Affirmitive to the first question, could get to 34p from 25p :D, to the second question i'll have to check on monday, your ML support is amazing Alex, wish I had time to give more regular updates.
On Thursday, May 10, 2012 11:51:25 PM UTC+12, Alex wrote:
> So, with old implementation, you could get 34p from 25p? I don't have > the 60D right now to check.
> If the image is wrong in any mode, simply increase timer A until you > get correct image, and tell me the value.
> On Thu, May 10, 2012 at 2:40 PM, Ash McKenzie <a...@pov.co.nz> wrote: > > Sorry Must have read wrong,
> > When the Camera is set to NTSC 30p ?? if im understanding correctly,
> > there is no frame buffer error in 3x Crop mode, the camera behaviours as > > expected, > > Magic Zoom fine for both, before and after entering 3x Crop Mode, > > Actual Max Frame rate is 30.415 All Modes > > Requesting a Frame Rate Higher than 30p, 35 being the first aviable > option > > results in the frame buffer error (am i discribing that correctly)
> > When the Camera is set to NTSC 24p
> > Same Issues as expressed in PAL 25p Mode
> > Quick question about the time base, if I use NTSC 30p as the time base > for > > over crank will i have to reinterprent my footage on the timeline, one > of > > the things I loved about that feb 18 build is that the footage played > down > > at 25p in camera and appeared in the timeline as 25p as it synced with > smart > > slate TC right off the bat?
> > On Thursday, May 10, 2012 11:19:26 PM UTC+12, Alex wrote:
> >> I've asked you about the 3x crop bug, not about the highest FPS limit.
> >> For highest FPS, I told you to override the 30p mode. You are still in > >> 25p.
> >> On Thu, May 10, 2012 at 2:14 PM, Ash McKenzie <a...@pov.co.nz> wrote: > >> > Hey Alex,
> >> > Since that build date reguardless of the mode, High Jello, Low FPS > 360, > >> > etc, > >> > as well as messing with A and B Timers, all values stop actual fps at > >> > 26.765
> >> > On Thursday, May 10, 2012 10:19:41 PM UTC+12, Alex wrote:
> >> >> Can you post a screenshot with FPS override submenu in that mode? > (to > >> >> see the numbers and the settings).
> >> >> Also, you can increase timer values manually until the image becomes > >> >> correct again, and tell me those values. There is a minimum safe > value > >> >> for timer A and a minimum safe value for timer B (they seem pretty > >> >> much independent).
> >> >> Changing optimization mode may also help. Tell me which modes work > and > >> >> which do not.
> >> >> On Thu, May 10, 2012 at 1:13 PM, Ash McKenzie <a...@pov.co.nz> > wrote: > >> >> > Hope this is useful info,
> >> >> > MagicZoom after 3x Crop mode is now working great,
> >> >> > However 3x Crop mode plus FPS Overide is Now Broken, both VF and > >> >> > Recorded > >> >> > image look like image buffer errors
> >> >> > Thanks Guys
> >> >> > On Thursday, May 10, 2012 6:03:10 AM UTC+12, Alex wrote:
> >> >> >> As you have noticed from the changesets, I've got my hands dirty > and > >> >> >> tried to rework a large part of the backend.
> >> >> >> Most important changes:
> >> >> >> - Memory issue from 550D is gone. > >> >> >> - Shutdown process is much better understood. Hopefully the > >> >> >> incomplete > >> >> >> shutdown problem is solved. > >> >> >> - Menu backend is much less obtrusive - it no longer uses Canon > >> >> >> dialog > >> >> >> API, but the same mechanism used for shortcut keys. Should feel a > >> >> >> bit > >> >> >> faster too. > >> >> >> - A few aggressive stability tests - which uncovered some very > >> >> >> subtle > >> >> >> bugs. > >> >> >> - Most CPU-intensive operations were moved to separate tasks, so > >> >> >> they > >> >> >> should no longer slowdown Canon tasks. > >> >> >> - Some interesting info in debug menu: list of tasks started by > >> >> >> Canon > >> >> >> and by ML (see screenshot), free memory in both malloc and system > >> >> >> heaps, CPU usage log.
> >> >> >> Sure, all of these probably have little effect on user > experience, > >> >> >> but > >> >> >> I think they represent a huge step regarding stability => we are > >> >> >> much > >> >> >> closer to the next big release.
> >> >> >> However, these backend changes may have caused small bugs on the > >> >> >> front-end side. I'll let you discover them. In particular, this > >> >> >> update > >> >> >> may save crash logs when something goes wrong - if you notice > them > >> >> >> on > >> >> >> the card, send them to me. Don't be scared if you see 100 crash > logs > >> >> >> on the card - it's not that bad :)
> >> >> >> Back on the user side. I've implemented a feature requested by > our > >> >> >> moderator from Vimeo (Lucas Pfaff). You've guessed it - focus > >> >> >> ramping > >> >> >> (like this: https://vimeo.com/41443896 ). Besides this, bulb > ramping > >> >> >> got some major improvements:
> >> >> >> - Manual exposure ramping. > >> >> >> - Dedicated sunset/sunrise modes. Idea from Tom Lowe - > >> >> >> http://bit.ly/lowe-ratchet . > >> >> >> - New feedback loop - not just something done with trial and > error, > >> >> >> but this time it's mathematically correct. Proof included: > >> >> >> http://bit.ly/bramp-math . > >> >> >> - Calibration is only done once, and then saved in config file. > >> >> >> Please > >> >> >> try it on 600D and let me know if it works. > >> >> >> - Configurable smoothing factor (for engineers: location of > >> >> >> closed-loop poles). Should result in very little flicker. Not > tested > >> >> >> (only simulated the mathematical model).
> >> >> >> Also you can find a couple of nice tricks:
> >> >> >> - Scrollwheel(s) working in ML menu on all cameras. > >> >> >> - Some LiveView zoom tricks for manual focus (Tweaks menu). > >> >> >> - An automatic HDR bracketing option (just point and shoot - ML > will > >> >> >> compute how many frames you need). > >> >> >> - 5D2, 50D, 500D: an option to adjust black level in movie mode. > Use > >> >> >> it if you get green/magenta shadows (I've noticed them on 5D2). > >> >> >> Newer > >> >> >> cameras don't seem to need this adjustment. > >> >> >> - Some new menu icons - should be more obvious where you have > >> >> >> submenus.
> >> >> >> Fixes:
> >> >> >> - Exposure override should be saved correctly now. > >> >> >> - On 600D, focus peaking and magic zoom should work when toggling > >> >> >> between 1x/3x (credits: Paul Nolan). > >> >> >> - Tried to fix the first second of very loud audio. Not tested. > >> >> >> - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> >> >> >> UniWB was disabled (not reliable).
> >> >> >> Latest changes only tested on my cameras (5D2, 550D, 60D). I'd > like > >> >> >> to > >> >> >> thank to people who compiled from source and pointed out bugs > that I > >> >> >> didn't even think about (like repeating keys in menu).
> >> >> >> Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port > >> >> >> goes > >> >> >> forward, slow but steady. Also take a look at Coutts' progress on > 5D > >> >> >> classic. However, for now, the stable release for 5D2 has the > >> >> >> highest > >> >> >> priority.
> >> >> > To post to this group, send email to ml-devel@googlegroups.com > >> >> > To unsubscribe from this group, send email to > >> >> > ml-devel+unsubscribe@googlegroups.com > >> >> > For more options, visit this group at > >> >> > http://groups.google.com/group/ml-devel?hl=en
> >> > To post to this group, send email to ml-devel@googlegroups.com > >> > To unsubscribe from this group, send email to > >> > ml-devel+unsubscribe@googlegroups.com > >> > For more options, visit this group at > >> > http://groups.google.com/group/ml-devel?hl=en
> > To post to this group, send email to ml-devel@googlegroups.com > > To unsubscribe from this group, send email to > > ml-devel+unsubscribe@googlegroups.com > > For more options, visit this group at > > http://groups.google.com/group/ml-devel?hl=en
On Thu, May 10, 2012 at 4:03 AM, Alex <broscutama...@gmail.com> wrote:
> As you have noticed from the changesets, I've got my hands dirty and
> tried to rework a large part of the backend.
> Most important changes:
> - Memory issue from 550D is gone.
> - Shutdown process is much better understood. Hopefully the incomplete
> shutdown problem is solved.
> - Menu backend is much less obtrusive - it no longer uses Canon dialog
> API, but the same mechanism used for shortcut keys. Should feel a bit
> faster too.
> - A few aggressive stability tests - which uncovered some very subtle bugs.
> - Most CPU-intensive operations were moved to separate tasks, so they
> should no longer slowdown Canon tasks.
> - Some interesting info in debug menu: list of tasks started by Canon
> and by ML (see screenshot), free memory in both malloc and system
> heaps, CPU usage log.
> Sure, all of these probably have little effect on user experience, but
> I think they represent a huge step regarding stability => we are much
> closer to the next big release.
> However, these backend changes may have caused small bugs on the
> front-end side. I'll let you discover them. In particular, this update
> may save crash logs when something goes wrong - if you notice them on
> the card, send them to me. Don't be scared if you see 100 crash logs
> on the card - it's not that bad :)
> Back on the user side. I've implemented a feature requested by our
> moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping
> (like this: https://vimeo.com/41443896 ). Besides this, bulb ramping
> got some major improvements:
> - Manual exposure ramping.
> - Dedicated sunset/sunrise modes. Idea from Tom Lowe -
> http://bit.ly/lowe-ratchet .
> - New feedback loop - not just something done with trial and error,
> but this time it's mathematically correct. Proof included:
> http://bit.ly/bramp-math .
> - Calibration is only done once, and then saved in config file. Please
> try it on 600D and let me know if it works.
> - Configurable smoothing factor (for engineers: location of
> closed-loop poles). Should result in very little flicker. Not tested
> (only simulated the mathematical model).
> Also you can find a couple of nice tricks:
> - Scrollwheel(s) working in ML menu on all cameras.
> - Some LiveView zoom tricks for manual focus (Tweaks menu).
> - An automatic HDR bracketing option (just point and shoot - ML will
> compute how many frames you need).
> - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use
> it if you get green/magenta shadows (I've noticed them on 5D2). Newer
> cameras don't seem to need this adjustment.
> - Some new menu icons - should be more obvious where you have submenus.
> Fixes:
> - Exposure override should be saved correctly now.
> - On 600D, focus peaking and magic zoom should work when toggling
> between 1x/3x (credits: Paul Nolan).
> - Tried to fix the first second of very loud audio. Not tested.
> - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> UniWB was disabled (not reliable).
> Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to
> thank to people who compiled from source and pointed out bugs that I
> didn't even think about (like repeating keys in menu).
> Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes
> forward, slow but steady. Also take a look at Coutts' progress on 5D
> classic. However, for now, the stable release for 5D2 has the highest
> priority.
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to
> ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/ml-devel?hl=en
> As you have noticed from the changesets, I've got my hands dirty and
> tried to rework a large part of the backend.
> Most important changes:
> - Memory issue from 550D is gone.
> - Shutdown process is much better understood. Hopefully the incomplete
> shutdown problem is solved.
> - Menu backend is much less obtrusive - it no longer uses Canon dialog
> API, but the same mechanism used for shortcut keys. Should feel a bit
> faster too.
> - A few aggressive stability tests - which uncovered some very subtle bugs.
> - Most CPU-intensive operations were moved to separate tasks, so they
> should no longer slowdown Canon tasks.
> - Some interesting info in debug menu: list of tasks started by Canon
> and by ML (see screenshot), free memory in both malloc and system
> heaps, CPU usage log.
> Sure, all of these probably have little effect on user experience, but
> I think they represent a huge step regarding stability => we are much
> closer to the next big release.
> However, these backend changes may have caused small bugs on the
> front-end side. I'll let you discover them. In particular, this update
> may save crash logs when something goes wrong - if you notice them on
> the card, send them to me. Don't be scared if you see 100 crash logs
> on the card - it's not that bad :)
> Back on the user side. I've implemented a feature requested by our
> moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping
> (like this: https://vimeo.com/41443896 ). Besides this, bulb ramping
> got some major improvements:
> - Manual exposure ramping.
> - Dedicated sunset/sunrise modes. Idea from Tom Lowe -
> http://bit.ly/lowe-ratchet .
> - New feedback loop - not just something done with trial and error,
> but this time it's mathematically correct. Proof included:
> http://bit.ly/bramp-math .
> - Calibration is only done once, and then saved in config file. Please
> try it on 600D and let me know if it works.
> - Configurable smoothing factor (for engineers: location of
> closed-loop poles). Should result in very little flicker. Not tested
> (only simulated the mathematical model).
> Also you can find a couple of nice tricks:
> - Scrollwheel(s) working in ML menu on all cameras.
> - Some LiveView zoom tricks for manual focus (Tweaks menu).
> - An automatic HDR bracketing option (just point and shoot - ML will
> compute how many frames you need).
> - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use
> it if you get green/magenta shadows (I've noticed them on 5D2). Newer
> cameras don't seem to need this adjustment.
> - Some new menu icons - should be more obvious where you have submenus.
> Fixes:
> - Exposure override should be saved correctly now.
> - On 600D, focus peaking and magic zoom should work when toggling
> between 1x/3x (credits: Paul Nolan).
> - Tried to fix the first second of very loud audio. Not tested.
> - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> UniWB was disabled (not reliable).
> Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to
> thank to people who compiled from source and pointed out bugs that I
> didn't even think about (like repeating keys in menu).
> Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes
> forward, slow but steady. Also take a look at Coutts' progress on 5D
> classic. However, for now, the stable release for 5D2 has the highest
> priority.
<nicol...@gmail.com> wrote:
> Thanks for the update!
> Crash logs attached from testing the sunset bulb ramping. (Still managed to
> work OK besides some crazy flicker)
> Cheers,
> Nick.
> On Thu, May 10, 2012 at 4:03 AM, Alex <broscutama...@gmail.com> wrote:
>> As you have noticed from the changesets, I've got my hands dirty and
>> tried to rework a large part of the backend.
>> Most important changes:
>> - Memory issue from 550D is gone.
>> - Shutdown process is much better understood. Hopefully the incomplete
>> shutdown problem is solved.
>> - Menu backend is much less obtrusive - it no longer uses Canon dialog
>> API, but the same mechanism used for shortcut keys. Should feel a bit
>> faster too.
>> - A few aggressive stability tests - which uncovered some very subtle
>> bugs.
>> - Most CPU-intensive operations were moved to separate tasks, so they
>> should no longer slowdown Canon tasks.
>> - Some interesting info in debug menu: list of tasks started by Canon
>> and by ML (see screenshot), free memory in both malloc and system
>> heaps, CPU usage log.
>> Sure, all of these probably have little effect on user experience, but
>> I think they represent a huge step regarding stability => we are much
>> closer to the next big release.
>> However, these backend changes may have caused small bugs on the
>> front-end side. I'll let you discover them. In particular, this update
>> may save crash logs when something goes wrong - if you notice them on
>> the card, send them to me. Don't be scared if you see 100 crash logs
>> on the card - it's not that bad :)
>> Back on the user side. I've implemented a feature requested by our
>> moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping
>> (like this: https://vimeo.com/41443896 ). Besides this, bulb ramping
>> got some major improvements:
>> - Manual exposure ramping.
>> - Dedicated sunset/sunrise modes. Idea from Tom Lowe -
>> http://bit.ly/lowe-ratchet .
>> - New feedback loop - not just something done with trial and error,
>> but this time it's mathematically correct. Proof included:
>> http://bit.ly/bramp-math .
>> - Calibration is only done once, and then saved in config file. Please
>> try it on 600D and let me know if it works.
>> - Configurable smoothing factor (for engineers: location of
>> closed-loop poles). Should result in very little flicker. Not tested
>> (only simulated the mathematical model).
>> Also you can find a couple of nice tricks:
>> - Scrollwheel(s) working in ML menu on all cameras.
>> - Some LiveView zoom tricks for manual focus (Tweaks menu).
>> - An automatic HDR bracketing option (just point and shoot - ML will
>> compute how many frames you need).
>> - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use
>> it if you get green/magenta shadows (I've noticed them on 5D2). Newer
>> cameras don't seem to need this adjustment.
>> - Some new menu icons - should be more obvious where you have submenus.
>> Fixes:
>> - Exposure override should be saved correctly now.
>> - On 600D, focus peaking and magic zoom should work when toggling
>> between 1x/3x (credits: Paul Nolan).
>> - Tried to fix the first second of very loud audio. Not tested.
>> - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
>> UniWB was disabled (not reliable).
>> Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to
>> thank to people who compiled from source and pointed out bugs that I
>> didn't even think about (like repeating keys in menu).
>> Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes
>> forward, slow but steady. Also take a look at Coutts' progress on 5D
>> classic. However, for now, the stable release for 5D2 has the highest
>> priority.
>> To post to this group, send email to ml-devel@googlegroups.com
>> To unsubscribe from this group, send email to
>> ml-devel+unsubscribe@googlegroups.com
>> For more options, visit this group at
>> http://groups.google.com/group/ml-devel?hl=en
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to
> ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/ml-devel?hl=en
> If it wasn't the 1/8 EV flicker (which you can't avoid), can you send
> me the exif info for the pictures, and the settings you have used?
> On Thu, May 10, 2012 at 3:00 PM, Nicolas Triantafillou
> <nicol...@gmail.com> wrote:
> > Thanks for the update!
> > Crash logs attached from testing the sunset bulb ramping. (Still managed
> to
> > work OK besides some crazy flicker)
> > Cheers,
> > Nick.
> > On Thu, May 10, 2012 at 4:03 AM, Alex <broscutama...@gmail.com> wrote:
> >> As you have noticed from the changesets, I've got my hands dirty and
> >> tried to rework a large part of the backend.
> >> Most important changes:
> >> - Memory issue from 550D is gone.
> >> - Shutdown process is much better understood. Hopefully the incomplete
> >> shutdown problem is solved.
> >> - Menu backend is much less obtrusive - it no longer uses Canon dialog
> >> API, but the same mechanism used for shortcut keys. Should feel a bit
> >> faster too.
> >> - A few aggressive stability tests - which uncovered some very subtle
> >> bugs.
> >> - Most CPU-intensive operations were moved to separate tasks, so they
> >> should no longer slowdown Canon tasks.
> >> - Some interesting info in debug menu: list of tasks started by Canon
> >> and by ML (see screenshot), free memory in both malloc and system
> >> heaps, CPU usage log.
> >> Sure, all of these probably have little effect on user experience, but
> >> I think they represent a huge step regarding stability => we are much
> >> closer to the next big release.
> >> However, these backend changes may have caused small bugs on the
> >> front-end side. I'll let you discover them. In particular, this update
> >> may save crash logs when something goes wrong - if you notice them on
> >> the card, send them to me. Don't be scared if you see 100 crash logs
> >> on the card - it's not that bad :)
> >> Back on the user side. I've implemented a feature requested by our
> >> moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping
> >> (like this: https://vimeo.com/41443896 ). Besides this, bulb ramping
> >> got some major improvements:
> >> - Manual exposure ramping.
> >> - Dedicated sunset/sunrise modes. Idea from Tom Lowe -
> >> http://bit.ly/lowe-ratchet .
> >> - New feedback loop - not just something done with trial and error,
> >> but this time it's mathematically correct. Proof included:
> >> http://bit.ly/bramp-math .
> >> - Calibration is only done once, and then saved in config file. Please
> >> try it on 600D and let me know if it works.
> >> - Configurable smoothing factor (for engineers: location of
> >> closed-loop poles). Should result in very little flicker. Not tested
> >> (only simulated the mathematical model).
> >> Also you can find a couple of nice tricks:
> >> - Scrollwheel(s) working in ML menu on all cameras.
> >> - Some LiveView zoom tricks for manual focus (Tweaks menu).
> >> - An automatic HDR bracketing option (just point and shoot - ML will
> >> compute how many frames you need).
> >> - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use
> >> it if you get green/magenta shadows (I've noticed them on 5D2). Newer
> >> cameras don't seem to need this adjustment.
> >> - Some new menu icons - should be more obvious where you have submenus.
> >> Fixes:
> >> - Exposure override should be saved correctly now.
> >> - On 600D, focus peaking and magic zoom should work when toggling
> >> between 1x/3x (credits: Paul Nolan).
> >> - Tried to fix the first second of very loud audio. Not tested.
> >> - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> >> UniWB was disabled (not reliable).
> >> Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to
> >> thank to people who compiled from source and pointed out bugs that I
> >> didn't even think about (like repeating keys in menu).
> >> Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes
> >> forward, slow but steady. Also take a look at Coutts' progress on 5D
> >> classic. However, for now, the stable release for 5D2 has the highest
> >> priority.
> >> To post to this group, send email to ml-devel@googlegroups.com
> >> To unsubscribe from this group, send email to
> >> ml-devel+unsubscribe@googlegroups.com
> >> For more options, visit this group at
> >> http://groups.google.com/group/ml-devel?hl=en
> > To post to this group, send email to ml-devel@googlegroups.com
> > To unsubscribe from this group, send email to
> > ml-devel+unsubscribe@googlegroups.com
> > For more options, visit this group at
> > http://groups.google.com/group/ml-devel?hl=en
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to
> ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/ml-devel?hl=en
I don´t know if i can understand it right.. so this means,that people
with 5D2 ML unified can replace the old autoexec.bin with this new one
too?
would it be possible to skip every second frame(only on the screen) in
HDR video mode..to avoid the flickering on display?
BTW: Thanks for your briliant work!!
Greetings from Prague
On May 9, 8:03 pm, Alex <broscutama...@gmail.com> wrote:
> As you have noticed from the changesets, I've got my hands dirty and
> tried to rework a large part of the backend.
> Most important changes:
> - Memory issue from 550D is gone.
> - Shutdown process is much better understood. Hopefully the incomplete
> shutdown problem is solved.
> - Menu backend is much less obtrusive - it no longer uses Canon dialog
> API, but the same mechanism used for shortcut keys. Should feel a bit
> faster too.
> - A few aggressive stability tests - which uncovered some very subtle bugs.
> - Most CPU-intensive operations were moved to separate tasks, so they
> should no longer slowdown Canon tasks.
> - Some interesting info in debug menu: list of tasks started by Canon
> and by ML (see screenshot), free memory in both malloc and system
> heaps, CPU usage log.
> Sure, all of these probably have little effect on user experience, but
> I think they represent a huge step regarding stability => we are much
> closer to the next big release.
> However, these backend changes may have caused small bugs on the
> front-end side. I'll let you discover them. In particular, this update
> may save crash logs when something goes wrong - if you notice them on
> the card, send them to me. Don't be scared if you see 100 crash logs
> on the card - it's not that bad :)
> Back on the user side. I've implemented a feature requested by our
> moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping
> (like this:https://vimeo.com/41443896). Besides this, bulb ramping
> got some major improvements:
> - Manual exposure ramping.
> - Dedicated sunset/sunrise modes. Idea from Tom Lowe -http://bit.ly/lowe-ratchet.
> - New feedback loop - not just something done with trial and error,
> but this time it's mathematically correct. Proof included:http://bit.ly/bramp-math.
> - Calibration is only done once, and then saved in config file. Please
> try it on 600D and let me know if it works.
> - Configurable smoothing factor (for engineers: location of
> closed-loop poles). Should result in very little flicker. Not tested
> (only simulated the mathematical model).
> Also you can find a couple of nice tricks:
> - Scrollwheel(s) working in ML menu on all cameras.
> - Some LiveView zoom tricks for manual focus (Tweaks menu).
> - An automatic HDR bracketing option (just point and shoot - ML will
> compute how many frames you need).
> - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use
> it if you get green/magenta shadows (I've noticed them on 5D2). Newer
> cameras don't seem to need this adjustment.
> - Some new menu icons - should be more obvious where you have submenus.
> Fixes:
> - Exposure override should be saved correctly now.
> - On 600D, focus peaking and magic zoom should work when toggling
> between 1x/3x (credits: Paul Nolan).
> - Tried to fix the first second of very loud audio. Not tested.
> - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> UniWB was disabled (not reliable).
> Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to
> thank to people who compiled from source and pointed out bugs that I
> didn't even think about (like repeating keys in menu).
> Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes
> forward, slow but steady. Also take a look at Coutts' progress on 5D
> classic. However, for now, the stable release for 5D2 has the highest
> priority.
> As you have noticed from the changesets, I've got my hands dirty and
> tried to rework a large part of the backend.
> Most important changes:
> - Memory issue from 550D is gone.
> - Shutdown process is much better understood. Hopefully the incomplete
> shutdown problem is solved.
> - Menu backend is much less obtrusive - it no longer uses Canon dialog
> API, but the same mechanism used for shortcut keys. Should feel a bit
> faster too.
> - A few aggressive stability tests - which uncovered some very subtle bugs.
> - Most CPU-intensive operations were moved to separate tasks, so they
> should no longer slowdown Canon tasks.
> - Some interesting info in debug menu: list of tasks started by Canon
> and by ML (see screenshot), free memory in both malloc and system
> heaps, CPU usage log.
> Sure, all of these probably have little effect on user experience, but
> I think they represent a huge step regarding stability => we are much
> closer to the next big release.
> However, these backend changes may have caused small bugs on the
> front-end side. I'll let you discover them. In particular, this update
> may save crash logs when something goes wrong - if you notice them on
> the card, send them to me. Don't be scared if you see 100 crash logs
> on the card - it's not that bad :)
> Back on the user side. I've implemented a feature requested by our
> moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping
> (like this:https://vimeo.com/41443896). Besides this, bulb ramping
> got some major improvements:
> - Manual exposure ramping.
> - Dedicated sunset/sunrise modes. Idea from Tom Lowe -http://bit.ly/lowe-ratchet.
> - New feedback loop - not just something done with trial and error,
> but this time it's mathematically correct. Proof included:http://bit.ly/bramp-math.
> - Calibration is only done once, and then saved in config file. Please
> try it on 600D and let me know if it works.
> - Configurable smoothing factor (for engineers: location of
> closed-loop poles). Should result in very little flicker. Not tested
> (only simulated the mathematical model).
> Also you can find a couple of nice tricks:
> - Scrollwheel(s) working in ML menu on all cameras.
> - Some LiveView zoom tricks for manual focus (Tweaks menu).
> - An automatic HDR bracketing option (just point and shoot - ML will
> compute how many frames you need).
> - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use
> it if you get green/magenta shadows (I've noticed them on 5D2). Newer
> cameras don't seem to need this adjustment.
> - Some new menu icons - should be more obvious where you have submenus.
> Fixes:
> - Exposure override should be saved correctly now.
> - On 600D, focus peaking and magic zoom should work when toggling
> between 1x/3x (credits: Paul Nolan).
> - Tried to fix the first second of very loud audio. Not tested.
> - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> UniWB was disabled (not reliable).
> Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to
> thank to people who compiled from source and pointed out bugs that I
> didn't even think about (like repeating keys in menu).
> Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes
> forward, slow but steady. Also take a look at Coutts' progress on 5D
> classic. However, for now, the stable release for 5D2 has the highest
> priority.
On Friday, May 11, 2012 6:14:59 AM UTC-4, Milan Drobný wrote:
> I don´t know if i can understand it right.. so this means,that people > with 5D2 ML unified can replace the old autoexec.bin with this new one > too? > would it be possible to skip every second frame(only on the screen) in > HDR video mode..to avoid the flickering on display?
> BTW: Thanks for your briliant work!! > Greetings from Prague
> On May 9, 8:03 pm, Alex <broscutama...@gmail.com> wrote: > > As you have noticed from the changesets, I've got my hands dirty and > > tried to rework a large part of the backend.
> > Most important changes:
> > - Memory issue from 550D is gone. > > - Shutdown process is much better understood. Hopefully the incomplete > > shutdown problem is solved. > > - Menu backend is much less obtrusive - it no longer uses Canon dialog > > API, but the same mechanism used for shortcut keys. Should feel a bit > > faster too. > > - A few aggressive stability tests - which uncovered some very subtle > bugs. > > - Most CPU-intensive operations were moved to separate tasks, so they > > should no longer slowdown Canon tasks. > > - Some interesting info in debug menu: list of tasks started by Canon > > and by ML (see screenshot), free memory in both malloc and system > > heaps, CPU usage log.
> > Sure, all of these probably have little effect on user experience, but > > I think they represent a huge step regarding stability => we are much > > closer to the next big release.
> > However, these backend changes may have caused small bugs on the > > front-end side. I'll let you discover them. In particular, this update > > may save crash logs when something goes wrong - if you notice them on > > the card, send them to me. Don't be scared if you see 100 crash logs > > on the card - it's not that bad :)
> > Back on the user side. I've implemented a feature requested by our > > moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping > > (like this:https://vimeo.com/41443896). Besides this, bulb ramping > > got some major improvements:
> > - Manual exposure ramping. > > - Dedicated sunset/sunrise modes. Idea from Tom Lowe -
> http://bit.ly/lowe-ratchet. > > - New feedback loop - not just something done with trial and error, > > but this time it's mathematically correct. Proof included:
> http://bit.ly/bramp-math. > > - Calibration is only done once, and then saved in config file. Please > > try it on 600D and let me know if it works. > > - Configurable smoothing factor (for engineers: location of > > closed-loop poles). Should result in very little flicker. Not tested > > (only simulated the mathematical model).
> > Also you can find a couple of nice tricks:
> > - Scrollwheel(s) working in ML menu on all cameras. > > - Some LiveView zoom tricks for manual focus (Tweaks menu). > > - An automatic HDR bracketing option (just point and shoot - ML will > > compute how many frames you need). > > - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use > > it if you get green/magenta shadows (I've noticed them on 5D2). Newer > > cameras don't seem to need this adjustment. > > - Some new menu icons - should be more obvious where you have submenus.
> > Fixes:
> > - Exposure override should be saved correctly now. > > - On 600D, focus peaking and magic zoom should work when toggling > > between 1x/3x (credits: Paul Nolan). > > - Tried to fix the first second of very loud audio. Not tested. > > - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> > UniWB was disabled (not reliable).
> > Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to > > thank to people who compiled from source and pointed out bugs that I > > didn't even think about (like repeating keys in menu).
> > Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes > > forward, slow but steady. Also take a look at Coutts' progress on 5D > > classic. However, for now, the stable release for 5D2 has the highest > > priority.
Under the shoot menu. Focus ramping only works in Liveview and it will shift focus a preset amount with each exposure. Bulb ramping is more complicated and you should read the calibration instructions on the wikia here. <http://magiclantern.wikia.com/wiki/Unified/UserGuide#id3>
On Friday, May 11, 2012 6:17:14 AM UTC-4, EmIls Geršinskis Geršinskis wrote:
> How to set Bulb/Focus Ramping ?
> On May 9, 9:03 pm, Alex <broscutama...@gmail.com> wrote: > > As you have noticed from the changesets, I've got my hands dirty and > > tried to rework a large part of the backend.
> > Most important changes:
> > - Memory issue from 550D is gone. > > - Shutdown process is much better understood. Hopefully the incomplete > > shutdown problem is solved. > > - Menu backend is much less obtrusive - it no longer uses Canon dialog > > API, but the same mechanism used for shortcut keys. Should feel a bit > > faster too. > > - A few aggressive stability tests - which uncovered some very subtle > bugs. > > - Most CPU-intensive operations were moved to separate tasks, so they > > should no longer slowdown Canon tasks. > > - Some interesting info in debug menu: list of tasks started by Canon > > and by ML (see screenshot), free memory in both malloc and system > > heaps, CPU usage log.
> > Sure, all of these probably have little effect on user experience, but > > I think they represent a huge step regarding stability => we are much > > closer to the next big release.
> > However, these backend changes may have caused small bugs on the > > front-end side. I'll let you discover them. In particular, this update > > may save crash logs when something goes wrong - if you notice them on > > the card, send them to me. Don't be scared if you see 100 crash logs > > on the card - it's not that bad :)
> > Back on the user side. I've implemented a feature requested by our > > moderator from Vimeo (Lucas Pfaff). You've guessed it - focus ramping > > (like this:https://vimeo.com/41443896). Besides this, bulb ramping > > got some major improvements:
> > - Manual exposure ramping. > > - Dedicated sunset/sunrise modes. Idea from Tom Lowe -
> http://bit.ly/lowe-ratchet. > > - New feedback loop - not just something done with trial and error, > > but this time it's mathematically correct. Proof included:
> http://bit.ly/bramp-math. > > - Calibration is only done once, and then saved in config file. Please > > try it on 600D and let me know if it works. > > - Configurable smoothing factor (for engineers: location of > > closed-loop poles). Should result in very little flicker. Not tested > > (only simulated the mathematical model).
> > Also you can find a couple of nice tricks:
> > - Scrollwheel(s) working in ML menu on all cameras. > > - Some LiveView zoom tricks for manual focus (Tweaks menu). > > - An automatic HDR bracketing option (just point and shoot - ML will > > compute how many frames you need). > > - 5D2, 50D, 500D: an option to adjust black level in movie mode. Use > > it if you get green/magenta shadows (I've noticed them on 5D2). Newer > > cameras don't seem to need this adjustment. > > - Some new menu icons - should be more obvious where you have submenus.
> > Fixes:
> > - Exposure override should be saved correctly now. > > - On 600D, focus peaking and magic zoom should work when toggling > > between 1x/3x (credits: Paul Nolan). > > - Tried to fix the first second of very loud audio. Not tested. > > - Tried to fix very low FPS issue 979/1228 on 500D. Not tested.
> > UniWB was disabled (not reliable).
> > Latest changes only tested on my cameras (5D2, 550D, 60D). I'd like to > > thank to people who compiled from source and pointed out bugs that I > > didn't even think about (like repeating keys in menu).
> > Meanwhile, with the help of Indy, g3gg0 and Marvin, the 5D3 port goes > > forward, slow but steady. Also take a look at Coutts' progress on 5D > > classic. However, for now, the stable release for 5D2 has the highest > > priority.
On Fri, May 11, 2012 at 7:24 PM, Stifa <ste...@gmail.com> wrote:
> Why is HDR bracketing making a pause after each bracketed shot ? It is
> annoying, in wasnt like that before . right ?
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/ml-devel?hl=en
Thnx 4 a quick reply,
On my 550D , it goes into 2 second beep and then takes first picture
and then displays it like on the PLAY mode, then after ~2 seconds it
takes another shot and like that in circle... It was like that in
previous update.
On Fri, May 11, 2012 at 7:30 PM, Stifa <ste...@gmail.com> wrote:
> Thnx 4 a quick reply,
> On my 550D , it goes into 2 second beep and then takes first picture
> and then displays it like on the PLAY mode, then after ~2 seconds it
> takes another shot and like that in circle... It was like that in
> previous update.
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/ml-devel?hl=en
I have found a small bug. I have set the camera to view the zoom when I change manual focus. After I switch on the 60D and went to movie mode (ISO + LV), I was shown the 5x zoom immediately. After I moved the focus manually on the camera and the 5x zoom disappeared, there was some error message with one of the built-in cropmarks (if I remember well it was Cinescope) and got Err 70. I tried the above again and the same happened (which is why there are 2 crash logs attached). This only happens when you change focus during the 5x zoom which appears when you enter movie mode. Changing focus after it disappears works perfectly.
Hope this helps. Other than that, great work! :-) Malcolm
It's actually worse on the 5D. If I enable the 'zoom on manual focus' feature, it works after enabling it. When you switch off and on live view (either double pressing the LV button, or restarting the camera), you get a black screen with the 5x text only, and it's stuck like that even if you switch off the camera. There is no Err on the top screen, and no crash log is produced. After turning it off (accessed the ML menu while not in live view and switched the feature off), the camera works normally.
On Friday, May 11, 2012 9:58:22 PM UTC+2, Malcolm Debono wrote:
> Hi Alex,
> I have found a small bug. I have set the camera to view the zoom when I > change manual focus. After I switch on the 60D and went to movie mode (ISO > + LV), I was shown the 5x zoom immediately. After I moved the focus > manually on the camera and the 5x zoom disappeared, there was some error > message with one of the built-in cropmarks (if I remember well it was > Cinescope) and got Err 70. I tried the above again and the same happened > (which is why there are 2 crash logs attached). This only happens when you > change focus during the 5x zoom which appears when you enter movie mode. > Changing focus after it disappears works perfectly.
> Hope this helps. Other than that, great work! :-) > Malcolm
Looks like the logs are very good for catching hidden bugs. So... time
for the fixes :)
- ML compiles without warnings (credits: Takashi Miyake).
- Logs are now saved silently, without prompting (since they are not
really crashes, just warnings). Still, if you find them on the card,
please send them.
- Startup with 64GB cards should no longer cause trouble. Also,
startup with smaller cards should be faster.
- Zoom on manual focus: I couldn't reproduce any of the two bugs, but
I've moved it to a different task. Please re-test. On 60D, make sure
the cropmarks are compressed.
For FPS on 60D/600D: I've tried to combine the two FPS methods, so now
you can get all the FPS features you had in the old implementation,
plus the new ones. This means:
- timelapse in low light (down to roughly 0.2 fps at 1/fps),
- timelapse in daylight (fast exposures at around 3 fps),
- slight overcranking (34fps),
- exact FPS, slitscan (high jello) mode, FPS fine tuning and so on.
On these two cameras, when you optimize for low light and you
undercrank, ML will use the DIGIC method (which increases shutter
speed at low frame rates). Otherwise, ML will use the
60D/600D-specific method (which does not have the side effect of
increased shutter speed).
FPS needs some thorough testing especially on 600D; I'm not really
sure it's rock solid. In particular, it may have to flip video modes
to apply the changes (they can no longer be applied while recording).
Minor stuff:
- Q button fixed on 600D (confirmed by Scrax).
- on 60D/600D, since Q button does not work while recording, you can
use the INFO button instead (but normally you shouldn't need this).
- LED feedback for bulb timer (for those who use it with main display off).
Warning: I didn't run any real-world test with these changes. Only
tried a few random things and ran the automatic tests. I'm sure there
are still rough edges - they need to be reported so I can fix them.
> Looks like the logs are very good for catching hidden bugs. So... time
> for the fixes :)
> - ML compiles without warnings (credits: Takashi Miyake).
> - Logs are now saved silently, without prompting (since they are not
> really crashes, just warnings). Still, if you find them on the card,
> please send them.
> - Startup with 64GB cards should no longer cause trouble. Also,
> startup with smaller cards should be faster.
> - Zoom on manual focus: I couldn't reproduce any of the two bugs, but
> I've moved it to a different task. Please re-test. On 60D, make sure
> the cropmarks are compressed.
> For FPS on 60D/600D: I've tried to combine the two FPS methods, so now
> you can get all the FPS features you had in the old implementation,
> plus the new ones. This means:
> - timelapse in low light (down to roughly 0.2 fps at 1/fps),
> - timelapse in daylight (fast exposures at around 3 fps),
> - slight overcranking (34fps),
> - exact FPS, slitscan (high jello) mode, FPS fine tuning and so on.
> On these two cameras, when you optimize for low light and you
> undercrank, ML will use the DIGIC method (which increases shutter
> speed at low frame rates). Otherwise, ML will use the
> 60D/600D-specific method (which does not have the side effect of
> increased shutter speed).
> FPS needs some thorough testing especially on 600D; I'm not really
> sure it's rock solid. In particular, it may have to flip video modes
> to apply the changes (they can no longer be applied while recording).
> Minor stuff:
> - Q button fixed on 600D (confirmed by Scrax).
> - on 60D/600D, since Q button does not work while recording, you can
> use the INFO button instead (but normally you shouldn't need this).
> - LED feedback for bulb timer (for those who use it with main display off).
> Warning: I didn't run any real-world test with these changes. Only
> tried a few random things and ran the automatic tests. I'm sure there
> are still rough edges - they need to be reported so I can fix them.
> Looks like the logs are very good for catching hidden bugs. So... time > for the fixes :)
> - ML compiles without warnings (credits: Takashi Miyake). > - Logs are now saved silently, without prompting (since they are not > really crashes, just warnings). Still, if you find them on the card, > please send them. > - Startup with 64GB cards should no longer cause trouble. Also, > startup with smaller cards should be faster. > - Zoom on manual focus: I couldn't reproduce any of the two bugs, but > I've moved it to a different task. Please re-test. On 60D, make sure > the cropmarks are compressed.
> For FPS on 60D/600D: I've tried to combine the two FPS methods, so now > you can get all the FPS features you had in the old implementation, > plus the new ones. This means: > - timelapse in low light (down to roughly 0.2 fps at 1/fps), > - timelapse in daylight (fast exposures at around 3 fps), > - slight overcranking (34fps), > - exact FPS, slitscan (high jello) mode, FPS fine tuning and so on.
> On these two cameras, when you optimize for low light and you > undercrank, ML will use the DIGIC method (which increases shutter > speed at low frame rates). Otherwise, ML will use the > 60D/600D-specific method (which does not have the side effect of > increased shutter speed).
> FPS needs some thorough testing especially on 600D; I'm not really > sure it's rock solid. In particular, it may have to flip video modes > to apply the changes (they can no longer be applied while recording).
> Minor stuff: > - Q button fixed on 600D (confirmed by Scrax). > - on 60D/600D, since Q button does not work while recording, you can > use the INFO button instead (but normally you shouldn't need this). > - LED feedback for bulb timer (for those who use it with main display > off).
> Warning: I didn't run any real-world test with these changes. Only > tried a few random things and ran the automatic tests. I'm sure there > are still rough edges - they need to be reported so I can fix them.
Thanks for the updated version! Will try it out in the next few days as I'm busy shooting some projects this weekend. Today I have been filming for around 10 hours, and have been using the previous build (the one found in this thread before the update), and I'm pleased to say that I haven't experienced any troubles at all both on the 60D and the 5D. If that doesn't prove that this build is stable, I don't know what else will! :-)
Sweet, having the 3x zoom working on the 600D is a nice improvement.
When using DISP+Zoom to enable digital zoom, the KEL and ISO shortcuts are displayed as if you'd pressed DISP on it's own.
This requires a second press of DISP to clear the screen.
> Looks like the logs are very good for catching hidden bugs. So... time
> for the fixes :)
> - ML compiles without warnings (credits: Takashi Miyake).
> - Logs are now saved silently, without prompting (since they are not
> really crashes, just warnings). Still, if you find them on the card,
> please send them.
> - Startup with 64GB cards should no longer cause trouble. Also,
> startup with smaller cards should be faster.
> - Zoom on manual focus: I couldn't reproduce any of the two bugs, but
> I've moved it to a different task. Please re-test. On 60D, make sure
> the cropmarks are compressed.
> For FPS on 60D/600D: I've tried to combine the two FPS methods, so now
> you can get all the FPS features you had in the old implementation,
> plus the new ones. This means:
> - timelapse in low light (down to roughly 0.2 fps at 1/fps),
> - timelapse in daylight (fast exposures at around 3 fps),
> - slight overcranking (34fps),
> - exact FPS, slitscan (high jello) mode, FPS fine tuning and so on.
> On these two cameras, when you optimize for low light and you
> undercrank, ML will use the DIGIC method (which increases shutter
> speed at low frame rates). Otherwise, ML will use the
> 60D/600D-specific method (which does not have the side effect of
> increased shutter speed).
> FPS needs some thorough testing especially on 600D; I'm not really
> sure it's rock solid. In particular, it may have to flip video modes
> to apply the changes (they can no longer be applied while recording).
> Minor stuff:
> - Q button fixed on 600D (confirmed by Scrax).
> - on 60D/600D, since Q button does not work while recording, you can
> use the INFO button instead (but normally you shouldn't need this).
> - LED feedback for bulb timer (for those who use it with main display off).
> Warning: I didn't run any real-world test with these changes. Only
> tried a few random things and ran the automatic tests. I'm sure there
> are still rough edges - they need to be reported so I can fix them.
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/ml-devel?hl=en<autoexec.bin>
Noticing an issue with Movie mode on both 550D and 600D, using the wheel to adjust shutter speed.
When moving the wheel in one direction (short exposure to longer exposure), it changes value for each click of the wheel (i.e. 1/64<-1/50<-1/42).
But when moving the wheel in the other direction, the first click has no effect, then the second causes it to jump two values (i.e. 1/42->[no change]->1/64).
So to get from 1/42 to 1/50, you have to click the wheel twice in one direction and once in the other direction (1/42->[no change]->1/64<-1/50).
> I noticed that on 600D the INFO button act like Q always, not only during rec. I can still open the help from the complete user guide menu.
> Il giorno sabato 12 maggio 2012 12:26:37 UTC+2, Alex ha scritto:
> Looks like the logs are very good for catching hidden bugs. So... time > for the fixes :)
> - ML compiles without warnings (credits: Takashi Miyake). > - Logs are now saved silently, without prompting (since they are not > really crashes, just warnings). Still, if you find them on the card, > please send them. > - Startup with 64GB cards should no longer cause trouble. Also, > startup with smaller cards should be faster. > - Zoom on manual focus: I couldn't reproduce any of the two bugs, but > I've moved it to a different task. Please re-test. On 60D, make sure > the cropmarks are compressed.
> For FPS on 60D/600D: I've tried to combine the two FPS methods, so now > you can get all the FPS features you had in the old implementation, > plus the new ones. This means: > - timelapse in low light (down to roughly 0.2 fps at 1/fps), > - timelapse in daylight (fast exposures at around 3 fps), > - slight overcranking (34fps), > - exact FPS, slitscan (high jello) mode, FPS fine tuning and so on.
> On these two cameras, when you optimize for low light and you > undercrank, ML will use the DIGIC method (which increases shutter > speed at low frame rates). Otherwise, ML will use the > 60D/600D-specific method (which does not have the side effect of > increased shutter speed).
> FPS needs some thorough testing especially on 600D; I'm not really > sure it's rock solid. In particular, it may have to flip video modes > to apply the changes (they can no longer be applied while recording).
> Minor stuff: > - Q button fixed on 600D (confirmed by Scrax). > - on 60D/600D, since Q button does not work while recording, you can > use the INFO button instead (but normally you shouldn't need this). > - LED feedback for bulb timer (for those who use it with main display off).
> Warning: I didn't run any real-world test with these changes. Only > tried a few random things and ran the automatic tests. I'm sure there > are still rough edges - they need to be reported so I can fix them.
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/ml-devel?hl=en
> Noticing an issue with Movie mode on both 550D and 600D, using the wheel > to adjust shutter speed. > When moving the wheel in one direction (short exposure to longer > exposure), it changes value for each click of the wheel (i.e. > 1/64<-1/50<-1/42). > But when moving the wheel in the other direction, the first click has no > effect, then the second causes it to jump two values (i.e. 1/42->[no > change]->1/64).
> So to get from 1/42 to 1/50, you have to click the wheel twice in one > direction and once in the other direction (1/42->[no change]->1/64<-1/50).
> Morgan.
> On 13/05/2012, at 4:27 AM, scrax wrote:
> I noticed that on 600D the INFO button act like Q always, not only during > rec. I can still open the help from the complete user guide menu.
> Il giorno sabato 12 maggio 2012 12:26:37 UTC+2, Alex ha scritto:
>> Looks like the logs are very good for catching hidden bugs. So... time >> for the fixes :)
>> - ML compiles without warnings (credits: Takashi Miyake). >> - Logs are now saved silently, without prompting (since they are not >> really crashes, just warnings). Still, if you find them on the card, >> please send them. >> - Startup with 64GB cards should no longer cause trouble. Also, >> startup with smaller cards should be faster. >> - Zoom on manual focus: I couldn't reproduce any of the two bugs, but >> I've moved it to a different task. Please re-test. On 60D, make sure >> the cropmarks are compressed.
>> For FPS on 60D/600D: I've tried to combine the two FPS methods, so now >> you can get all the FPS features you had in the old implementation, >> plus the new ones. This means: >> - timelapse in low light (down to roughly 0.2 fps at 1/fps), >> - timelapse in daylight (fast exposures at around 3 fps), >> - slight overcranking (34fps), >> - exact FPS, slitscan (high jello) mode, FPS fine tuning and so on.
>> On these two cameras, when you optimize for low light and you >> undercrank, ML will use the DIGIC method (which increases shutter >> speed at low frame rates). Otherwise, ML will use the >> 60D/600D-specific method (which does not have the side effect of >> increased shutter speed).
>> FPS needs some thorough testing especially on 600D; I'm not really >> sure it's rock solid. In particular, it may have to flip video modes >> to apply the changes (they can no longer be applied while recording).
>> Minor stuff: >> - Q button fixed on 600D (confirmed by Scrax). >> - on 60D/600D, since Q button does not work while recording, you can >> use the INFO button instead (but normally you shouldn't need this). >> - LED feedback for bulb timer (for those who use it with main display >> off).
>> Warning: I didn't run any real-world test with these changes. Only >> tried a few random things and ran the automatic tests. I'm sure there >> are still rough edges - they need to be reported so I can fix them.
> To post to this group, send email to ml-devel@googlegroups.com > To unsubscribe from this group, send email to > ml-devel+unsubscribe@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/ml-devel?hl=en
I noticed one bug on the 5d MKII, when I format the card the progress bar
will get to the end, and then the red light indicating the card is being
accessed starts flashing, locking up the entire UI until it's done. It
appears either the formatting progress bar finishes too early or completes
before the formatting is even started.
On Sat, May 12, 2012 at 8:26 PM, Alex <broscutama...@gmail.com> wrote:
> Looks like the logs are very good for catching hidden bugs. So... time
> for the fixes :)
> - ML compiles without warnings (credits: Takashi Miyake).
> - Logs are now saved silently, without prompting (since they are not
> really crashes, just warnings). Still, if you find them on the card,
> please send them.
> - Startup with 64GB cards should no longer cause trouble. Also,
> startup with smaller cards should be faster.
> - Zoom on manual focus: I couldn't reproduce any of the two bugs, but
> I've moved it to a different task. Please re-test. On 60D, make sure
> the cropmarks are compressed.
> For FPS on 60D/600D: I've tried to combine the two FPS methods, so now
> you can get all the FPS features you had in the old implementation,
> plus the new ones. This means:
> - timelapse in low light (down to roughly 0.2 fps at 1/fps),
> - timelapse in daylight (fast exposures at around 3 fps),
> - slight overcranking (34fps),
> - exact FPS, slitscan (high jello) mode, FPS fine tuning and so on.
> On these two cameras, when you optimize for low light and you
> undercrank, ML will use the DIGIC method (which increases shutter
> speed at low frame rates). Otherwise, ML will use the
> 60D/600D-specific method (which does not have the side effect of
> increased shutter speed).
> FPS needs some thorough testing especially on 600D; I'm not really
> sure it's rock solid. In particular, it may have to flip video modes
> to apply the changes (they can no longer be applied while recording).
> Minor stuff:
> - Q button fixed on 600D (confirmed by Scrax).
> - on 60D/600D, since Q button does not work while recording, you can
> use the INFO button instead (but normally you shouldn't need this).
> - LED feedback for bulb timer (for those who use it with main display off).
> Warning: I didn't run any real-world test with these changes. Only
> tried a few random things and ran the automatic tests. I'm sure there
> are still rough edges - they need to be reported so I can fix them.
> To post to this group, send email to ml-devel@googlegroups.com
> To unsubscribe from this group, send email to
> ml-devel+unsubscribe@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/ml-devel?hl=en
