is going to look blocky like an NES video game. Now let's say you put 5mil pixels on it. Great, now you have a sharp pic. Now you cram 8mpx. or 10. Does it get any sharper? At some point, you have diminishing returns. Your optics (points 1 and 2 above) become the limiting factor. In my experience, the cell phone camera's optics don't justify more than 2 megapixels.
Now, if you could move the sensor farther back, essentially lengthening the focal length, the cone of light shines a larger circle, so you can enlarge the entire sensor. Once you do that, you can justify cramming more pixels on the sensor because you have to cover more area. Don't know if this is getting too complicated, but basically, you have to weigh the number of pixels to the total area of the sensor. Too dense means too many pixels are not doing anything.
Ok, so why do some cell phone cameras take sharper pics? Focal length is short, and aperture is tiny. What's the difference? It's the quality of the pixel sensors on the sensor chip. Higher fidelity color/brightness reproduction. And the quality comes at a cost. The difference in cost between an entry-level dSLR and a professional one is not the added features on the body. A lot of that is just firmware crippling on the lower-end cameras. The main difference in cost is the quality of the pixel sensors used on the sensor chip, the size of the sensor chip, and the total number of pixel sensors.
Here's the most important part of this long-winded post: unfortunately, due to the megapixel marketing sham, consumers would laugh off a 2mpx camera, even if it costs as much as a crappy 10mpx camera. The image quality would be SO much better on the 2mpx because the 8-million reduction in pixel sensors could subsidize higher quality sensors.
So... educate yourselves. Don't just demand more pixel sensors (megapixels). Don't just demand CMOS or CCD. Demand higher quality / low noise sensors.
OmniVision
It is CMOS technology. I had assumed it was CCD since you had wishlisted CMOS for the Evo 3D. For comparison, the same company provides Apple's iPhone 4 rear camera:
OmniVision
In terms of marketing to the mass public, people have been trained to identify "megapixel" with how decent the camera is. With that word so ingrained, it's not necessary to throw around distinctions like CMOS vs CCD, although both terms are also as useless as megapixel in determining camera quality.
So here's the breakdown on how to tell if a camera is good or not:
1) focal length. All cell phone cameras have a fundamental physical limitation: they have to be narrow. No one wants a thick phone. This really reduces the quality of light entering the lens.
2) aperture. This is how big the hole is behind the lens. Bigger aperture means more light gets in, resulting in better picture. Cell phone camera lenses are tiny, and therefore the aperture behind it is tiny/tinier. You pretty much have a pinhole camera, the kind kids used to make with oatmeal cans.
These two factors above all else really determine picture quality. Put a professional dSLR sensor behind a cell phone camera's optics, and you still have a crappy picture.
3) Pixel sensor quality. Each pixel in your image is rendered by a sensor. The speed in which the sensor can correctly capture the light's intensity and color determines its quality. With cheap sensors, the color/brightness reproduction is very bad especially in low light; that's why low light photos are always grainy. The grain is the inconsistency between individual pixel sensors.
4) pixel sensor density. This is the one not many people think about. If your sensor matrix is a fixed size (in the Evo, it's roughly a 1/3" x 1/3" square), and you cram 256 pixels on it, the resulting image is going to look blocky like an NES video game. Now let's say you put 5mil pixels on it. Great, now you have a sharp pic. Now you cram 8mpx. or 10. Does it get any sharper? At some point, you have diminishing returns. Your optics (points 1 and 2 above) become the limiting factor. In my experience, the cell phone camera's optics don't justify more than 2 megapixels.
Now, if you could move the sensor farther back, essentially lengthening the focal length, the cone of light shines a larger circle, so you can enlarge the entire sensor. Once you do that, you can justify cramming more pixels on the sensor because you have to cover more area. Don't know if this is getting too complicated, but basically, you have to weigh the number of pixels to the total area of the sensor. Too dense means too many pixels are not doing anything.
Ok, so why do some cell phone cameras take sharper pics? Focal length is short, and aperture is tiny. What's the difference? It's the quality of the pixel sensors on the sensor chip. Higher fidelity color/brightness reproduction. And the quality comes at a cost. The difference in cost between an entry-level dSLR and a professional one is not the added features on the body. A lot of that is just firmware crippling on the lower-end cameras. The main difference in cost is the quality of the pixel sensors used on the sensor chip, the size of the sensor chip, and the total number of pixel sensors.
Here's the most important part of this long-winded post: unfortunately, due to the megapixel marketing sham, consumers would laugh off a 2mpx camera, even if it costs as much as a crappy 10mpx camera. The image quality would be SO much better on the 2mpx because the 8-million reduction in pixel sensors could subsidize higher quality sensors.
So... educate yourselves. Don't just demand more pixel sensors (megapixels). Don't just demand CMOS or CCD. Demand higher quality / low noise sensors.
Evo 5oh
Android Enthusiast
The specs are great the 3D another feature but i want it ....people say no biggie but having something other people dont have is great 
and of course competition you know how that works.... Im going to PR this summer hopefully the phone is cheap enough that i can get it. I would record my trip in 3D
and of course competition you know how that works.... Im going to PR this summer hopefully the phone is cheap enough that i can get it. I would record my trip in 3D
novox77
Leeeroy Jennnkinnns!
Points 1 and 2 are pretty much permanent limitations, since people expect thin cameras and there's already so much other hardware crammed inside your phone. Therefore, the only place we can really improve the camera is by using fewer but better pixel sensors.
Droid X... look at that phone on its profile. Notice it is fatter where the camera is. That may have helped picture quality a bit, assuming the thickness was to increase the focal length a bit.
I'm really hoping that since the Evo 3D is placing lots of emphasis on the camera (with 3D and the dedicated shutter button) that it uses higher quality sensors. I'm thinking that the central part of the panoramic pictures will look great because it's the average of two images. Error in any given sensor is nullified somewhat because its result is averaged with another sensor on the other camera.
I guess I'll go on record and wager that the full price of this phone will cost more than the Evo due to the 2 cameras and other hardware. Whether or not Sprint will subsidize more to keep it in the $200-250 range is up in the air. I suspect they will.
jdsingle
Android Expert
I'll call your record and raise you one. I don't think it will cost more even unsubsidized. Despite giving something more features, the $500-$600 is basically the customary pricing point for almost all smart phones. Whoa hold on there, for some reason I thought off contract pricing on EVO 4G was $599. Guess I'm used to the Verizon BS.
Anyways, I'm hoping this thing takes decent quality pictures. My phone is the main camera at the party. Maybe if Novox PM's BSOD and tells him all this the EVO 3D that gets shipped to everyone will get updated
Vanquished
Android Expert
Without the 3D element it becomes the HTC Pyramid
But with more ram and a better GPU.
If they both use the Qualcomm 8660, then the gpu is same.
Despite what blogs claim ZERO phones have separate or dedicated or whatever gpus.
Despite what blogs claim ZERO phones have separate or dedicated or whatever gpus.
Vanquished
Android Expert
The Pyramid has the 8260 and the E3D has the 8660; you appear to be right that they have the same GPU though. The difference between the two is the HSPA+ and the multi-mode HSPA+/CDMA2000 1xEV-DO Rev.B.
SaleenFiend
Android Enthusiast
It's the idea of being a game changer... think EVO 4G was first ever 4G phone and they also touted the kickstand.
Now EVO 3D will be first ever 3D but I will desperately miss my kickstand.. hopefully someone (such as HTC) will see the need for a case to sell seperately with a strong factory EVO4G like kickstand.
As dumb as some of you see it... 3D is going to be the next move. 3D TVs are rolling out pretty quickly and the technology is def getting better and better with more advancement. I'm not a fan of 3D but it's still cool to have.
Now EVO 3D will be first ever 3D but I will desperately miss my kickstand.. hopefully someone (such as HTC) will see the need for a case to sell seperately with a strong factory EVO4G like kickstand.
As dumb as some of you see it... 3D is going to be the next move. 3D TVs are rolling out pretty quickly and the technology is def getting better and better with more advancement. I'm not a fan of 3D but it's still cool to have.
Cobravision
Android Expert
The question is, can the Evo 3D use two different areas of a single sensor to create 2 separate images? That seems to be a logical explanation, but whether or not it's even possible is another story.
The other possibility is that the images are taken in quick succession on one sensor, quick enough that the motion change has little effect on the 3D effect (and if the motion change were fast enough, the image would be blurred anyway). Two apertures are needed for parallax, but both record to a single chip. Due to the speed needed to capture and process the image, only 2MP are captured at a time.
The reason why I tend to think there is only one chip is the 2MP limit for 3D pictures. If there were two separate 5MP sensors, then why aren't there 5MP 3D images? However, there may be another reason for the 2MP limit or, perhaps, there is no 2MP limit and that information is erroneous.
novox77
Leeeroy Jennnkinnns!
To answer all your points:
The sensor works via direct contact with light. Since the two cameras are pretty far apart, that would require a very special-shaped sensor (long rectangular shape, possibly without pixel sensors in between). Given the surplus of sensors manufactured in more traditional shapes, I tend to discount this type of sensor that spans the distance between the two apertures.
Your 2nd point about 2 quick successive shots on one sensor: that would imply that a standard-sized sensor chip quickly slides from one aperture to another to capture the parallax. I doubt it would be able to move over fast enough, and if there is even a little camera shake, the two images wouldn't align properly. Plus, it's a lot of mechanical engineering to have a sliding sensor, and IMO would be prone to failure. Again, the sensor chips works by having light shine directly on it. There's simply no space to squeeze in a ton of mirrors to direct light to some centralized location. Plus, the more mirrors you add, the more distortion you introduce.
Finally, regarding the 2mp: to me this is a software 'limitation.' It's the same as taking a 1mp picture with our current Evo camera, even though the sensor is capable of 8. This is a post-processing step that either the hardware behind the sensor is performing, or the OS is doing the downsampling. My best guess is that the 3D looks just as good at 2mp as 5mp. So rather than having to store/compress 10mp, 3D images only need to deal with 4mp.
And going back to your theory that the pixels are somehow shared on one 5mp sensor, 2mp per camera still doesn't add up to 5.
Also, as far as I know, these cell phone cameras come as a complete package: meaning that the lens and the sensor are all one unit. Again, that's not to say someone didn't make a special 3D setup just for Evo, but more likely than not, HTC used two regular ol cameras and programmed them to take pictures together, and the resulting two frames are stitched together via software to form the 3D image. It's the simplest (and therefore most cost-effective) scenario.
D
Deleted User
Guest
They've already said they're going to provide a case with a built in kickstand, and from the pics, it looks to be much more sturdy.
Vanquished
Android Expert
And less Memory (1gb vs. 4gb)