Experiences with Celestron NexImage cameras?

I've looked at cameras, and I have seen a few deals on NexImage cameras lately which look on the surface like they'd be good deals. However, I can find almost nothing about how well, or poorly they perform under Linux and INDI. The few that I do find either seem to refer to the old ones, with a Phillips chip. With Linux, so often things are either: No mention, because you plug it in and have 0 problems, or no mention, because no one has ever looked at it.

So I'm wondering if anyone has a NexImage, and if so, which one, and how well do they function? I'm guessing they'd be v4l devices, but based on a few posts on the internet I could find. It looks like the original color one works fine.

So if people could comment on that it would help me, or someone else searching the internet before asking questions.
Models, that I'm aware of:
NexImage Solar System Imager Model: 93709 (Sensor: ??, can't find it on Celestron's site, from comments elsewhere seems to be a Phillips chip also used in webcams?)
NexImage Burst Color Model: 95518 (Aptina AR0132 color sensor)
NexImage Burst Monochrome Model: 95519 (Aptina AR0132 mono sensor)
NexImage 5 Model: 93711 (Micron MT9P031, Some reference I can no longer find, I saw that it worked with the QHY5 drivers? Same chip as the QHY5P-II-M)
NexImage 10 Model: 93708 (ON Semi MT9J003 Color CMOS, USB 3)

(Useful reference: s3.amazonaws.com/celestron-site-marketin...ras+Compare_2017.pdf because of this, and having some technical details, I'll mention the models of Skyris too, in case that goes away, and some dead forum post is useful)

Skyris 132C Model: 95508 (Aptina AR0132 color, USB 3)
Skyris 132M Model: 95509 (Aptina AR0132 mono, USB 3)
Skyris 236C Model: 95506 (Sony IMX236LQJ color, USB 3)
Skyris 236M Model: 95507 (Sony IMX236LQJ mono, USB 3)

I did find a seemingly great deal on a NexImage 5, so I did get it.

Here is the USB ID:
Bus 003 Device 055: ID 199e:8207 The Imaging Source Europe GmbH

Based on the driver package prior to seeing it, I was able to track back it, and most other Celestron cameras back to The Imaging Source. Which lead me to here: github.com/TheImagingSource/tiscamera/wi...USB-2.0-CMOS-Cameras

I followed that procedure, with the commands in bold, and the serial number replaced with <SERIAL>
sudo ./firmware-update -l

Found 1 device(s).

Name - ID - Serialnumber
NexImage 5 - 199e:8207 - <SERIAL>
sudo ./firmware-update -id <SERIAL>

Device manufacturer: Celestron
Product name: NexImage 5
Serial number: <SERIAL>
VendorID:ProductID: 199e:8207
Firmware version: 129
UVC mode is: off
Camera EEPROM size: 32768

sudo ./firmware-update -ud <SERIAL> -f ../../../data/firmware/usb2/dfk72uc02_3012.euvc

!!! Attention !!!
This action could break your camera.

Do you really want to proceed? [y/N] y
Firmware Size: 19968 EEPROM Size: 32768
100 %

Upload successful!
Please reconnect your camera.

sudo ./firmware-update -id <SERIAL>

Device manufacturer: Celestron
Product name: NexImage 5
Serial number: <SERIAL>
VendorID:ProductID: 199e:8207
Firmware version: 196
UVC mode is: on
Camera EEPROM size: 32768


Following that, the camera worked, and at low exposure time, gave me a great picture of some garden flowers, along with insect, via a Celestron First Scope Telescope (using v4l2 and guvview. A 76mm/300mm FL telescope. I'd see about a picture of stars, but there's the issue of these things called clouds that show up as an extra on seemingly any astronomy related purchase. This one came with a very large free package of wind as well!

Unforuntely, INDI doesn't seem to be as cooperative as guvcview.

This appears to be because of the formats supported:
ioctl: VIDIOC_ENUM_FMT
Index : 0
Type : Video Capture
Pixel Format: 'GREY'
Name : 8-bit Greyscale

Index : 1
Type : Video Capture
Pixel Format: 'GRBG'
Name : 8-bit Bayer GRGR/BGBG

Which looks like a 2 line pattern, if that's to be believed. Based on looking at guvcview's colorspaces.c it looks like that's how it's handling it.

If trying to stream, you get ... basically static.
For now, greyscale seems to work. I've been piddling around with a new conversion:


I've added in the case to adjust the size of the rgb24_buffer based on pixel format: case V4L2_PIX_FMT_SGRBG8: in v4l_builtin_decoder.cpp, allocBuffers() but I'm a bit at a loss on it. Sometimes, it runs for a few minutes, and sometimes it fails immediately. It's probably something stupid that someone will see immediately above.

Unless the *src has some size issue.

Ok, I've got it working. Amusingly helped by some networked RGB lights called Filimins that a friend started a business building.

Or did last night, I let my computer go to sleep and lost some controls (v4l2 problem not INDI) Like the R/G/B gains.

Here's the branch if you can use it, via some git magic, otherwise I'll make a clean pull request at some point in the next few days. github.com/james-lan/indi/tree/neximage (I'm based off of azwing's so that causes some issues I think.)

Some more notes for anyone else who tries to use this: I got the controls back, apparently, there's a kernel->userspace transfer or something going on with uvc type things.

You'll need the file from the The Imaging Source github above, you need uvcdyntrl installed, then run this:
uvcdynctrl -i data/uvc-extensions/tisEUVC.xml -d /dev/video1

It goes from this:
v4l2-ctl -d /dev/video1 -l
brightness (int) : min=0 max=255 step=1 default=12 value=12
gain (int) : min=4 max=63 step=1 default=16 value=36
exposure_absolute (int) : min=1 max=300000 step=1 default=127 value=127
focus_absolute (int) : min=0 max=1000 step=1 default=0 value=0
privacy (bool) : default=0 value=0

to this:
v4l2-ctl -d /dev/video1 -l
brightness (int) : min=0 max=255 step=1 default=12 value=12
gain (int) : min=4 max=63 step=1 default=16 value=36
trigger_global_reset_shutter (bool) : default=0 value=0
exposure_absolute (int) : min=1 max=300000 step=1 default=127 value=127
focus_absolute (int) : min=0 max=1000 step=1 default=0 value=0
privacy (bool) : default=0 value=0
trigger (bool) : default=0 value=0
software_trigger (button) :
gainr (int) : min=0 max=63 step=1 default=1 value=36
gaing (int) : min=0 max=63 step=1 default=1 value=36
gainb (int) : min=0 max=63 step=1 default=1 value=36
binning (int) : min=1 max=4 step=1 default=1 value=1
x_offset (int) : min=0 max=1794 step=2 default=0 value=0
y_offset (int) : min=0 max=1464 step=2 default=0 value=0

Of those:
Brightness does nothing that I can tell.
Binning breaks it.
Trigger of all types is are somewhat untested by me, though they seem related to the snapshot mode.
Focus does nothing.
Exposure Absolute is in 0.1ms intervals. (Up to 300000 = 30 seconds, which seems to work. (Haven't tried it a lot yet, that's based on the white as this thing's background that showed up on streaming about 30 seconds after!))
R/G/B gains do work. There's probably an ideal value... but that either takes a patient approach or is the thing that drives people mad trying to properly figure out.

Testing with an ETX-70, I can see a fair number of stars. Sadly I can see stars, but I start to sink in the mud, and I'm exhausted.

Pull request submitted: github.com/indilib/indi/pull/587

That should add support for the NexImage 5 to INDI as well as likely the NexImage 10 (Based on this: github.com/openastroproject/openastro/issues/252 it appears to use the same format.)

(The more of git I use the more it's both more powerful and really annoying.)

Just to note, I did get another NexImage5, and following my own instructions worked. (Always great when that happens.)

I’m having an issues getting mine going. I mirror my install on a netbook and on RPi. Everything was installed a week or so ago. I’ve had the v4l2 working on both. Rpi ccd works (poorly) and netbooks screen cam streams.
Running Ubuntu mate 19
Installed indi full

Neximage shows up running lsusb as in your example. It does not show up under /dev/video1 (video0 are other cams). Disabled the ccd on rpi off in bios on the netbook and rebooted but no joy.

All kinds of issues trying follow through the TIS instructions. Dependency not installing and files missing errors all over the place.

Oh actually reading the readme actually worked.