Running a Presentation from a Propeller - putting a BMP in VGA?
Ken Gracey
Posts: 7,400
Hey all,
Just returned from a trip to Taiwan, Hong Kong and China with part of the Parallax team (Jessica and Aristides). Hopefully I'll have time over the next couple of days to dribble out some pieces of the trip with everybody on the forums. The main purpose of the trip was to make arrangements for Propeller 2 packaging and testing in Taiwan.
Another purpose of the trip was to make presentations on the Propeller at Taiwan National University and National Tsing Hua University. The first presentation went very well - it was all demo based examples with *.wav file playback, speech synthesis robotics and quadcopters. Mixed in was a fair amount of presentation run from Keynote (the Mac PPT equivalent). The second presentation was loaded with some challenges when VMFusion wouldn't run the Propeller tool on my Mac (had disconnected a USB port during a prior session and this didn't restart well), requiring a system reboot and rearrangement of our plan. Thankfully there were two of us, giving us the ability to have Jessica sort out the computer issue while I took them outside for an ELEV-8 flight.
While doing these presentations I started thinking about the need for simplicity. We run some fairly complex presentations with switching between presentation software (PPT, Keynote, etc), the Propeller tool, PST, VGA and sometimes even video. I had the idea that the presentation should be run from the Propeller next time. I think this discussion has even come up on the forums before. It would require a simplification of the presentation methods, but that's fine - less is often more. Using this approach the PC would be a supplement to the presentation instead of the star of the show.
Some interesting things could be done with audience involvement. I could hand out some PropBOEs with heart rate sensors, color sensors, accelerometers, humidity/temperature, all loaded with XBees. The audience could become the demo, and the Propeller needn't be relegated to a slide when it could be shown. How many times have companies shown up at your workplace and taken too long to show you the sales/company/introduction stuff when you really want to see the demo?
We can count on VGA, but not composite at almost every location.
And so my question - could you direct me towards the right way of displaying single BMP (or other format, if appropriate) images into a VGA text driver?
This would help make the presentation more visual than displaying only text. I'd like to be able to show a Parallax board, a sensor, a graph, or anything else I could fit into this format.
In case I'm not asking the right question, I'll explain what I'm after as the end result. I'd like my slide show to be run by the Propeller. Each "slide" should contain some large text (equivalent to 30 pt font size or so) and a single image. I don't need to move images around, but be able to display them. Some slides would actually be real demos, using sensors/XBee or anything else of interest - haven't given this too much thought yet.
I can usually take care of the Propeller programming if I have a starting point. I'd love to be able to load up the text in DAT statements (could also use microSD) for simplicity and have a nice scalable way of achieving this goal. Ultimately it would be very easy to build a list of slides from an outline.
Any direction here would be appreciated, whether towards starter objects, other threads, or anything else. I'll be busy at Parallax the next couple of days but I'll check in on my thread at least a few times. I will always make time to load and run any samples.
My next major presentation is at the end of July.
Thanks everybody!
Ken Gracey
Just returned from a trip to Taiwan, Hong Kong and China with part of the Parallax team (Jessica and Aristides). Hopefully I'll have time over the next couple of days to dribble out some pieces of the trip with everybody on the forums. The main purpose of the trip was to make arrangements for Propeller 2 packaging and testing in Taiwan.
Another purpose of the trip was to make presentations on the Propeller at Taiwan National University and National Tsing Hua University. The first presentation went very well - it was all demo based examples with *.wav file playback, speech synthesis robotics and quadcopters. Mixed in was a fair amount of presentation run from Keynote (the Mac PPT equivalent). The second presentation was loaded with some challenges when VMFusion wouldn't run the Propeller tool on my Mac (had disconnected a USB port during a prior session and this didn't restart well), requiring a system reboot and rearrangement of our plan. Thankfully there were two of us, giving us the ability to have Jessica sort out the computer issue while I took them outside for an ELEV-8 flight.
While doing these presentations I started thinking about the need for simplicity. We run some fairly complex presentations with switching between presentation software (PPT, Keynote, etc), the Propeller tool, PST, VGA and sometimes even video. I had the idea that the presentation should be run from the Propeller next time. I think this discussion has even come up on the forums before. It would require a simplification of the presentation methods, but that's fine - less is often more. Using this approach the PC would be a supplement to the presentation instead of the star of the show.
Some interesting things could be done with audience involvement. I could hand out some PropBOEs with heart rate sensors, color sensors, accelerometers, humidity/temperature, all loaded with XBees. The audience could become the demo, and the Propeller needn't be relegated to a slide when it could be shown. How many times have companies shown up at your workplace and taken too long to show you the sales/company/introduction stuff when you really want to see the demo?
We can count on VGA, but not composite at almost every location.
And so my question - could you direct me towards the right way of displaying single BMP (or other format, if appropriate) images into a VGA text driver?
This would help make the presentation more visual than displaying only text. I'd like to be able to show a Parallax board, a sensor, a graph, or anything else I could fit into this format.
In case I'm not asking the right question, I'll explain what I'm after as the end result. I'd like my slide show to be run by the Propeller. Each "slide" should contain some large text (equivalent to 30 pt font size or so) and a single image. I don't need to move images around, but be able to display them. Some slides would actually be real demos, using sensors/XBee or anything else of interest - haven't given this too much thought yet.
I can usually take care of the Propeller programming if I have a starting point. I'd love to be able to load up the text in DAT statements (could also use microSD) for simplicity and have a nice scalable way of achieving this goal. Ultimately it would be very easy to build a list of slides from an outline.
Any direction here would be appreciated, whether towards starter objects, other threads, or anything else. I'll be busy at Parallax the next couple of days but I'll check in on my thread at least a few times. I will always make time to load and run any samples.
My next major presentation is at the end of July.
Thanks everybody!
Ken Gracey
Comments
Exciting stuff. You have a great job.
Regarding your question, I think our DVI Graphics Shield would be great for that:
http://forums.parallax.com/showthread.php?138121
You can show any kind of photo you want. You can simply save your Power Point slides as VGA bitmaps and show them as a slideshow...
It is possible also to make a great looking o'scope display (although I haven't done it yet).
If you really need VGA and not DVI or HDMI, then we will have a VGA plugin for that ready in about 2 weeks.
Then, you can do the same thing over regular analog VGA.
But, if you really want to show a bitmap in good o'l 6-bit color, I'm sure I've posted something sometime on that too...
Wouldn't look nearly as nice though...
I've done a few experiments that could be useful. I've got standard bitmaps on a $20 touchscreen which comes with a SD socket as a bonus.
which is 320x240 with 16 bits per pixel. In the process I wrote some code to decode the .bmp header. This is fairly simple and in then end just means extracting the width and height of the picture from the header and some code to read the file from the end as the bitmap format starts with the last row first. Oh and a sneaky thing with rounding that you need to correct for if an odd number of pixels in the row. The video above uses just 5 bitmap files to do all the graphics (background, slider background, knob, switch on and switch off).
Code to read a .bmp file into an external ram chip (would work storing to hub ram but only for small files)
PUB LoadBMPtoRamdisk(stringptr) | width,height,w,i ' load and display a bitmap file ' http://en.wikipedia.org/wiki/BMP_file_format scroll down about 1/3 of the way for the header formats ' get the width, height ' store these as 2 longs at the beginning of the ram file ' the bitmap format starts at the last row and works up so need to reverse the order ' on the ram store a 0x36 byte header - the bitmap header but put in the filename at location 0 ' this becomes a rudimentary file system that can be searched to find pictures 'fat.openfile(stringptr,"R") OpenFileRead(stringptr) fat.readdata(@sdbuffer,$36) ' get the header 0x36 hex bytes width := sdbuffer[$12] + sdbuffer[$13] << 8 ' only read in two bytes as never will be >64k wide or high height := sdbuffer[$16] + sdbuffer[$17] << 8 w := width * 3 ' this number of bytes per row, and round up to nearest 4 bytes et 327 goes to 328 The size of each row is rounded up to a multiple of 4 bytes (a 32-bit DWORD) by padding. w +=3 ' add 3 and w &= %11111111_11111111_11111111_11111100 ' round down repeat i from 0 to 11 ' add the file name at the beginning of the header sdbuffer[i] := byte[stringptr][i] docmd("S",@sdbuffer,RamDiskPointer,$36) ' store header to ram i := $36 + RamDiskPointer + ((height - 1) * width) ' start + header and on the last row and work back repeat height fat.readdata(@sdbuffer,w) ' read in the first row docmd("F",@sdbuffer,@rambuffer,width) HubToRam(i,width) ' send to ram i -= width ' subtract the width, move up the picture fat.closefile RamDiskPointer += $36 + (width * height) ' increment the ramdiskpointer for next picture RamDiskFiles += 1 ' increment number of entries in the ram diskThen there is the VGA resolution. Kye has a 160x120 display with all the colors and with Floyd-Steinberg dithering it doesn't look too bad. I've done experiments reducing the number of colors to 16 and increasing the pixels and then dithering but there isn't much difference in the end.
I'm just following along with what Rayman did a while back. He is way ahead of me and looking at the post above looks like he will have something ready for you soon which will be just perfect.
Maybe combine Rayman's VGA driver and my .bmp decoder?
Rayman and Drac definately are the ones to help with the VGA and BMP sections. Kye has also done some VGA code that is worth a look.
We have a number of Prop OSes that you can use with an SD card, so no problems there.
This should be doable with a standard VGA configuration, if you aren't choosy about how many colors you want.
Roadster's VGA version of the Ranquest driver can handle 256x192. Can anyone else go higher?
@Ken: Post an example slide for us to work on..
OBC
The portable VGA thing was mentioned by Ahle last year in the new product contest.
Notable that Linus Akesson used Propeller Font in a presentation once.
After chatting a bit over Skype with OBC he suggested I upload a couple of images to see what's possible. It sounds like it's best to keep the text and graphics coupled together as a BMP image, and that mixing the two is probably not all that practical. Might even yield better graphics in the end, too, so I could handle that. I exported these from MacBook Air using Keynote.
Couple more requirements:
- Use standard hardware with no extra chips, so that guests to a presentation could run the same presentation on their PropBOE and microSD card. I need to look closer at Rayman's design to see if there's a compatibility or not.
- Would also be super nice to have two displays: both a VGA output to the projector (which must support these lower resolutions BTW) and a portable display so I can see what's going on.
Could you guys tell me what could be accomplished with the images I've attached to this post? Would they look pretty much the same as they do when viewed on the computer?
Can you imagine how cool it would be to suddenly have a slide that is actually a demo, running on the same Propeller? Wow.
If it's easy enough to do, maybe somebody could post a sample code I could run on the PropBOE. I promise to be independent once I understand how to use the code.
Thanks,
Ken Gracey
(I just got back from Hawaii a couple of days ago)
Sorry. Not possible with standard hardware (ie no extra parts) on a Prop1 to a projector / vga monitor.
Insufficient memory to hold bitmaps that big, and too few colors.
The best you can do with a Prop1 is 160x120 pixels (displayed as 640x480) with 64 colors per pixel.
Ray's DVI shield could do what you want, with enough colors.
I've been working on and off on a 512x384 driver for Morpheus with 256 colors per pixel that would get close, but I think for your application Ray's board is better at this time.
Regards,
Bill
Surely we should be able to get close with a VGA driver...
OBC
Best case, IMHO is a C program that runs off SD card, EEPROM, something, that then displays a sequence of tiled displays. A tool would be required to compose these. Fetch all the tiles, display them, on next slide, take the screen white or black, fetch the other tiles, display them, etc... A very large number of tiles would fit on a SD card.
Meanwhile the local parts stores are still touting 8951s and such for hobbyist. I've long suspected that the retail market is primarily made up of excess inventory from the manufacturing side of the economy. Hon Hai Precision Industries, Ltd (also known as Foxconn) is the biggest retail parts supplier in Taiwan (and they may be the biggest in the world), but they have never heard of Parallax in the Kaohsiung store. If you come back, you might find an excuse to visit thier retail outlet in Taipei.
http://biz.yahoo.com/ic/58/58638.html
Chinese love complexity, so I wouldn't worry too much that your presentation was too much for them. It may have actually been more enticing. It is likely your audience was a mix of extremely savy individuals and students that want to be.
But, trying to do graphics out of VGA with no external chips is going to be dissappointing, I think...
If you'd allow a Flashpoint SQI chip to be plugged into the PropBOE, maybe something reasonably nice could be done.
Or, rigging up a TV output in the PropBOE proto area could improve the BMP quality.
(I have a BMP image viewer for TV posted in this forum that could do a good job).
I have a tiny, $15 adapter from Amazon.com that can convert TV to VGA, if it has to be VGA.
But, if you want VGA native output and no external anything, it's going to be tough.
I think what I would do compose each slide from several 1-bit and 2-bit bitmaps and have as much of the slide blank as possible.
1. A modified version of Kye's 160x120 VGA driver (which is signalled as 640x480), that can put
- a single 160x120 bitmap anywhere within the 640x480 area (filling 1/16 of the screen area), or
- a "double pixel" (320x240) bitmap anywhere in the 640x480 area (filling 1/4 of the screen area, useful for presentations), or
- a "quad pixel" bitmap filling the entire 640x480 screen.
Ie the bitmap/buffer is always the same 160x120, but scaling can be 1x, 2x or 4x, and the X and Y offset of the starting location of that bitmap can be manually set for the 1x or 2x cases. This could also allow for cool side of screen scrolling entry of the bitmap (yeah just like real powerpoint).
2. A "slave" text driver that operates in another cog, and watches the H and V sync of Kye's 640x480 driver, and inserts big White text by overdriving P18..23 high where text is required. The text would ideally have at least 2 sizes (eg 16 or 32 pixels tall at 640x480).
3. For bonus points, a counter to duty cycle (PWM) the blue LSB according to the Y line, so we have a nice blue fade to black render in the background of the white text.
I was going to have a crack at the "slave" video cog when Kuroneko came out with his brilliant 128x64 VGA text driver, where each character can have its own foreground and background color. Combined with half block characters, this allows up to a 128x128 "bitmap" with the bonus that each cell can have half blocks or text character.
But there is still a need for big text, and a picture that doesn't fill the whole screen that can be placed arbitrarily...
Ron Nollet and myself gave a talk late last year at a local university. We plugged a proto board with fresh 9v battery into the lecture hall projector and ran Linus's turbulence demo. Would have loved to run the presentation off it too (may need a bigger battery! ). And there are many other applications if we get this going...
http://www.linusakesson.net/music/elements/presentation.php
It is a presentation tool Linus Akesson used in this
http://www.linusakesson.net/music/elements/index.php
lecture about chip music
Maybe it's time to do bitplanes and tiles with more than one signal COG!
Instead of stuffing the whole byte through the waitvid, restricting it to 4 pixels essentially, continue to use the 2 bits per pixel mode, so 16 pixels can be done. That will get us the 640 pixel clock, and all the colors. Several COGS working together, and synchronized, could then get full color output with a smaller pixel size. This is key to getting picture placement with small pixels. IMHO, doing that makes the most of the 64 colors, because smaller pixels tend to blend in the eye better than larger ones do. Dr_A was exploring this at length, and his early icon work demonstrated how that dynamic worked. At that time, a three COG driver was too expensive for his goals, but not for these.
Doing it this way would also allow the use of the ROM Font for a considerable savings of HUB Ram. I like this idea because the ROM font is very distinctive, in that it's all Parallax, and because it's got a nice set of characters, which can do a lot of basic diagrams.
Break that into 16x16 tiles, so that ROM font text can be placed at 16 pixel precision. Because multiple tile-maps would be needed (3), maybe go with the 32 pixel tiles, just to make the most of the HUB memory.
A slide then would consist of:
Three tilemaps, one for each of the signal COGS. They would each read their tile maps, each contributing their color. One each Red, Blue, Green, with one of them also delivering sync, and acting as the primary reference signal COG.
ROM Font for text.
Tiles!!
For simple graphics, a few tiles could be dedicated to the shapes needed. For product pictures, they would need to be rendered in limited colors. This might be effective on things like circuit boards, which really don't need to be shown in true color to be understood. It won't be all that effective for a face, or some real object.
Store lots of tiles on the SD Card, and fetch them into the HUB as needed.
Picture "area" would be restricted to 1/8 or maybe 1/4 of the screen depending on content. The rest can be mixed text and or simple graphics, a fair amount of it blank.
Ideally, the input to this program would be a bitmap to be placed in the tile-maps, and or a tool to help generate and position the ROM font characters in the tile map. Use a PC for this task.
Some work color reducing the images would have to be done, as well as breaking them into the three slices needed. Red, Green, Blue.
http://www.rayslogic.com/propeller/Programming/6-Bit%20Bitmap%20App/6BitBitmap.htm
Had a bit of trouble getting the cogs perfectly aligned, but still it worked OK.
The obvious solution here is the Solomon SSD1963 -> VGA that Rayman has.
You really want a top-down solution, with good image quality, so it can work on a projector to a large screen.
This also makes image preparation easier, and the wipe speeds should be good enough you may be able to 'dress them up' with a Morph/blend type slide change.
Not sure of the total data-rates, but a quick calc suggests 100-200ms wipes should be possible ?
Rayman will have exact numbers
Rayman has pushed the graphics out to the limit. He says on post #17 which is the same as the Floyd-Steinberg dithering I was using. I was writing it in vb.net but I see Rayman has found Adobe Acrobat can do the same thing. Other graphics programs might be able to as well.
All the cogs are free so some cogs could be doing the graphics while others could be putting the text part together on the fly. So like potatohead says, have a cog (or more if needed) that draws the text using the internal ROM font. So you might need an 80x25 text screen and that will take 2000 bytes. So that decreases the space for tiles a little. But you can be smart with the layout and maybe have some text lines at the top of the screen, and then graphics go in the middle, and then more text under the graphic. Most powerpoint slides work this way, and indeed, usually the text is big, simple and maybe only 3-4 lines with bullet points. So instead of 2000 bytes maybe all the text could fit in a few hundred bytes.
Then the images can all be PC prepared, (leaving a little room for the Prop status line), and new images are just a drag/drop operation, and the slide player is just a memory shuffle ?
Rayman, maybe you could confirm the following for my meager mind:
- It's the BMPs that require all of the RAM, and therefore dictate the image display resolution.
- A tiled approach could make the most effective use of our limited resources by fixing areas for some graphics and an image. This could constrain BMPs to a fixed location, of fixed size, and allow area above and below to be used for text.
- Using the Flashpiont SQI design it probably wouldn't be necessary to separate text and graphics anyway, right?
Would a design approach like this be possible using one of your Flashpoint SQI designs on a PropBOE breadboard? Seems like this video circuitry is built into your display PCBs and not all that portable to a PropBOE without a new breadboard.I'm not in the position to do lots of development on my own, nor am I really all that capable to do so. I have an assortment of other responsibilities around here, so I'm wondering [out loud now] if there is some way to entice our user community to lend me a coding hand.
Thanks,
Ken Gracey
Ken, I think that's what I'm saying...
Text and simple graphics are fine, but
showing a photo without extra memory on native VGA is not terribly rewarding...
Perhaps the SSD1963 isn't the best solution for what Ken wants to do, especially since you can do it with other microcontroller too...
I think a Flashpoint Superquad module can plug into PropBOE though and be a way to show a higher resolution image from low cost, external memory.
I've done it for TV, so maybe it can be done for VGA too...
Maybe I should buy a PropBOE so I can test that out... I still have some robot contest 2nd place award money to burn...
Ray would you share this with us?
OBC
http://www.amazon.com/dp/B00351VWKI (was $15 at the time) that convertes VGA to TV.
But, the box and instructions also talk about going from TV to VGA. But, I think now that they use the same box and instructions for 2 different products.
So, I imagine there is an identical looking thing that does the reverse conversion... But, I don't see it right away...
Ok, I found it:
http://www.amazon.com/dp/B006FS0F92
There are a ton of VGA to TV things that all look the same... Not too many TV to VGA though...
But, I think the cases are all the same...
http://www.amazon.com/Sabrent-TV-VGA2-RCA-S-VIDEO-Converter/dp/B002R35IVY
As I recall there can be some herringbone effects with very fine lines, but you can minimize it by playing with the resolution.
On edit, better price here (I think it's the same unit, with different name):
http://sewelldirect.com/Composite-RCA-to-VGA-Converter.asp
-- Gordon
So the challenge is to fit it all in a standard prop with no external ram - ie the BOE.
160x120 using 1 byte per pixel (6 bits actually) is 19200 bytes. That should leave sufficient space for a text buffer (maybe 1k) and the code (maybe 10k).
Standardising on this format would make designing slides easier too. Each slide has a heading, a picture, and several lines of text.
For the video driver, would it be a hybrid of existing drivers? eg - for 640x480 vga, maybe the first 48 lines you run the text vga driver code in a cog. Then run the graphics driver code, then finally back to text driver code. So it would be a matter of fitting text code and graphics code into one cog. OR - running in separate cogs and doing handovers in the back and front porch.
Rayman, confirm your shipping address to me via kgracey@parallax.com and we'll take care of your PropBOE hardware shortage [without any obligation on your behalf, of course]. Nobody should be in that position, especially a UPE roadie as yourself who contributes so much to our customers.
Ken Gracey
Way to go Rayman! Santa is arriving early this year.
Just did a quick test of what our 6-bit bitmap app would do for one of Ken's pictures:
This is what it would look like on regular Prop VGA. I guess it's not horrible.
This takes up 17kB. So, there's still a little room left for some text. If that were done as 1-bit bitmap, maybe enough room...
Is there enough room to shoehorn some SD drivers and turn it into a multi-file reader?
OBC