So then you agree with me that it's not the file size, but the number of pixels and bit depth? Good, I'm glad we agree on something.
Because the only thing that would make a file bigger--other than pixels or bit depth--is layers. So, to repeat, a huge, 200 MB multilayered Photoshop file from a 3000 pixel x 5000 pixel image will take no longer to process after loading (saturation, contrast, etc.) in Aperture than will a 1 MB JPEG file of the same dimensions and bit depth.
Again, it's not file size that matters.
I really don't know how to get this across because I'm positive that you know this but somehow we are just not getting together on it.
File pixel dimensions and bit depth are what determine the file size. This is true no matter what format the file is in, though the format will determine the size in that format.
Let's take a TIFF because that's standard and is simple to calculate.
A 24 bit 2,000 x 3,000 image will be 18MB's in size. 6 Mpixel image
A 48 bit 2,000 x 3,000 image will be 36MB's in size. 6 Mpixel image.
A 24 bit 4,000 x 6,000 image will be 72MB's in size. 24 Mpixel image.
A 48 bit 4,000 x 6,000 image will be 144MB's in size. 24 Mpixel image.
And so on.
If you compress each of those images 10 to 1 as a JPEG, then each file will be 10% of the above size, but the relationship between the sizes will remain the same.
If each of those images were to be RAW instead, then they would be about 75% of the number of pixels in the image. But again the relationship would remain the same.
Create a new, blank file of large dimensions (say 6000 x 6000 pixels).
Save it as a PSD. Save a copy as a JPEG. The PSD will be 60+ MB, the JPEG will be 2 MB or less.
Now apply a rendering filter in Photoshop (say, Clouds) to each image. Which takes longer?
Surprise! They take the same amount of time.
Of course they do!!!
If you look at the image size of the JPEG in the information list at the lower left side of the image window, you will notice that the JPEG is not 2MB, but 60MB.
You're not doing an adjustment on a saved 2MB image file, but a decompressed 60MB image file.
I just hope that you don't resave it as a JPEC after you work on it. You know that each time you open a JPEG, modify it, and resave it as a JPEG again, you are recompressing it. It's a big no no.
I have a question that is slighly unrelated, but since you seem to really be in the-know, I'll ask here anyways
Is it possible to store lossless JPEG ? I mean, if you choose to store the JPEG at 100% is it then stored lossless, or is there still a loss of quality. I know JPEG is a lossy compression algorithm, but i'm curious to know if 100% is "same quality as original" or if it means "as good as it gets".
BTW: It seems like you (melgross and bikertwin) are agreeing here, you are just using different wordings and different analogies
"8 bits" is 8 bits per channel. Since there are 3 channels (R, G, , an 8-bit-per-channel file is actually 24 bits per pixel.
Likewise, a 16-bit-per-channel file has a total of 48 bits per pixels.
This is very on the side here
I didn't realise it was 8 bits per channel... Thinking about it though, I did kinnda know 2^8 = 256, and I'm used to working with web colors from 0 - 255 for each red, green and blue
Does this mean that a CMYK file will be 4/3 the size of the same RGB file ?
If you look back a page or two, you'll see that my original response was to this:
Quote:
Originally posted by melgross
It has nothing to do with how many images are being stored on your HD. It has to do with operations on the images you have open, and on the screen.
The images used at the Photo Expo were small. No more than about 20MB each.
My average image size is over 50MB in 24 bit. Many images are in 48 bit - 4 times the file size.
Several times a month I would get file sizes over 300MB, 24 bit. working with just one of those would slow the program down. Remember that CI is not instant. It just seems that way with certain image sizes. PS is also instant with most adjustments on smaller file sizes.
What file size would the GPU and associated memory choke on?
And my response was--and still is--that the file size or file type itself is not the determiner of processing speed in Aperture. Melgross was saying here that it all depends on file size. And I'm saying, no, file size is determined by:
1. pixels
2. color depth
3. compression
4. layers
3 and 4 affect the file size, but they do *NOT* affect processing time in Aperture. (3 doesn't affect processing time Photoshop, but 4 certainly does).
So, a 200 MB multi-layer file in Photoshop that takes *forever* to process will be much quicker in Aperture because it only deals with a single layer. Photoshop has to load and juggle all those layers in memory; Aperture does not.
Please don't relate Photoshop horror stories and think they'll apply to Aperture. Some things will work differently.
I have a question that is slighly unrelated, but since you seem to really be in the-know, I'll ask here anyways
Is it possible to store lossless JPEG ? I mean, if you choose to store the JPEG at 100% is it then stored lossless, or is there still a loss of quality. I know JPEG is a lossy compression algorithm, but i'm curious to know if 100% is "same quality as original" or if it means "as good as it gets".
BTW: It seems like you (melgross and bikertwin) are agreeing here, you are just using different wordings and different analogies
You can't save a JPEG with any normal program because they all compress to some minimun extent. The usual minimum is normally 10 to 1.
It can get confusing because the number you often go by can be listed starting with either 10 or 100%. Neither of those numbers mean compression amount or percentage. They mean highest quality. It means, as you said, "as good as it gets".
By the way, there is a lossless JPEG. It's Adobe's Digital Negative.
If you look back a page or two, you'll see that my original response was to this:
And my response was--and still is--that the file size or file type itself is not the determiner of processing speed in Aperture. Melgross was saying here that it all depends on file size. And I'm saying, no, file size is determined by:
1. pixels
2. color depth
3. compression
4. layers
3 and 4 affect the file size, but they do *NOT* affect processing time in Aperture. (3 doesn't affect processing time Photoshop, but 4 certainly does).
So, a 200 MB multi-layer file in Photoshop that takes *forever* to process will be much quicker in Aperture because it only deals with a single layer. Photoshop has to load and juggle all those layers in memory; Aperture does not.
Please don't relate Photoshop horror stories and think they'll apply to Aperture. Some things will work differently.
That's what I'm saying.
I wish you would stop talking about compression and layers.
When a compressed file is opened in either Aperture or PS it is at full size. We agree on that at least.
You keep bringing up layers. Why are you doing that? I'm not talking about layers at all.
I've made it quite clear that I'm not.
Do you somehow think that file size doesn't exist except when layers are used? Are you having some problem understanding that I'm talking about the pure file?
You post as though you haven't read a thing I've said, otherwise you would understand what I mean. I've been quite clear in my explanations.
I wish you would stop talking about compression and layers.
...
You post as though you haven't read a thing I've said, otherwise you would understand what I mean. I've been quite clear in my explanations.
Not every post of mine was in reply to you. Is that what you were thinking?
I also replied to BoeManE, Teno Bell, and the cool gut. The topic has ranged from JPEGs and compression to large files in Photoshop--which could be raw, JPEG, TIFF, or PSDs. Some might have layers.
By the way, there is a lossless JPEG. It's Adobe's Digital Negative.
JPEG can compresses images both lossy AND losslessly. Typically BOTH lossy and lossless JPEG compression are applied to images (as in standard JPEG images that we are all familiar with). The lossless compression that JPEG does can be applied independantly of any lossy compression. Hence, DNG does infact incorporate lossless JPEG compression, as do most RAW files (at least the ones I know of which include Canon, Nikon, etc.) However, DNG is much more than "lossless JPEG" as you said. It could more accurately be coined "Adobe Raw", as all sensor data is maintained in the file (according to Adobe). DNG is actually being promoted as an alternative to proprietary RAW files... Though the likelihood of this actually happening is relatively slim IMO.
Not every post of mine was in reply to you. Is that what you were thinking?
I also replied to BoeManE, Teno Bell, and the cool gut. The topic has ranged from JPEGs and compression to large files in Photoshop--which could be raw, JPEG, TIFF, or PSDs. Some might have layers.
JPEG can compresses images both lossy AND losslessly. Typically BOTH lossy and lossless JPEG compression are applied to images (as in standard JPEG images that we are all familiar with). The lossless compression that JPEG does can be applied independantly of any lossy compression. Hence, DNG does infact incorporate lossless JPEG compression, as do most RAW files (at least the ones I know of which include Canon, Nikon, etc.) However, DNG is much more than "lossless JPEG" as you said. It could more accurately be coined "Adobe Raw", as all sensor data is maintained in the file (according to Adobe). DNG is actually being promoted as an alternative to proprietary RAW files... Though the likelihood of this actually happening is relatively slim IMO.
Which programs apply lossless JPEG compression other than some esoteric ones that I acknowledged earlier?
And definitely, Adobe's Digital Negative is more than JUST lossless JPEC. I just mentioned that it was an example of a lossless JPEG.
In case we are still on the discussion of whether or not RAW images can be "processed" as quickly as JPEGs or TIFFs in Aperature, please take the following into consideration.
You have mentioned that once either JPEG, PSD, TIFF, etc. images have been opened in an image editing app, they are decompressed and thus baring any discrepencies (image size, bit depth, etc) will take the same amount of time to process. I would agree based on how Photoshop handles file and would assume Aperature will handle these files similarly.
However, RAW files are different beasts entirely. "Demosaicing" a RAW file is not in any way similar decompressing a compressed file such as JPEG or PSD. It involves interpolating the color information based on the color filter mosaic that the camera has recorded, but that is after it decompresses the mosaic from any lossless compression.
But here is the real caveat... While you adjust a RAW image, reinterpolation occurs. This is the power of RAW. This is what Apple is promoting that Aperture is good at(and what Photoshop can already do in Camera RAW). All adjustments that you can do in-camera (white balance, exposure (which is really exposure compensation), brightness, contrast, sharpness, etc.) and that can be done while processing RAW files in Photoshop or Aperature, happen before reinterpolation.
At any rate, my point is that RAW files once "loaded" into Aperture will certainly take more processing power to process that other types of files, simply because of the nature of RAW - which has to do with interpolation, NOT compression.
BTW, what is cool about Aperture is that all adjustments (including the "heal brush" or whatever it is in Aperture) are simply instructions saved alongside the RAW file, which makes it darn efficient!
In case we are still on the discussion of whether or not RAW images can be "processed" as quickly as JPEGs or TIFFs in Aperature, please take the following into consideration.
You have mentioned that once either JPEG, PSD, TIFF, etc. images have been opened in an image editing app, they are decompressed and thus baring any discrepencies (image size, bit depth, etc) will take the same amount of time to process. I would agree based on how Photoshop handles file and would assume Aperature will handle these files similarly.
However, RAW files are different beasts entirely. "Demosaicing" a RAW file is not in any way similar decompressing a compressed file such as JPEG or PSD. It involves interpolating the color information based on the color filter mosaic that the camera has recorded, but that is after it decompresses the mosaic from any lossless compression.
But here is the real caveat... While you adjust a RAW image, reinterpolation occurs. This is the power of RAW. This is what Apple is promoting that Aperture is good at(and what Photoshop can already do in Camera RAW). All adjustments that you can do in-camera (white balance, exposure (which is really exposure compensation), brightness, contrast, sharpness, etc.) and that can be done while processing RAW files in Photoshop or Aperature, happen before reinterpolation.
At any rate, my point is that RAW files once "loaded" into Aperture will certainly take more processing power to process that other types of files, simply because of the nature of RAW - which has to do with interpolation, NOT compression.
BTW, what is cool about Aperture is that all adjustments (including the "heal brush" or whatever it is in Aperture) are simply instructions saved alongside the RAW file, which makes it darn efficient!
While I wasn't trying to equate demosaicing with decompression (I just mentioned that they both happen when loading the file), I wasn't thinking about the constant re-demosaicing (is that a word?!) after each step. As you say, that would make a big difference in processing time.
That extra step would be evident whenever modifying the file (changing contrast, brightness, etc.), but presumably wouldn't affect things like the Loupe or Lightbox, where--hopefully!--Aperture would use a fully cached, demosaicized image.
From reading the posts from you guys I'm pretty confused.
A JPEG is uncompressed to be manipulated, I didn't know that. I'm assuming this is saying the JPEG is uncompressed to reveal the native image of the CMOS sensor. Does this include restoring any luminance or color sub-sampling?
If a JPEG can be uncompressed to its original pixel count and bit depth. It seems the only real advantage of dealing with a huge RAW file is for large detailed images or extremely detail critical images. Other than that a RAW file is more trouble than its worth. What becomes the disadvantage or advantage of RAW vs JPEG?
In my world if an image is taken as 10 bit HD with mild compression. That is the way it stays unless you convert it to another video format. That 10 bit HD image can never be uncompressed or reveal more color or contrast than was recorded.
There are HD cameras that can record the RAW native image from the camera's CCD or CMOS sensor. The problem is how do you efficiently store 24 uncompressed 1920 X 1080 images per second. HD manufactures are at work on this because compression and color sub sampling are so detrimental.
From reading the posts from you guys I'm pretty confused.
A JPEG is uncompressed to be manipulated, I didn't know that. I'm assuming this is saying the JPEG is uncompressed to reveal the native image of the CMOS sensor. Does this include restoring any luminance or color sub-sampling?
If a JPEG can be uncompressed to its original pixel count and bit depth. It seems the only real advantage of dealing with a huge RAW file is for large detailed images or extremely detail critical images. Other than that a RAW file is more trouble than its worth. What becomes the disadvantage or advantage of RAW vs JPEG?
In my world if an image is taken as 10 bit HD with mild compression. That is the way it stays unless you convert it to another video format. That 10 bit HD image can never be uncompressed or reveal more color or contrast than was recorded.
There are HD cameras that can record the RAW native image from the camera's CCD or CMOS sensor. The problem is how do you efficiently store 24 uncompressed 1920 X 1080 images per second. HD manufactures are at work on this because compression and color sub sampling are so detrimental.
A JPEG doesn't reveal the "native image" off the sensor. Only RAW files do that. It's why they are called RAW. A JPEG is a HIGHLY processed lossy representation of the original image file.
Comments
Create a new, blank file of large dimensions (say 6000 x 6000 pixels).
Save it as a PSD. Save a copy as a JPEG. The PSD will be 60+ MB, the JPEG will be 2 MB or less.
Now apply a rendering filter in Photoshop (say, Clouds) to each image. Which takes longer?
Surprise! They take the same amount of time.
Originally posted by bikertwin
So then you agree with me that it's not the file size, but the number of pixels and bit depth? Good, I'm glad we agree on something.
Because the only thing that would make a file bigger--other than pixels or bit depth--is layers. So, to repeat, a huge, 200 MB multilayered Photoshop file from a 3000 pixel x 5000 pixel image will take no longer to process after loading (saturation, contrast, etc.) in Aperture than will a 1 MB JPEG file of the same dimensions and bit depth.
Again, it's not file size that matters.
I really don't know how to get this across because I'm positive that you know this but somehow we are just not getting together on it.
File pixel dimensions and bit depth are what determine the file size. This is true no matter what format the file is in, though the format will determine the size in that format.
Let's take a TIFF because that's standard and is simple to calculate.
A 24 bit 2,000 x 3,000 image will be 18MB's in size. 6 Mpixel image
A 48 bit 2,000 x 3,000 image will be 36MB's in size. 6 Mpixel image.
A 24 bit 4,000 x 6,000 image will be 72MB's in size. 24 Mpixel image.
A 48 bit 4,000 x 6,000 image will be 144MB's in size. 24 Mpixel image.
And so on.
If you compress each of those images 10 to 1 as a JPEG, then each file will be 10% of the above size, but the relationship between the sizes will remain the same.
If each of those images were to be RAW instead, then they would be about 75% of the number of pixels in the image. But again the relationship would remain the same.
Originally posted by bikertwin
Here's a simple test, even in Photoshop.
Create a new, blank file of large dimensions (say 6000 x 6000 pixels).
Save it as a PSD. Save a copy as a JPEG. The PSD will be 60+ MB, the JPEG will be 2 MB or less.
Now apply a rendering filter in Photoshop (say, Clouds) to each image. Which takes longer?
Surprise! They take the same amount of time.
Of course they do!!!
If you look at the image size of the JPEG in the information list at the lower left side of the image window, you will notice that the JPEG is not 2MB, but 60MB.
You're not doing an adjustment on a saved 2MB image file, but a decompressed 60MB image file.
I just hope that you don't resave it as a JPEC after you work on it. You know that each time you open a JPEG, modify it, and resave it as a JPEG again, you are recompressing it. It's a big no no.
Is it possible to store lossless JPEG ? I mean, if you choose to store the JPEG at 100% is it then stored lossless, or is there still a loss of quality. I know JPEG is a lossy compression algorithm, but i'm curious to know if 100% is "same quality as original" or if it means "as good as it gets".
BTW: It seems like you (melgross and bikertwin) are agreeing here, you are just using different wordings and different analogies
Originally posted by bikertwin
Yeah, it's confusing at first!
"8 bits" is 8 bits per channel. Since there are 3 channels (R, G,
Likewise, a 16-bit-per-channel file has a total of 48 bits per pixels.
This is very on the side here
I didn't realise it was 8 bits per channel... Thinking about it though, I did kinnda know 2^8 = 256, and I'm used to working with web colors from 0 - 255 for each red, green and blue
Does this mean that a CMYK file will be 4/3 the size of the same RGB file ?
Originally posted by melgross
It has nothing to do with how many images are being stored on your HD. It has to do with operations on the images you have open, and on the screen.
The images used at the Photo Expo were small. No more than about 20MB each.
My average image size is over 50MB in 24 bit. Many images are in 48 bit - 4 times the file size.
Several times a month I would get file sizes over 300MB, 24 bit. working with just one of those would slow the program down. Remember that CI is not instant. It just seems that way with certain image sizes. PS is also instant with most adjustments on smaller file sizes.
What file size would the GPU and associated memory choke on?
And my response was--and still is--that the file size or file type itself is not the determiner of processing speed in Aperture. Melgross was saying here that it all depends on file size. And I'm saying, no, file size is determined by:
1. pixels
2. color depth
3. compression
4. layers
3 and 4 affect the file size, but they do *NOT* affect processing time in Aperture. (3 doesn't affect processing time Photoshop, but 4 certainly does).
So, a 200 MB multi-layer file in Photoshop that takes *forever* to process will be much quicker in Aperture because it only deals with a single layer. Photoshop has to load and juggle all those layers in memory; Aperture does not.
Please don't relate Photoshop horror stories and think they'll apply to Aperture. Some things will work differently.
That's what I'm saying.
Originally posted by BoeManE
I have a question that is slighly unrelated, but since you seem to really be in the-know, I'll ask here anyways
Is it possible to store lossless JPEG ? I mean, if you choose to store the JPEG at 100% is it then stored lossless, or is there still a loss of quality. I know JPEG is a lossy compression algorithm, but i'm curious to know if 100% is "same quality as original" or if it means "as good as it gets".
BTW: It seems like you (melgross and bikertwin) are agreeing here, you are just using different wordings and different analogies
You can't save a JPEG with any normal program because they all compress to some minimun extent. The usual minimum is normally 10 to 1.
It can get confusing because the number you often go by can be listed starting with either 10 or 100%. Neither of those numbers mean compression amount or percentage. They mean highest quality. It means, as you said, "as good as it gets".
By the way, there is a lossless JPEG. It's Adobe's Digital Negative.
Originally posted by bikertwin
If you look back a page or two, you'll see that my original response was to this:
And my response was--and still is--that the file size or file type itself is not the determiner of processing speed in Aperture. Melgross was saying here that it all depends on file size. And I'm saying, no, file size is determined by:
1. pixels
2. color depth
3. compression
4. layers
3 and 4 affect the file size, but they do *NOT* affect processing time in Aperture. (3 doesn't affect processing time Photoshop, but 4 certainly does).
So, a 200 MB multi-layer file in Photoshop that takes *forever* to process will be much quicker in Aperture because it only deals with a single layer. Photoshop has to load and juggle all those layers in memory; Aperture does not.
Please don't relate Photoshop horror stories and think they'll apply to Aperture. Some things will work differently.
That's what I'm saying.
I wish you would stop talking about compression and layers.
When a compressed file is opened in either Aperture or PS it is at full size. We agree on that at least.
You keep bringing up layers. Why are you doing that? I'm not talking about layers at all.
I've made it quite clear that I'm not.
Do you somehow think that file size doesn't exist except when layers are used? Are you having some problem understanding that I'm talking about the pure file?
You post as though you haven't read a thing I've said, otherwise you would understand what I mean. I've been quite clear in my explanations.
Originally posted by melgross
I wish you would stop talking about compression and layers.
...
You post as though you haven't read a thing I've said, otherwise you would understand what I mean. I've been quite clear in my explanations.
Not every post of mine was in reply to you. Is that what you were thinking?
I also replied to BoeManE, Teno Bell, and the cool gut. The topic has ranged from JPEGs and compression to large files in Photoshop--which could be raw, JPEG, TIFF, or PSDs. Some might have layers.
Originally posted by melgross
By the way, there is a lossless JPEG. It's Adobe's Digital Negative.
JPEG can compresses images both lossy AND losslessly. Typically BOTH lossy and lossless JPEG compression are applied to images (as in standard JPEG images that we are all familiar with). The lossless compression that JPEG does can be applied independantly of any lossy compression. Hence, DNG does infact incorporate lossless JPEG compression, as do most RAW files (at least the ones I know of which include Canon, Nikon, etc.) However, DNG is much more than "lossless JPEG" as you said. It could more accurately be coined "Adobe Raw", as all sensor data is maintained in the file (according to Adobe). DNG is actually being promoted as an alternative to proprietary RAW files... Though the likelihood of this actually happening is relatively slim IMO.
Originally posted by bikertwin
Not every post of mine was in reply to you. Is that what you were thinking?
I also replied to BoeManE, Teno Bell, and the cool gut. The topic has ranged from JPEGs and compression to large files in Photoshop--which could be raw, JPEG, TIFF, or PSDs. Some might have layers.
I'm just responding to your posts that quote me.
Originally posted by 4fx
JPEG can compresses images both lossy AND losslessly. Typically BOTH lossy and lossless JPEG compression are applied to images (as in standard JPEG images that we are all familiar with). The lossless compression that JPEG does can be applied independantly of any lossy compression. Hence, DNG does infact incorporate lossless JPEG compression, as do most RAW files (at least the ones I know of which include Canon, Nikon, etc.) However, DNG is much more than "lossless JPEG" as you said. It could more accurately be coined "Adobe Raw", as all sensor data is maintained in the file (according to Adobe). DNG is actually being promoted as an alternative to proprietary RAW files... Though the likelihood of this actually happening is relatively slim IMO.
Which programs apply lossless JPEG compression other than some esoteric ones that I acknowledged earlier?
And definitely, Adobe's Digital Negative is more than JUST lossless JPEC. I just mentioned that it was an example of a lossless JPEG.
Originally posted by melgross
I'm just responding to your posts that quote me.
Dude, I said we agreed way back on the bottom of page 36. 8)
The only time I quoted you on page 37 was in reply to BoeManE. Not sure what you're getting so animated about.
You have mentioned that once either JPEG, PSD, TIFF, etc. images have been opened in an image editing app, they are decompressed and thus baring any discrepencies (image size, bit depth, etc) will take the same amount of time to process. I would agree based on how Photoshop handles file and would assume Aperature will handle these files similarly.
However, RAW files are different beasts entirely. "Demosaicing" a RAW file is not in any way similar decompressing a compressed file such as JPEG or PSD. It involves interpolating the color information based on the color filter mosaic that the camera has recorded, but that is after it decompresses the mosaic from any lossless compression.
But here is the real caveat... While you adjust a RAW image, reinterpolation occurs. This is the power of RAW. This is what Apple is promoting that Aperture is good at(and what Photoshop can already do in Camera RAW). All adjustments that you can do in-camera (white balance, exposure (which is really exposure compensation), brightness, contrast, sharpness, etc.) and that can be done while processing RAW files in Photoshop or Aperature, happen before reinterpolation.
At any rate, my point is that RAW files once "loaded" into Aperture will certainly take more processing power to process that other types of files, simply because of the nature of RAW - which has to do with interpolation, NOT compression.
BTW, what is cool about Aperture is that all adjustments (including the "heal brush" or whatever it is in Aperture) are simply instructions saved alongside the RAW file, which makes it darn efficient!
Here's a great article article on RAW capture: http://www.adobe.com/digitalimag/pdf...rawcapture.pdf
Originally posted by bikertwin
Dude, I said we agreed way back on the bottom of page 36. 8)
The only time I quoted you on page 37 was in reply to BoeManE. Not sure what you're getting so animated about.
I'm not animated kiddo, it's 3:00 am here!
Originally posted by melgross
Which programs apply lossless JPEG compression other than some esoteric ones that I acknowledged earlier?
I believe Native PSD files also use lossless JPEG compression, though I am not 100% sure. They might use more than one type of compression.
Originally posted by 4fx
I believe Native PSD files also use lossless JPEG compression, though I am not 100% sure. They might use more than one type of compression.
No, they RLE. They used to use another one, long ago.
Originally posted by 4fx
In case we are still on the discussion of whether or not RAW images can be "processed" as quickly as JPEGs or TIFFs in Aperature, please take the following into consideration.
You have mentioned that once either JPEG, PSD, TIFF, etc. images have been opened in an image editing app, they are decompressed and thus baring any discrepencies (image size, bit depth, etc) will take the same amount of time to process. I would agree based on how Photoshop handles file and would assume Aperature will handle these files similarly.
However, RAW files are different beasts entirely. "Demosaicing" a RAW file is not in any way similar decompressing a compressed file such as JPEG or PSD. It involves interpolating the color information based on the color filter mosaic that the camera has recorded, but that is after it decompresses the mosaic from any lossless compression.
But here is the real caveat... While you adjust a RAW image, reinterpolation occurs. This is the power of RAW. This is what Apple is promoting that Aperture is good at(and what Photoshop can already do in Camera RAW). All adjustments that you can do in-camera (white balance, exposure (which is really exposure compensation), brightness, contrast, sharpness, etc.) and that can be done while processing RAW files in Photoshop or Aperature, happen before reinterpolation.
At any rate, my point is that RAW files once "loaded" into Aperture will certainly take more processing power to process that other types of files, simply because of the nature of RAW - which has to do with interpolation, NOT compression.
BTW, what is cool about Aperture is that all adjustments (including the "heal brush" or whatever it is in Aperture) are simply instructions saved alongside the RAW file, which makes it darn efficient!
Here's a great article article on RAW capture: http://www.adobe.com/digitalimag/pdf...rawcapture.pdf
While I wasn't trying to equate demosaicing with decompression (I just mentioned that they both happen when loading the file), I wasn't thinking about the constant re-demosaicing (is that a word?!) after each step. As you say, that would make a big difference in processing time.
That extra step would be evident whenever modifying the file (changing contrast, brightness, etc.), but presumably wouldn't affect things like the Loupe or Lightbox, where--hopefully!--Aperture would use a fully cached, demosaicized image.
Oh, and good link. Bruce's book is excellent.
A JPEG is uncompressed to be manipulated, I didn't know that. I'm assuming this is saying the JPEG is uncompressed to reveal the native image of the CMOS sensor. Does this include restoring any luminance or color sub-sampling?
If a JPEG can be uncompressed to its original pixel count and bit depth. It seems the only real advantage of dealing with a huge RAW file is for large detailed images or extremely detail critical images. Other than that a RAW file is more trouble than its worth. What becomes the disadvantage or advantage of RAW vs JPEG?
In my world if an image is taken as 10 bit HD with mild compression. That is the way it stays unless you convert it to another video format. That 10 bit HD image can never be uncompressed or reveal more color or contrast than was recorded.
There are HD cameras that can record the RAW native image from the camera's CCD or CMOS sensor. The problem is how do you efficiently store 24 uncompressed 1920 X 1080 images per second. HD manufactures are at work on this because compression and color sub sampling are so detrimental.
Originally posted by TenoBell
From reading the posts from you guys I'm pretty confused.
A JPEG is uncompressed to be manipulated, I didn't know that. I'm assuming this is saying the JPEG is uncompressed to reveal the native image of the CMOS sensor. Does this include restoring any luminance or color sub-sampling?
If a JPEG can be uncompressed to its original pixel count and bit depth. It seems the only real advantage of dealing with a huge RAW file is for large detailed images or extremely detail critical images. Other than that a RAW file is more trouble than its worth. What becomes the disadvantage or advantage of RAW vs JPEG?
In my world if an image is taken as 10 bit HD with mild compression. That is the way it stays unless you convert it to another video format. That 10 bit HD image can never be uncompressed or reveal more color or contrast than was recorded.
There are HD cameras that can record the RAW native image from the camera's CCD or CMOS sensor. The problem is how do you efficiently store 24 uncompressed 1920 X 1080 images per second. HD manufactures are at work on this because compression and color sub sampling are so detrimental.
A JPEG doesn't reveal the "native image" off the sensor. Only RAW files do that. It's why they are called RAW. A JPEG is a HIGHLY processed lossy representation of the original image file.