@ all Random Nerds,
This morning I did arrange a new ESP32CAM board. – Long story short:

Connecting………………………………..
A fatal error occurred: Failed to connect to ESP32: No serial data received.
For troubleshooting steps visit: https://docs.espressif.com/projects/esptool/en/latest/troubleshooting.html
Failed uploading: uploading error: exit status 2

The new board I can’t even flash anymore! – I feel awkward about such product quality and non existing QA processes.
Anyway, let’s swap the new camera module and retry flashing with the old board. –
Jepp, finally there is the picture stream!
2 days of my life gone, caused by poor product quality.
Anyhow, that’s a big part of a Random Nerd Life and I am so thankful to have learned how to cope with the core dump error. – Simply look at the partion table, if there is a core dump space already alocated.
Happy Hackin’
-P

Hi.
 
What version of the ESP32 boards do you have installed?
Go to Tools > Board > Boards Manager, search for ESP32 and check your version.
 
What board are you selecting to upload code in Tools > Boards?
 
Regards,
Sara

Hi Sara,
Both boards are identical (although different sources) and both attempted to flash as AI Tinker ESP32-CAM.
The ‘old’ board from this weekend could be flashed, but no camera stream.
The ‘new’ board can’t be flashed (please see error message above), but serial connection did work looking at the board details within IDE.
Anyway, it is finally solved: The camera module from the old board did not work and I continue with ‘old’ board, but new camera module.   J
Best,
-P 

Ok.
For the new board, try to select another board “ESP Wrover Module” to upload.
Let me know if there’s any change..
Regards,
Sara

Thank you for your idea to try with a ESP-WROOM-32 board.
Here I do experience the same situation: Initial test failed, because the file is too large. So I swapped to a non OTA setup.

Now flashing results again in the following failure:

esptool.py v4.5.1
Serial port /dev/cu.usbserial-340
Connecting………………………………..
A fatal error occurred: Failed to connect to ESP32: No serial data received.
For troubleshooting steps visit: https://docs.espressif.com/projects/esptool/en/latest/troubleshooting.html
Failed uploading: uploading error: exit status 2

This board can’t be flashed, but I was happy to combine both boards and cameras to finally have one working system. Now I could run through your well written ESP32-CAM tutorial!
Thanks much and happy coding,
LoetLuemmel!

Yes.. it seems that your board can’t go into flashing mode… I’m not sure why that is happening. 
 
Nonetheless, I’m glad you have at least one working board and I hope you enjoy the eBook.
Regards,
Sara

Thank you Sara,
Yes all good and I’ve already passed the book with high excitement.
Just one comment or suggestion: Could you consider also adding RTSP info and code snippets to the book? I wanted to test a RTSP stream for existing equipment and was surprised not fining anything.

But it’s summer time, maybe something for a dull day in next fall or winter.   J
Best,
LoetLuemmel

Hi again.
 
About RTSP, one of our readers created a project using RTSP. I asked him to share his resources with us.
 
This is the email that he sent me with all the information:
“”CameraRainGauge.ino” project currently uses webRTC; servered from a Raspberry Pi running a Docker container of Wyze-Bridge.   Camera used is a Wyze Cam v3, basically an ip camera.  Version on Github still uses RTSP.
Project can be found on Github:  CameraRainGauge.ino.  The HTML page for “CameraRainGauge.ino”  Gist of RTSP embedded web page   Additional documentation is available on Hackster.io: Data Logger with live video feed ”
 
There’s also an entry on our forum about this subject: Embed RTSP in html
 
I hope this is helpful.
Regards,
Sara

Hi Sara,
Hi Rui,
Thank you for the link to the RTSP Raspi server!
FYI: I do not want to make advertisements for competitive content, so I direct this message just to you admins.
I found RTSP code ready to run even on the ESP32-CAM itself:

github.com/rvjenya/rtsp-esp32cam
It compiles OK and after the usual SSID and Passwd changes I could stream to a VLC client on rtsp://.

Enjoy your evening!
-Pit

Hi @everyone,
I finally re-discovered my ‘dead’ ESP32-CAM boards!!!
In case you can’t flash them, …

• connect the board
• select the correct COM port for your USB cable
• open the serial console with 115200
• press- down the IO0 button on your ESP32-CAM-MB
• press-down the RST button on your ESP32CAM (!!! not on the MB32-CAM-MB!!!)
• You should see a message in the Serial Console:

  rst:0x1 (POWERON_RESET),boot:0x3 (DOWNLOAD_BOOT(UART0/UART1/SDIO_REI_REO_V2))
  

  waiting for download
  

  ets Jul 29 2019 12:21:46
  
  
  

• Now flash as usual
• The program (at least with me) doesn’t start after flashing, I need to press RST on the ESP32-CAM in order to start the program.

Intersting: I have 3 ESP32-CAM boards and one behaves completely different. No flash problems, after flash program starts automatically. PCB and HW looks slightly differrent, so there are different versions out there.
Anyhow, I am so glad that finally all three boards can be flashed!
One camera module is still dead, but this is another story and doesn’t touch the flashing.
 
Happy Hackin’
LoetLuemmel 

Hey William,

Thank you for that link!!
I was also experimenting with RTSP and also found inspiring sketches here.
Whow, your link even has OTA on top.
Why is RTSP so compelling? Don’t know, I have a 8MP CAM running (no ESP) to see, if the postman brings new toys. But the system transmits huge loads of MJPEG and is very memory intense on the RX side. Now I want to experiment with video streaming, if I can achieve an acceptable resolution and a ‘fire & forget’ picture stream.
In addition, I am working on a racing horse with 512 GPU cores. It offers me Real Time Web Classification of a RTSP stream with unbelievable ~120FPS(!). I would like to process object learning, so I would receive an alarm, when the yellow post car enters my drive way, ha, ha!   J 
But that Nvidia machine is too energy hungry and I’d like to explore, how far we could get just using the inexpensive ESP32-CAM module.
 
I wish all of us an extremely interesting coding weekend.
Yours LoetLuemmel!

Good Morning Loet,

Camera I use for video streaming is a Wyse Cam v3.  My method of video streaming is a little different; with a Raspberry Pi 4 running Portainer with a Docker container, “Wyze-Bridge.”  This “Wyze-Bridge” interfaces with your “Wyze” account; using your “Wyze” account credentials to capture the H.264 video stream and provide URL’s that can be embedded in a web page.  “Wyze-Bridge” is specific to the “Wyze” cameras and does not need “Wyze” RTSP firmware.

Previously I used a Windows 11 computer to run Unreal Media Server, 24/7 consuming around 300 Watts per hour.  Switching to Rapberry Pi saves 290 Watts per hour!  UMS provided RSTP URL that was embedded in HTML web page.

Current Live view of Wetland Pond behind our home.

Live view web page is served from ESP32 running Arduino code: “CameraRainGauge.” , which receives URL from Docker container, “Wyze-Bridge” running on Raspberry Pi 4.

Regards,
William

Hi Loet.
Thanks for clarifying the uploaded methods for the other ESP32-CAM boards.
 
And thank you William for the useful information provided.
Regards,
Sara

Searching for compatiable ESP32-CAM cameras; asked Google’s Bard:

“ESP32-cam compatiable 5 meg pixel camera?

Sure, here are some 5MP cameras that are compatible with the ESP32-CAM:

• Arducam Mega 5MP Camera: This camera is a great option for those who need high-resolution images. It has a resolution of 2592 x 1944 pixels and can output images in JPEG, BMP, or RAW format. The Arducam Mega Camera is also compatible with both 3.3V and 5V systems.
• HiLetgo 5MP Camera Module: This camera is a more affordable option than the Arducam Mega Camera. It has a resolution of 2592 x 1944 pixels and can output images in JPEG format. The HiLetgo 5MP Camera Module is also compatible with both 3.3V and 5V systems.
• SparkFun OV5640 Camera Module: This camera is a good option for those who want a camera that is easy to use. It comes with a pre-soldered PCB and a variety of libraries and drivers that make it easy to get started. The SparkFun OV5640 Camera Module has a resolution of 2592 x 1944 pixels and can output images in JPEG or BMP format.

When choosing a 5MP camera for the ESP32-CAM, it is important to consider the following factors:

• Image resolution: The resolution of the camera will determine the quality of the images that it can produce.
• Cost: The cost of the camera will vary depending on the model and features.
• Availability: Some cameras are more widely available than others.
• Support: The availability of libraries and drivers for the camera will make it easier to develop applications with it.

Once you have considered these factors, you can choose the 5MP camera that is best suited for your needs.”

–from Google’s Bard

Here is a ov5640AF (auto focus) from AliExpress:

OV5647 Camera Module for Raspberry Pi 3B 4B 3B+ Adjustable Focus 120 130 200 160 Degree 3.6MM HD 5 Million Pixel Night Vision

Selection of differing degrees of “capture,” viewing angles; choose according to your need.

Availiable from Amazon:  HiLetgo OV5647 5MP Camera OV5647 Camera Module Raspberry Pi Camera for Raspberry Pi A/B+/2 Model B with Cable

Amazon Hi-Letgo ov5640 camera is fixed-focus.

Library for Auto focus feature of ov5640AF:  Enable OV5640’s autofocus function on ESP32 AI-THINKER Board

Regards,
William

Hi William,
Thank you so much for this valuable information!
I wasn’t expecting real alternatives for available camera modules, because the ESP32 is a low cost approach. – But obviously you did correct that statement!
Especially I like the OV5640 autofocus option (thanks for the link!).
Interesting, that Google’s Bard can also serve as shopping assistant. – I will give it a try with openAI for the European user or shopping space.   J

Thanks again for your very details camera option research and have a great start of your week!
-Loet

Quick update: Here‘s a comment from Bard‘s competitor: 
The ESP32-CAM can indeed be used with a 5MP OV5640 camera module. This camera has a better resolution compared to the commonly used 2MP OV2640 module, which allows for higher quality images and video.
The OV5640 is a 1/4-inch CMOS image sensor from OmniVision that combines a high-quality 5MP (2592×1944) camera and an integrated ISP in a single chip. Please ensure to update your firmware and libraries accordingly to support this camera on your ESP32-CAM board. You’ll need to set the right configurations in your software to use the OV5640 instead of the OV2640.
Also, be aware that increased image quality could lead to slower processing times and higher memory usage, so ensure that your project can handle this.
–
Wise words. 

There are some available with berrybase in Berlin, but the price tag is quite delicate with 44.90€.
On the other side, if it is good there is always a price for it!
Happy coding!

Good Morning,
Just a FYI; there appears to be at least two versions of the OV5640 camera, one being fixed focus (OV5460), the other auto focus (OV5640AF).  Inaddition to the camera is available with differing angles of view; not sure that the differing angle of view cameras are auto focus!.
“
Hardware Modification
“To use the AF function of the OV5640 on the ESP32 AI-Thinker board, you need to connect 3.3v to the last 24th pin(AF-VCC). I share what I’ve been working on. You can also see how AF works.” — from 0015 Eric’s Github for the ov5640AF Library.

OV6840AF is available from Amazon.

OV5640AF Camera only is available from AliExpress.

Regards,
William

Are there alternative ESP32-CAM cameras greater than 5 Mega Pixels?

“There are a few 5MP and greater cameras that are compatible with the ESP32-CAM:

• HiLetgo 8MP Camera Module: This camera has a resolution of 3264 x 2448 pixels and can output images in JPEG format. The HiLetgo 8MP Camera Module is also compatible with both 3.3V and 5V systems.
• SparkFun OV5647 Camera Module: This camera has a resolution of 2592 x 1944 pixels and can output images in JPEG or BMP format. The SparkFun OV5647 Camera Module is also compatible with both 3.3V and 5V systems.
• Arducam Mini 8MP Plus Camera: This camera has a resolution of 3264 x 2448 pixels and can output images in JPEG, BMP, or RAW format. The Arducam Mini 8MP Plus Camera is also compatible with both 3.3V and 5V systems.

If you need a camera with even higher resolution, there are a few 12MP cameras that are compatible with the ESP32-CAM:

• HiLetgo 12MP Camera Module: This camera has a resolution of 4000 x 3000 pixels and can output images in JPEG format. The HiLetgo 12MP Camera Module is also compatible with both 3.3V and 5V systems.
• SparkFun OV5648 Camera Module: This camera has a resolution of 4000 x 3000 pixels and can output images in JPEG or BMP format. The SparkFun OV5648 Camera Module is also compatible with both 3.3V and 5V systems.
• Arducam Mini 12MP Plus Camera: This camera has a resolution of 4000 x 3000 pixels and can output images in JPEG, BMP, or RAW format. The Arducam Mini 12MP Plus Camera is also compatible with both 3.3V and 5V systems.

When choosing a camera for the ESP32-CAM, it is important to consider the following factors:

• Image resolution: The resolution of the camera will determine the quality of the images that it can produce.
• Cost: The cost of the camera will vary depending on the model and features.
• Availability: Some cameras are more widely available than others.
• Support: The availability of libraries and drivers for the camera will make it easier to develop applications with it.

Once you have considered these factors, you can choose the camera that is best suited for your needs.
Can these be used for video streaming?

Yes, the cameras I mentioned can be used for video streaming. The HiLetgo 8MP Camera Module, SparkFun OV5647 Camera Module, and Arducam Mini 8MP Plus Camera can all output video in JPEG format at a resolution of 3264 x 2448 pixels. The HiLetgo 12MP Camera Module, SparkFun OV5648 Camera Module, and Arducam Mini 12MP Plus Camera can all output video in JPEG format at a resolution of 4000 x 3000 pixels.

The video streaming quality will depend on the resolution of the camera, the frame rate, and the bandwidth of the network connection. For example, if you are streaming video at 3264 x 2448 pixels at 30 frames per second, you will need a network connection with a bandwidth of at least 10 Mbps.

Here are some additional factors to consider when streaming video from an ESP32-CAM:

Frame rate: The frame rate determines how many frames are captured per second. A higher frame rate will result in smoother video, but it will also require more bandwidth.Compression: The compression algorithm used to encode the video will affect the quality of the video and the bandwidth required.

Codec: The codec is the format used to store the video data. Some codecs are more efficient than others and will require less bandwidth.

Once you have considered these factors, you can choose the settings that are best suited for your needs.”  –from Google’s Bard

Regards,
William

Just learned of the Arducam Mega!

Arducam Mega – A Camera Solution for ANY Microcontroller
Arducam Wiki –MEGA SPI Camera

Arducam –ESP32-CAM: Machine Vision Tips, Camera Guides and Projects

William

Hi William,
I’m impressed about the wide spectrum of possible CAM solutions to expand the ESP32 with camera solutions for all budgets. Your consideration of bandwidth is the core thing to worry about, I’d suggest as a starting point to plan new  projects. 
Right now my modules all just have the built-in WiFi antenna. Along with the 2MP cam, this is a great and cheap solution for small projects close to the existing WiFi infrastructure.
A couple of years ago I started with the brand new 8MP RasPi cam connected remotely by WiFi. The cam took great pictures and I couldn’t believe how much bandwidth was required. For the planned location, WiFi was simply not possible with the existing access point setup. – Long story short: I finally wired the cam with an ethernet cable.
Usually for projects it is a balance of what do I want to achieve and how much I am willing to sacrifice budget-wise.
Here the ESP32-CAM did really impress me with it’s simplicity, availability and unbelievable price tag.
Now I found super cheap offerings for just the OV5640 module with small ribbon cable for just ~6US$! – Maybe these are fake ones, but I am scratching my head if it might be worth a try…
Q: Have you already been successful accessing a 3rd party cam via SPI? I did an OLED conversion and remember it was very time consuming.
Plus great finding: I already bought a OV5647 camera kit about half a year ago and didn’t use it (…too bizzy!). It came along with IR-LED lights, so ready to go for a day and night job.
Only problem is the MIPI connector, I don’t know if I can find ways to connect it to the ESP32 boards. Maybe there is an adapter board existing, next step would be finding or creating a driver.
But that’s the interesting thing with our hobby: Endless ocean waiting for you to explore, ha, ha!   J
Thanks for your comprehensive CAM research and happy coding,
Loet

Morning Loet,

Responses from my prompts are from Google’s Bard (AI).

Just a FYI; there appears to be at least two versions of the OV5640 camera, one being fixed focus (OV5460), the other auto focus (OV5640AF).  Inaddition to the camera is available with differing angles of view; not sure that the differing angle of view cameras are auto focus!.

Hardware Modification

To use the AF function of the OV5640 on the ESP32 AI-Thinker board, you need to connect 3.3v to the last 24th pin(AF-VCC).

I share what I’ve been working on. You can also see how AF works.” — YouTube video from Eric’s, Github for the OV5640AF Library.

Eric’s Github for the OV5640AF Library.

Might get more WiFi bandwidth with the ESP32-C6-DevKitC-1-N8.

Details ESP32-C6-DevKitC-1-N8; max WiFi bandwidth?

The ESP32-C6-DevKitC-1-N8 is a development board based on the ESP32-C6 chip from Espressif Systems. The ESP32-C6 chip is a powerful Wi-Fi and Bluetooth module that supports up to 150 Mbps of data throughput.

The ESP32-C6-DevKitC-1-N8 board has a number of features that make it ideal for IoT applications, including:
A 240 MHz dual-core Xtensa LX7 microcontroller
4 MB of flash memory
520 KB of RAM
Support for Wi-Fi and Bluetooth Low Energy
A wide range of GPIO pins
Low power consumption

The ESP32-C6-DevKitC-1-N8 board is a great choice for applications that require a powerful and efficient Wi-Fi module. It is also a good choice for applications that require a wide range of GPIO pins.

Here is a table that summarizes the key features of the ESP32-C6-DevKitC-1-N8 board:
| Feature | ESP32-C6-DevKitC-1-N8 |
|—|—|—|
| Processor | ESP32-C6 |
| Clock speed | 240 MHz |
| Flash memory | 4 MB |
| RAM | 520 KB |
| Wireless connectivity | Wi-Fi and Bluetooth Low Energy |
| Wi-Fi speed | 150 Mbps |
| GPIO | 42 pins |
| Analog input/output | 11 pins |
| Power consumption | 100 mA |

Does ESP32-C6-DevKitC-1-N8 have psram?

Yes, the ESP32-C6-DevKitC-1-N8 has 2 MB of PSRAM. PSRAM is a type of memory that is faster than the on-board RAM. It is used to store data that is frequently accessed, such as image buffers and audio buffers.

Here is a table that summarizes the memory on the ESP32-C6-DevKitC-1-N8 board:
Type      Size
On-board RAM  520 KB
PSRAM 2 MB

The ESP32-C6-DevKitC-1-N8 board is a great choice for applications that require high-speed data access. It is also a good choice for applications that require a lot of memory, such as image processing and audio processing applications.

Does ESP32-C6-DevKitC-1-N8 have extnal antenna connector?

Yes, the ESP32-C6-DevKitC-1-N8 has an external antenna connector. The antenna connector is a U.FL type connector, which is a small, low-profile connector that is commonly used for Wi-Fi and Bluetooth antennas.

The ESP32-C6-DevKitC-1-N8 board comes with a PCB antenna, but you can also connect an external antenna for improved performance. To connect an external antenna, you will need a U.FL to SMA adapter cable.

No I have not experimented with other 3rd party cameras via SPI.  I am planning to purchase an Arducam Mega 5MP, Autofocus (SPI) camera.

Regards,
William

Hi William,
The possibilities are great, without any doubts!

>>No I have not experimented with other 3rd party cameras via SPI.  I am planning to purchase an >>Arducam Mega 5MP, Autofocus (SPI) camera.

Same here. I’ll continue with the AI Thinker CAM to implement some simple RTSP streams. These should be processed by Raspi with a Coral TPU. For the 8 and 12 MP CAMs, I need to find an application. They are too heavy for a microcontroller.
With my best regards over to yours,
Loet

Good Afternoon,

Intersting tutorial on Arducam MegaCamera from iotprojectsideas.com; includes Gerber files for ordering a printed circuit board…

Arducam Mega,  5 MP Camera Tutorial

William

Good morning William,
That’s really straight forward. – I am excited to read about 1st hand experience with the 5MP cam.
Plus I like the PCB offering for a ready-to-go solution.
Thanks for sharing,
Loet

Good morning William,
That’s really straight forward. – I am excited to read about 1st hand experience with the 5MP cam.
Plus I like the PCB offering for a ready-to-go solution.
Thanks for sharing,
Loet