Archive 2020 · EOS-R5....will it live up to the hype?
p.10 #1EOS-R5....will it live up to the hype?
greenfield wrote:
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.We still don’t know under what circumstances it will reach 20fps, but I’m pretty sure it will not be when shooting lossless RAW 45MP stills. The 24MP Sony A9 has a stacked sensor (fastest readout there is) and can only reach 20fps shooting compressed RAW, but it does have a huge buffer that gets around the relative bottleneck of UHS-II SD cards to some extentGood point. But it only reinforces the fact that high speed cards are useful for stills. Higher speed cards could mean less limitations on possible file quality.
Fast cards is the way to go. Huge buffers will always suffer from long clearing times.I think it's going to be large cards. If that picture isn't fake, this huge door must be hiding something big.
Maybe a heatsink
Nah, probably just the pipe terminals for the external water cooling.
alundeb wrote
wordfool wrote
alundeb wrote
Huh? The CFExpress cards are crucial for the unlimited buffer and instant buffer clearing for stills on the 1DXIII. Very important for a pro stills camera. If the R5 uses SD UHS II cards it will be a huge (uh narrow) bottleneck for the buffer for 20 fps stills shooting.Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.
Fast cards is the way to go. Huge buffers will always suffer from long clearing times.
Maybe a heatsink
Nah, probably just the pipe terminals for the external water cooling.
Feb 20, 2020 at 10:27 AM
p.10 #2EOS-R5....will it live up to the hype?
ChrisMak wrote:
Because the existing EF lenses adapt very well to the EOS-R, everybody seems to think that Canon will be content to develop a mirrorless top body without the neccessary EF lenses, but that is not like Canon.
Look at the EF lenses developed so far: they all fit a 5DIV like pro allround mirrorless body like a glove. Canon has most likely with intent worked on the lenses for the EOS-R5, and now they're here, it's time for the body. I don't expect super sports wildlife performance yet though, that is for what comes after, when they have the lenses ready.The R5 should be an all around body like the 5D IV, but at 45MP it certainly is not. 
EBHActually, for me using the 7DII and looking to upgrade, a 45mp R5 would be ideal. The 32,5mp 90D is too much for my liking, so I would hope that Canon does not bring that sensor along to a possible mirrorless 7DII, which is unsure to ever materialize anyhow. So my personal hope is for a 45-50mp FF R body. I would not know otherwise what to put the 400DOII on.
But I agree, it will not be to everyone's liking. Perhaps Canon decided that they had to put something against the Nikon Z7 and the Sony A7RIV, and could not...Show more →
I haven't really followed the R5 discussion because it does not look like the camera I would be interested in. But I guess it would fit Canon marketing strategy to introduce a 40-50MP ML camera with improved AF/tracking compared to EOS R that can run even faster (framerate, EVF, storage etc.) in APS-C crop mode, making it a potential 7D2 replacement plus it looks good on the spec sheet. And then maybe at later time a cheaper version with real APS-C sensor and even later a hires version that has the same high performance running in FF mode. Milking the market, they are really good at that
EB-1 wrote
ChrisMak wrote
The EOS-R5 should logically be an allround mirrorless pro body like the 5DIV is an allround pro dslr. There is not yet a reason to release a 1DX like mirrorless AF and speed monster because Canon still needs a couple of years to make the RF lenses for such a body, like a 300mm f2.8, 500mm f4, 400mm f2.8Because the existing EF lenses adapt very well to the EOS-R, everybody seems to think that Canon will be content to develop a mirrorless top body without the neccessary EF lenses, but that is not like Canon.
Look at the EF lenses developed so far: they all fit a 5DIV like pro allround mirrorless body like a glove. Canon has most likely with intent worked on the lenses for the EOS-R5, and now they're here, it's time for the body. I don't expect super sports wildlife performance yet though, that is for what comes after, when they have the lenses ready.
EBH
But I agree, it will not be to everyone's liking. Perhaps Canon decided that they had to put something against the Nikon Z7 and the Sony A7RIV, and could not...Show more →
I haven't really followed the R5 discussion because it does not look like the camera I would be interested in. But I guess it would fit Canon marketing strategy to introduce a 40-50MP ML camera with improved AF/tracking compared to EOS R that can run even faster (framerate, EVF, storage etc.) in APS-C crop mode, making it a potential 7D2 replacement plus it looks good on the spec sheet. And then maybe at later time a cheaper version with real APS-C sensor and even later a hires version that has the same high performance running in FF mode. Milking the market, they are really good at that
Feb 20, 2020 at 10:35 AM
p.10 #3EOS-R5....will it live up to the hype?
alundeb wrote:
Not to mention that the card slot remained SD ... i.e. current consumer oriented (which is fine for the price point). As others are moving toward faster card types, I couldn't quite see investing in something that wasn't future oriented. Again, nice enough for the price point (kinda sorta) ... i.e. still not a home run.
Pondering if a used R might become a backup bargain once the 5R hits the stores. Imo, anything that I'd be considering as a "backup bargain" ... well, that just doesn't quite conjure up the notion of Home Run, either.On a camera like this it makes almost NO difference that the cards are SD. SD are cheap , easy to find and more than fast enough.
On the 1Dx ll and lll the only reason to need a faster card is not to shoot images but to shoot full-blown either 4K or 8K video.
If you are not gonna shoot more than HD video then it's a waste of money to use ultra expensive CFast and CFexpress cards!!! Being a consumer camera or a pro camera has nothing to do with it. Unfortunately the cameras have to be spec'd for their highest use, and that's video!
JohnHuh? The CFExpress cards are crucial for the unlimited buffer and instant buffer clearing for stills on the 1DXIII. Very important for a pro stills camera. If the R5 uses SD UHS II cards it will be a huge (uh narrow) bottleneck for the buffer for 20 fps stills shooting.
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.I would be very surprised if the DIGIC inside the R5 could sustain 20 fps beyond the initial DRAM frame buffer. Shuffling data off the image sensor into memory @ 20fps is the easy part. Getting 45MP of data subsequently processed through DIGIC and ready to write out to CFE is a different matter. In other words, I would predict the media card wont be the bottleneck for the sustained fps.That doesn't sound right to me. If the buffer is before Digic, why do all cameras have unlimited buffer for jpg but very limited buffer for RAW?
Few reasons. Mirrorless bodies place more demands on internal bus and memory bandwidth because they're constantly having to move and process image data to support LV and the on-sensor focusing system. Next, the reduced size of the jpg means much lower bandwidth consumption on the shared internal memory when writing out to the media card interface, which means more internal bandwidth available for the multiple trips the intermediate raw data has to move in and out of DIGIC to be processed.
snapsy wrote
alundeb wrote
JohnSil wrote
RustyBug wrote
artsupreme wrote
You forgot about the lovely single card slot, the mega speed fps, the amazing touch bar, etc....the R was a walk to first base (batter was hit by a pitch)...Pondering if a used R might become a backup bargain once the 5R hits the stores. Imo, anything that I'd be considering as a "backup bargain" ... well, that just doesn't quite conjure up the notion of Home Run, either.
On the 1Dx ll and lll the only reason to need a faster card is not to shoot images but to shoot full-blown either 4K or 8K video.
If you are not gonna shoot more than HD video then it's a waste of money to use ultra expensive CFast and CFexpress cards!!! Being a consumer camera or a pro camera has nothing to do with it. Unfortunately the cameras have to be spec'd for their highest use, and that's video!
John
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.
Few reasons. Mirrorless bodies place more demands on internal bus and memory bandwidth because they're constantly having to move and process image data to support LV and the on-sensor focusing system. Next, the reduced size of the jpg means much lower bandwidth consumption on the shared internal memory when writing out to the media card interface, which means more internal bandwidth available for the multiple trips the intermediate raw data has to move in and out of DIGIC to be processed.
Feb 20, 2020 at 10:35 AM
p.10 #4EOS-R5....will it live up to the hype?
snapsy wrote:
Pondering if a used R might become a backup bargain once the 5R hits the stores. Imo, anything that I'd be considering as a "backup bargain" ... well, that just doesn't quite conjure up the notion of Home Run, either.On a camera like this it makes almost NO difference that the cards are SD. SD are cheap , easy to find and more than fast enough.
On the 1Dx ll and lll the only reason to need a faster card is not to shoot images but to shoot full-blown either 4K or 8K video.
If you are not gonna shoot more than HD video then it's a waste of money to use ultra expensive CFast and CFexpress cards!!! Being a consumer camera or a pro camera has nothing to do with it. Unfortunately the cameras have to be spec'd for their highest use, and that's video!
JohnHuh? The CFExpress cards are crucial for the unlimited buffer and instant buffer clearing for stills on the 1DXIII. Very important for a pro stills camera. If the R5 uses SD UHS II cards it will be a huge (uh narrow) bottleneck for the buffer for 20 fps stills shooting.
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.I would be very surprised if the DIGIC inside the R5 could sustain 20 fps beyond the initial DRAM frame buffer. Shuffling data off the image sensor into memory @ 20fps is the easy part. Getting 45MP of data subsequently processed through DIGIC and ready to write out to CFE is a different matter. In other words, I would predict the media card wont be the bottleneck for the sustained fps.That doesn't sound right to me. If the buffer is before Digic, why do all cameras have unlimited buffer for jpg but very limited buffer for RAW?Few reasons. Mirrorless bodies place more demands on internal bus and memory bandwidth because they're constantly having to move and process image data to support LV and the on-sensor focusing system. Next, the reduced size of the jpg means much lower bandwidth consumption on the shared internal memory when writing out to the media card interface, which means more internal bandwidth available for the multiple trips the intermediate raw data has to move in and out of DIGIC to be processed.
It still doesn't add up for me. A consequence of what you say, is that moving raw data to the card would only be a small part the total bandwidth requirement.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
alundeb wrote
snapsy wrote
alundeb wrote
JohnSil wrote
RustyBug wrote
Not to mention that the card slot remained SD ... i.e. current consumer oriented (which is fine for the price point). As others are moving toward faster card types, I couldn't quite see investing in something that wasn't future oriented. Again, nice enough for the price point (kinda sorta) ... i.e. still not a home run.Pondering if a used R might become a backup bargain once the 5R hits the stores. Imo, anything that I'd be considering as a "backup bargain" ... well, that just doesn't quite conjure up the notion of Home Run, either.
On the 1Dx ll and lll the only reason to need a faster card is not to shoot images but to shoot full-blown either 4K or 8K video.
If you are not gonna shoot more than HD video then it's a waste of money to use ultra expensive CFast and CFexpress cards!!! Being a consumer camera or a pro camera has nothing to do with it. Unfortunately the cameras have to be spec'd for their highest use, and that's video!
John
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.
It still doesn't add up for me. A consequence of what you say, is that moving raw data to the card would only be a small part the total bandwidth requirement.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
Feb 20, 2020 at 10:48 AM
p.10 #5EOS-R5....will it live up to the hype?
alundeb wrote:
On the 1Dx ll and lll the only reason to need a faster card is not to shoot images but to shoot full-blown either 4K or 8K video.
If you are not gonna shoot more than HD video then it's a waste of money to use ultra expensive CFast and CFexpress cards!!! Being a consumer camera or a pro camera has nothing to do with it. Unfortunately the cameras have to be spec'd for their highest use, and that's video!
JohnHuh? The CFExpress cards are crucial for the unlimited buffer and instant buffer clearing for stills on the 1DXIII. Very important for a pro stills camera. If the R5 uses SD UHS II cards it will be a huge (uh narrow) bottleneck for the buffer for 20 fps stills shooting.
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.I would be very surprised if the DIGIC inside the R5 could sustain 20 fps beyond the initial DRAM frame buffer. Shuffling data off the image sensor into memory @ 20fps is the easy part. Getting 45MP of data subsequently processed through DIGIC and ready to write out to CFE is a different matter. In other words, I would predict the media card wont be the bottleneck for the sustained fps.That doesn't sound right to me. If the buffer is before Digic, why do all cameras have unlimited buffer for jpg but very limited buffer for RAW?Few reasons. Mirrorless bodies place more demands on internal bus and memory bandwidth because they're constantly having to move and process image data to support LV and the on-sensor focusing system. Next, the reduced size of the jpg means much lower bandwidth consumption on the shared internal memory when writing out to the media card interface, which means more internal bandwidth available for the multiple trips the intermediate raw data has to move in and out of DIGIC to be processed.It still doesn't add up for me. A consequence of what you say, is that moving raw data to the card would only be a small part the total bandwidth requirement.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
I recently completed a similar analysis on the Nikon Z bodies, where I found they were Expeed//system bandwidth limited. I haven't compared them to the D850 since I don't own one but IR's tests show the D850 can shoot limitless jpgs whereas the Z's do not. There may be some unique image processing requirements on the Z's that the R5 won't have or for which DIGIC is better optimized for but nonetheless the bandwidth demands for mirrorless are much higher. Here's a link to my analysis on the Z:
https://www.dpreview.com/forums/post/63645424
snapsy wrote
alundeb wrote
snapsy wrote
alundeb wrote
JohnSil wrote
On a camera like this it makes almost NO difference that the cards are SD. SD are cheap , easy to find and more than fast enough.On the 1Dx ll and lll the only reason to need a faster card is not to shoot images but to shoot full-blown either 4K or 8K video.
If you are not gonna shoot more than HD video then it's a waste of money to use ultra expensive CFast and CFexpress cards!!! Being a consumer camera or a pro camera has nothing to do with it. Unfortunately the cameras have to be spec'd for their highest use, and that's video!
John
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
I recently completed a similar analysis on the Nikon Z bodies, where I found they were Expeed//system bandwidth limited. I haven't compared them to the D850 since I don't own one but IR's tests show the D850 can shoot limitless jpgs whereas the Z's do not. There may be some unique image processing requirements on the Z's that the R5 won't have or for which DIGIC is better optimized for but nonetheless the bandwidth demands for mirrorless are much higher. Here's a link to my analysis on the Z:
https://www.dpreview.com/forums/post/63645424
Feb 20, 2020 at 10:59 AM
p.10 #6EOS-R5....will it live up to the hype?
Unlikely that the 20FPS e-shutter will be usable for action without encountering distortion. The 1DXIII reportedly got to 1/60 readout speed but I doubt this R5 will improve on that. You need to get to A9 territory at 1/160.
Feb 20, 2020 at 11:05 AM
p.10 #7EOS-R5....will it live up to the hype?
evertdoorn wrote:
Not sure how compressed raw on the A9 is but c-raw (lossy compressed) on Canon is great because there’s virtually no visible difference in IQ in real world practice.
It’s probably about the same as Canon’s c-RAW — potential artifacts in certain situations if you look hard enough but generally no visible difference for the vast majority of situations (yet still subject to rancorous debate, of course). The big problem with Sony is that it doesn’t offer lossless compression, so the options are either giant uncompressed RAW files or smaller lossy RAW files.
wordfool wrote
We still don’t know under what circumstances it will reach 20fps, but I’m pretty sure it will not be when shooting lossless RAW 45MP stills. The 24MP Sony A9 has a stacked sensor (fastest readout there is) and can only reach 20fps shooting compressed RAW, but it does have a huge buffer that gets around the relative bottleneck of UHS-II SD cards to some extentIt’s probably about the same as Canon’s c-RAW — potential artifacts in certain situations if you look hard enough but generally no visible difference for the vast majority of situations (yet still subject to rancorous debate, of course). The big problem with Sony is that it doesn’t offer lossless compression, so the options are either giant uncompressed RAW files or smaller lossy RAW files.
Feb 20, 2020 at 11:07 AM
p.10 #8EOS-R5....will it live up to the hype?
If you don't enjoy sarcasm, please don't watch this and please don't complain about it...you have been warned 
Feb 20, 2020 at 11:11 AM
p.10 #9EOS-R5....will it live up to the hype?
snapsy wrote:
Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.I would be very surprised if the DIGIC inside the R5 could sustain 20 fps beyond the initial DRAM frame buffer. Shuffling data off the image sensor into memory @ 20fps is the easy part. Getting 45MP of data subsequently processed through DIGIC and ready to write out to CFE is a different matter. In other words, I would predict the media card wont be the bottleneck for the sustained fps.That doesn't sound right to me. If the buffer is before Digic, why do all cameras have unlimited buffer for jpg but very limited buffer for RAW?Few reasons. Mirrorless bodies place more demands on internal bus and memory bandwidth because they're constantly having to move and process image data to support LV and the on-sensor focusing system. Next, the reduced size of the jpg means much lower bandwidth consumption on the shared internal memory when writing out to the media card interface, which means more internal bandwidth available for the multiple trips the intermediate raw data has to move in and out of DIGIC to be processed.It still doesn't add up for me. A consequence of what you say, is that moving raw data to the card would only be a small part the total bandwidth requirement.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.I recently completed a similar analysis on the Nikon Z bodies, where I found they were Expeed//system bandwidth limited. I haven't compared them to the D850 since I don't own one but IR's tests show the D850 can shoot limitless jpgs whereas the Z's do not. There may be some unique image processing requirements on the Z's that the R5 won't have or for which DIGIC is better optimized for but nonetheless the bandwidth demands for mirrorless are much higher. Here's a link to my analysis on the Z:
https://www.dpreview.com/forums/post/63645424
So you don't know how these processors are designed. You try to deduce it from performance analysis. Sure a fun exercise.
alundeb wrote
snapsy wrote
alundeb wrote
snapsy wrote
alundeb wrote
Huh? The CFExpress cards are crucial for the unlimited buffer and instant buffer clearing for stills on the 1DXIII. Very important for a pro stills camera. If the R5 uses SD UHS II cards it will be a huge (uh narrow) bottleneck for the buffer for 20 fps stills shooting.Encoded 8K video in 16:9 at 30 fps is a piece of cake compared to the full sensor in lossless RAW at 20 fps.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
https://www.dpreview.com/forums/post/63645424
So you don't know how these processors are designed. You try to deduce it from performance analysis. Sure a fun exercise.
Feb 20, 2020 at 11:12 AM
p.10 #10EOS-R5....will it live up to the hype?
alundeb wrote:
That doesn't sound right to me. If the buffer is before Digic, why do all cameras have unlimited buffer for jpg but very limited buffer for RAW?Few reasons. Mirrorless bodies place more demands on internal bus and memory bandwidth because they're constantly having to move and process image data to support LV and the on-sensor focusing system. Next, the reduced size of the jpg means much lower bandwidth consumption on the shared internal memory when writing out to the media card interface, which means more internal bandwidth available for the multiple trips the intermediate raw data has to move in and out of DIGIC to be processed.It still doesn't add up for me. A consequence of what you say, is that moving raw data to the card would only be a small part the total bandwidth requirement.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.I recently completed a similar analysis on the Nikon Z bodies, where I found they were Expeed//system bandwidth limited. I haven't compared them to the D850 since I don't own one but IR's tests show the D850 can shoot limitless jpgs whereas the Z's do not. There may be some unique image processing requirements on the Z's that the R5 won't have or for which DIGIC is better optimized for but nonetheless the bandwidth demands for mirrorless are much higher. Here's a link to my analysis on the Z:
https://www.dpreview.com/forums/post/63645424So you don't know how these processors are designed. You try to deduce it from performance analysis. Sure a fun exercise.
We can't see black holes but can deduce their presence and behavior by analyzing how they interact with the matter around them
snapsy wrote
alundeb wrote
snapsy wrote
alundeb wrote
snapsy wrote
I would be very surprised if the DIGIC inside the R5 could sustain 20 fps beyond the initial DRAM frame buffer. Shuffling data off the image sensor into memory @ 20fps is the easy part. Getting 45MP of data subsequently processed through DIGIC and ready to write out to CFE is a different matter. In other words, I would predict the media card wont be the bottleneck for the sustained fps.The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
https://www.dpreview.com/forums/post/63645424
We can't see black holes but can deduce their presence and behavior by analyzing how they interact with the matter around them
Feb 20, 2020 at 11:17 AM
p.10 #12EOS-R5....will it live up to the hype?
arbitrage wrote:
If you don't enjoy sarcasm, please don't watch this and please don't complain about it...you have been warned
This guy is hilarious!!
He's better than watching old Dr. Who re-runs!!! LoL
I hope Canon hires him as their official Canon HAMMER!!! LoL
John
If you don't enjoy sarcasm, please don't watch this and please don't complain about it...you have been warned
This guy is hilarious!!
He's better than watching old Dr. Who re-runs!!! LoL
I hope Canon hires him as their official Canon HAMMER!!! LoL
John
Feb 20, 2020 at 12:30 PM
p.10 #13EOS-R5....will it live up to the hype?
Jesse Evans wrote:
https://www.jpl.nasa.gov/edu/news/2019/4/19/how-scientists-captured-the-first-image-of-a-black-hole/
"Accomplishing what was previously thought to be impossible, a team of international astronomers has captured an image of a black hole’s silhouette."
https://www.jpl.nasa.gov/edu/news/2019/4/19/how-scientists-captured-the-first-image-of-a-black-hole/
"Accomplishing what was previously thought to be impossible, a team of international astronomers has captured an image of a black hole’s silhouette."
Feb 20, 2020 at 12:37 PM
p.10 #14EOS-R5....will it live up to the hype?
snapsy wrote:
alundeb wrote
snapsy wrote
alundeb wrote
snapsy wrote
alundeb wrote
That doesn't sound right to me. If the buffer is before Digic, why do all cameras have unlimited buffer for jpg but very limited buffer for RAW?The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
https://www.dpreview.com/forums/post/63645424
I'm not against that. You just started off as if you knew how it looks inside. I just see a lot of behavior around the black hole that doesn't match what you say.
Feb 20, 2020 at 12:49 PM
p.10 #15EOS-R5....will it live up to the hype?
alundeb wrote:
It still doesn't add up for me. A consequence of what you say, is that moving raw data to the card would only be a small part the total bandwidth requirement.
The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.I recently completed a similar analysis on the Nikon Z bodies, where I found they were Expeed//system bandwidth limited. I haven't compared them to the D850 since I don't own one but IR's tests show the D850 can shoot limitless jpgs whereas the Z's do not. There may be some unique image processing requirements on the Z's that the R5 won't have or for which DIGIC is better optimized for but nonetheless the bandwidth demands for mirrorless are much higher. Here's a link to my analysis on the Z:
https://www.dpreview.com/forums/post/63645424So you don't know how these processors are designed. You try to deduce it from performance analysis. Sure a fun exercise.We can't see black holes but can deduce their presence and behavior by analyzing how they interact with the matter around them I'm not against that. You just started off as if you knew how it looks inside. I just see a lot of behavior around the black hole that doesn't match what you say.
Mistakes are always possible. Likely even, considering we're working off indirect and incomplete information. Which aspect of DIGIC's behavior that I theorized about do you disagree with?
snapsy wrote
alundeb wrote
snapsy wrote
alundeb wrote
snapsy wrote
Few reasons. Mirrorless bodies place more demands on internal bus and memory bandwidth because they're constantly having to move and process image data to support LV and the on-sensor focusing system. Next, the reduced size of the jpg means much lower bandwidth consumption on the shared internal memory when writing out to the media card interface, which means more internal bandwidth available for the multiple trips the intermediate raw data has to move in and out of DIGIC to be processed.The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
https://www.dpreview.com/forums/post/63645424
Mistakes are always possible. Likely even, considering we're working off indirect and incomplete information. Which aspect of DIGIC's behavior that I theorized about do you disagree with?
Feb 20, 2020 at 01:05 PM
p.10 #16EOS-R5....will it live up to the hype?
snapsy wrote:
alundeb wrote
snapsy wrote
alundeb wrote
snapsy wrote
alundeb wrote
It still doesn't add up for me. A consequence of what you say, is that moving raw data to the card would only be a small part the total bandwidth requirement.The other part is that the if the primary buffer really is the buffer, it would not have room for all the raw data. Or the size of the buffer in number of images would be identical for raw and jpeg.
https://www.dpreview.com/forums/post/63645424
In short, two things. That the bottleneck is the rate at which raw data can be processed, and that the primary buffer is the main buffer. The proof is that jpg images can be written at a higher rate than the raw writing bottleneck speed, so they have to be processed at a higher rate. Second that the number of images that will fill the buffer with an extremely slow card, is higher for jpg than raw. What I did not disagree with, is that the processor architecture can be the bottleneck.
Feb 20, 2020 at 01:18 PM
p.10 #17EOS-R5....will it live up to the hype?
ChrisMak wrote:
The EOS-R5 should logically be an allround mirrorless pro body like the 5DIV is an allround pro dslr. There is not yet a reason to release a 1DX like mirrorless AF and speed monster because Canon still needs a couple of years to make the RF lenses for such a body, like a 300mm f2.8, 500mm f4, 400mm f2.8
Because the existing EF lenses adapt very well to the EOS-R, everybody seems to think that Canon will be content to develop a mirrorless top body without the neccessary EF lenses, but that is not like Canon.
Look at the EF lenses...Show more →
EB-1 wrote:
The R5 should be an all around body like the 5D IV, but at 45MP it certainly is not.
EBH
alundeb wrote:
I am at a loss regarding your resolution requirements. There will be three resolution classes from Canon in the new generation RF mount cameras. 20, 45 and 84 MP (roughly). Three resolutions with sensible doubling between. How can 45 in the middle not be an allround body?
There will be an 80MP camera!? That doens't seem to be the current rumor of 45MP.
45MP is too noisy for general purpose IME.
EBH
The EOS-R5 should logically be an allround mirrorless pro body like the 5DIV is an allround pro dslr. There is not yet a reason to release a 1DX like mirrorless AF and speed monster because Canon still needs a couple of years to make the RF lenses for such a body, like a 300mm f2.8, 500mm f4, 400mm f2.8
Because the existing EF lenses adapt very well to the EOS-R, everybody seems to think that Canon will be content to develop a mirrorless top body without the neccessary EF lenses, but that is not like Canon.
Look at the EF lenses...Show more →
EB-1 wrote:
The R5 should be an all around body like the 5D IV, but at 45MP it certainly is not.
EBH
alundeb wrote:
I am at a loss regarding your resolution requirements. There will be three resolution classes from Canon in the new generation RF mount cameras. 20, 45 and 84 MP (roughly). Three resolutions with sensible doubling between. How can 45 in the middle not be an allround body?
There will be an 80MP camera!? That doens't seem to be the current rumor of 45MP.
45MP is too noisy for general purpose IME.
EBH
Feb 20, 2020 at 07:00 PM
p.10 #18EOS-R5....will it live up to the hype?
arbitrage wrote:
If you don't enjoy sarcasm, please don't watch this and please don't complain about it...you have been warned
This guy is the best, but it takes a few watches to get on board.
If you don't enjoy sarcasm, please don't watch this and please don't complain about it...you have been warned
This guy is the best, but it takes a few watches to get on board.
Feb 20, 2020 at 07:47 PM
p.10 #19EOS-R5....will it live up to the hype?
In terms of Canon 'home-runs' there have been a lot from MY POV:
- the Digital Rebel (EOS 300) was the first DLSR I could afford, and at the time the image quality was extremely good. I sold all my Nikon Film gear to afford it
- The 20D - clearly better than anything Nikon had, who where struggling with sensor striping and transitioning from CCD to CMOS (I could be a bit off with the timelines here)
- The 40D - while it was an evolution from the 20D, it was big leap in AF and I loved it.
- The 7D/7D mk2. Extremely rugged and capable.
Feb 20, 2020 at 07:58 PM
p.10 #20EOS-R5....will it live up to the hype?
'This guy is the best'
Based and red-pilled. Lots of spergy material to work with.
Based and red-pilled. Lots of spergy material to work with.
Feb 20, 2020 at 08:12 PM