Tag - Tutorials

Using the Douglas Little 3D Routines

I thought it would be a good idea to include this additional information I posted up to the STOS Coders FB page. It’s to help those who want to make new models for DML’s 3D extension. A JPG screenshot has been included, showing a possible workflow in action.

I’d advise not using The Pixel Twin’s Utility Disk version that’s floating around in Floppyshop PDL (ie. UTL-3242.zip). It seems to be an earlier 1991 version and somewhat incomplete, so please use the later 1992 version included within this archive (which Doug posted to Atari Forum in 2014). You can modify the code within DML’s newer example file, STOS3D3.BAS much more easily to view/use your own models now too.

So, I have come up with a workflow that works well so far. I’m keeping my fingers crossed that we can eventually do more than just view a newly created model!

1) Model a low poly model in Blender 3D (https://www.blender.org/)

2) Export as a Wavefront *.OBJ file (you can probably try other formats too, I just used this one because I figured it would work fine and it did)

3) Load the *.OBJ file into 3D Object Converter v7.0 (http://3doc.i3dconverter.com/index.html) (799 3D object file formats supported!)

4) Save as *.3D2 format (Cyber Studio CAD-3D)

5) Load the *.3D2 model file into DML’s “CADCONV.PRG”

6) See DML’s own original doc for further info and get ready to be patient with all of this!
Use the arrow keys to navigate around the model, Z & X keys to cycle through each triangle. Clean up each surface by deleting the first triangle (with “DEL FACE”), then expand the 2nd triangle to make all surfaces appear like your original Blender 3D model. You will have to click on “EXP FACE” a couple of times, sometimes more. Then “EXP DONE” when finished. Recolour all surfaces to your liking. Only use the first 16 colours. If you mess up, reload the *.3D2 file and start again.

7) Click on the floppy disk icon “O” with the arrow pointing towards it, ie. 4th one in. Export a new object file set (*.X, *.Y, *.Z, *.OBJ) to a blank floppy disk image in drive A (not your HDD partition, as I don’t think it can save to HDD). The doc says it saves off a *.S file somewhere, it doesn’t, it’s an *.OBJ file (so remember to keep your original Wavefront *.OBJ files within a seperate folder from your output here, to avoid overwriting!). You just type in the name you want for the whole file set, but no extension required as it saves out 4 new files at once.

8) Modify the example file listing (STOS3D3.BAS), so you can load in your new model and view it. In line 10 you can modify the palette, I think this may need experimenting with to get the colours to match up correctly.

Then on line 46, you can change each filename similar to what I have done here:

46 I(X_LIST)=start(6) : I(Y_LIST)=start(7) : I(Z_LIST)=start(8) : I(S_LIST)=start(9) : I(FIL_PATS)=start(10) : bload “objects\cybercar.x”,start(6) : bload “objects\cybercar.y”,start(7) : bload “objects\cybercar.z”,start(8) : bload “objects\cybercar.obj”,start(9)

It should now display your new model, at least I hope so. Elon Musk won’t be hiring me for my new car design, that’s for sure!

Mike K

ps. Many thanks to Neil Halliday for tracking down The Pixel Twin’s Utility Disk! We were able to include the docs from that disk, with the later version posted to AF which I found later.

DOWNLOAD

For reference, here is an image of the workflow

Converting Audio for the STE

By Neil Halliday

As we all know, the STE has far superior sound capabilities than the standard ST, and for many years this has been put to good use in games and demos. However, audio processing technology and software has massively changed since the day of 8 bit sampling, and audio playback quality is now pretty much as good as it is going to get. The question is, can we put the new world of audio processing to use with our beloved Atari STE?  Well, the answer is yes, we can.  We can use our modern devices to create and process our required audio, and then resample it for use on the STE. This post will talk you through the process, and show you the settings you need to use in order to do it.

Audacity

One of my favourite audio programs on the PC (also available on the MAC) is Audacity. Why is it my favourite? Well, it’s free for a start! Not to mention that it also has an abundance of plugins that can be used to further extend it’s features.  I’m quite familiar with Audacity and as a result I can get the results I want quite quickly.  When it comes to sound and music composition, I turn to other applications, but the final results always go through Audacity to convert them into the format that will work on the STE.

If you’ve not got Audacity, you can download it here.

So, what do I need to do?

Well, it’s pretty simple really to convert your audio, but here are the steps to follow.

Load your audio into Audacity

Firstly, you need to get your audio into Audacity. My preferred way is to just open Audacity and drag my audio file into it. Whatever way you do it, you’ll basically end up with your audio in the application and ready to start processing.

Make sure you set the correct Project Rate!

By default, Audacity will choose standard CD quality as it’s project rate. Notice how it says 44100 in the screen shot above. This is the frequency of the audio, which is measured in Hertz (Hz). This represents the number of samples of audio taken each second. When working with audio files, this can also be represented in kilohertz (kHz), so 44100 Hz can also be represented as 44.10 kHz.

The Atari STE’s DMA chip has 8 different modes that it can operate in. These are 6.258 kHz, 12.517 kHz, 25.033 kHz and 50.066 kHz in both monophonic or stereophonic playback. So, the first thing we need to do is set Audacity to recognise the frequency that we want to work in. We do this by setting the Project Rate at the bottom left hand side of the Audacity window. We can simply type over the value that is there by clicking in the box. Remember that Audacity works in Hz, so we need to convert our kilohertz into hertz. To do that, simply multiply the kHz rates listed above by 1000.

Time to resample

Now we need to resample our audio file so that it is in the correct rate to be replayed. This is done by using the Tracks pull down menu at the top of the Audacity window, and selecting Resample.

You will notice that Audacity knows what the project rate should be and therefore recommends to resample at that rate. If we want a different rate, we can simply over-type.

Click OK for Audacity to do it’s stuff.

Time to export the file

This is the most important step, and the one that has to be correct; otherwise the resulting sample will be inaudible when played through the STE hardware.  Go to the File menu at the top of the window and select Export. Then, from the sub-menu that appears, choose Export Audio. It is very important to use this option as it means we can change the export settings to be compatible with the STE.

You will now be presented with the standard file selector, but with some configurable options. Make sure you have the following set:

Save as type = Other Uncompressed Files

Header = RAW (Header-less)

Encoding = Signed 8-bit PCM

Click Save to save your converted file.

Well done!

Congratulations, you have now converted the audio to a format that will work with the STE. The file can be loaded into a standard STOS bank, and be replayed using the “e play” function from my GBP extension. You will notice that because of processing the audio via a PC/MAC that the result is much better than simply doing it via the STE. I much prefer doing it this way, but would be interested to hear how you get your digital samples into the STE.

Have fun!