Helicon Soft Forum

I just recently started using Helicon Remote with my Canon R5, and have noticed some issues that vary between "merely irritating" and "show stopper".

First, the merely irritating issue: significant inaccuracy in position when moving between points A and B

• I start by finding position A, and then finding position B,
  • Direction "towards infinity", using "towards camera" does not seem to change the behaviour

• When returning from B to A, either by clicking on A in the UI, or by starting shooting the stack, the position A is not re-acquired precisely:
  • The new stopping point, "new A", is in number of steps, somewhere from 10% to 30%, closer to the camera than the original A
  • Point B, at the end of the shooting, is also closer to the camera by approx the same amount
  • The effect accumulates - shooting a new stack again moves A (and thus B) somewhat closer to the camera

• The error is not constant, but rather seems to vary between around 10% and around 30%

• The effect appears to be independent of the various settings for Canon shooting.
  • "Big step" and "small step" show the same behaviour, with "Big step" maybe having a smaller error
  • "Waiting time between focus steps" also do not seem to have any/much influence, waiting for a long time causes the error in the same way as waiting for a short time does
  • The number of steps to perform for "medium" and "large" also does not appear to have a significant influence

• Work around: set A and B so far away from the proper points that my stack at least has all necessary images
  • This will work for one stack
  • But it will not work multiple stacks in the same situation, e.g. when I've varied the lighting, because the error accumulates over stacks

The show stopper is way worse: loss of images in certain situations, either by saving the wrong version (JPEG instead of RAW) or corrupting the memory card

• I have now lost images twice during a shoot, in the following ways:
  • Wrong image type saved to the wrong place - JPEG save to Mac instead of RAW saved to the SD card
  • Corrupted SD that could not be read on a computer afterwards, or by the camera, during the shoot

• These two problems have occurred twice, now, in the following situation:
  • First shooting single images without Helicon Remote being connected, using a standard wired remote RS-80 N3
    • These images were in the correct place and type (RAW on SD card in the camera)
  • The connecting Helicon Remote via USB-C, shooting a couple of stacks with the RF100
    • These images were also in the correct place and type (RAW on SD card in the camera, a copy sent to the Macbook)
  • Then changing lens to EF-65, using a rail to manually position the lens and the RS-80 N3 to trigger exposures
    • These images were not saved to the camera in any form, but sent to the Macbook as JPEGs
  • No settings in camera or in HeliconRemote were changed, either before or during
  • All images saved to the SD card ended up getting lost due to corruption
    • By using a disk utility on the SD card I could in one situation recover those images from the the card, but not in the second
  • I can avoid this problem by simply not using Helicon Remote in this way (remote viewer), although I had hoped to be able to.

I am aware that these are not necessarily problems that stem solely from Helicon Remote. Having, in a prior life, been professionally involved with the Canon SDKs I know that they are, to put it lightly, "very temperamental" and "not easy to program against"

If somebody knows what I can do about the first problem, I would be very happy to hear about it. My current workaround is not that satisfying The second problem I can for now work around by simply not getting in to that situation.

Setup ("latest versions" as of this weekend)

• Canon R5 with latest firmware

• Canon RF100 macro, latest firmware

• Helicon Remote, latest version, on either iPad or Macbook Air, also on most recent OS
  • The described behaviours do not vary between iPad and Macbook
  • Wired connection, via USB-C, not wireless

Regards,
Michael

Hello Michael,

1. Sorry about the image format switching to JPEG. That's not supposed to happen, and we weren't aware of such issues with R5, but we did fix a similar problem for the new R6 Mark II. Hopefully the new version solves that, please give it a try, and if it happens again - send a bug report as soon as you notice (menu - Help - Report a bug). Here's the new version.

2. As for the memory card being corrupted, that can't be caused by the software as the card is 100% managed by the camera itself (its firmware). Could it be that the card itself is defective?

3. Check that your "pause between focus movement" is set to 200 ms, that's a 100% safe value. If the value is too short - you will get the exact problem you've described, but that's not the only possible reason for it.
Also check that only the most basic autofocus mode is enabled - no tracking/servo/continuous/face detection etc.
If nothing else helps, this is just the lens, it's not a high precision mechanism (nor does it have to be for the purposes of regular camera use). Try a different lens if you have one, try mounting the camera upside down. Some lenses exhibit this problem to a much larger degree than others. The ultimate solution is a macro rail.

1. Yes, that is a newer version than the one I've been using. I wonder why it hasn't shown up, when I did a Help->Check for updates this weekend? It also didn't show up, when I did it just before. Curious. Do note though, that the issue was not just "switching to JPEG", it also did not leave a copy on the card. My settings are "Copy to computer", unchanged. I'll give the new version a try

2. Card itself is faintly possible, but I very much doubt it. a) it was not the same card the second time (I threw the first one out and used a different, newer card the second time) and b) I have in 20+ years had exactly two cases of corrupted cards - these two cases . Coincidences are, of course, possible... but...

"can't be caused by software" that I am not too sure about - how about the situation where the camera attempts to do one thing, the SDK attempts to do another, and the camera FW is not programmed to prevent that, and tries to do both things at the same time? I've seem weirder things... (And I know that neither Canon FW nor the Canon SDKs are particularly well guarded against race conditions).

3. "200 ms, that's a 100% safe value" Too short was the first thing I thought of, and thus tried numerous values between 50ms and 500ms, and I saw the same general behaviour. I cannot remember if I explicitly tried 200ms (unless that is the default value) - is there something special about the exact number? Again, I can try again.

I also tried to shoot "towards camera" to compensate for the lens being pointed slightly downwards as I thought gravitation might cause the focusing elements to "slip" more in one direction that in the other. Didn't make any difference. I can try upside down next time.

The camera is of course in full manual, not automatism of any kind activated for this kind of shooting.

"not a high precision mechanism" Problem is more that they are built to be have high precision relative to a current position mechanism (otherwise PDAF would not work), but not to have high absolute precision. The latter being necessary for what Helicon Focus does.

"The ultimate solution is a macro rail." I know, I know That is the reason why I switched to a manual high-precision macro rail. And lost all images as a consequence (second case, image corruption)

Thank you for the quick answers. I'll (retry) some of the suggestions next time.

Regards,
Michael

Unfortunately, I don't have any useful suggestions for the lens movement inconsistency. It's not a problem for most users with most lenses, but it starts becoming a problem if you want to reproduce the same shoot a number of times, and gets a lot worse if you happen to have a particularly unfortunate lens.

1. This is not a version that has been released yet so you couldn't possibly have discovered it. On top of that, the automated update notifications are disabled right now, but that's temporary.

We know why the image format has changed (which should be fixed now), but we don't know why a copy wasn't saved to the memory card. This normally only happens when no card is inserted. The only way we can investigate that is if you (or someone else) sends a bug report after that happened, i. e. at least one shot was taken and didn't make it to the memory card.

2. There is no race condition possible because there are no racing processes. The camera is requested to take a photo, sends a notification once the image is ready for download, then the software starts downloading it. If the memory card really does get corrupted - it is a firmware bug, and we can only guess what triggers it, i. e. what we could do differently to avoid it.

3. 200 ms is not a special value, just one that we're pretty sure to be enough in the worst case. So if you had at least 200 or more, then it won't help, the problem is in the lens itself.
Autofocus works in a feedback loop, so it's not necessary for the lens to always move the same amount for the same input signal. No feedback on the current lens position is available to the software.

I'm using the exact same setup for my auction house catalogue project since the Canon R5 launch. So far I didn't run into any problems as you described. One suggestion from my experience is, mark point B first then point A. You should be able to get a very consistent result. Hope this helps a little.