How do you convert samples to Amiga?

[Trapscallion]

Would like to hear how everyone process`s their sounds and then bumps them down onto the Amiga.

So far i have edited some breaks in Logic then either bounce them down to 16 bit aiff or wav. Then i transfer them via pcmia onto my A1200 then in Octamed SS i load them up, have to change the type from 16bit to Sample, then up the sample 2 octaves and then double the volume.

Is there a better way to get the samples sounding better on the Amiga ?

[jimbob]

some good tips here -

http://eab.abime.net/showthread.php?t=66065

[adrdesign]

There's a thread here and so many others
http://eab.abime.net/showthread.php?p=842458

I usually load the 16 bit sample in Octamed SS, apply max volume (uncheck clip), apply 133% volume, upsample _one time_ (this sets the rate at 22050), uncheck the stereo attribute and check sample (in instrument type) to convert to 8 bit.

This method is fast, but it clips the sample and the resample quality is rubbish, so I did a batch conversion script for use with SOX that keeps the samples crystal clear, with some compression and the best antialiasing.

You only need to extract it in X and create a subfolder in X where you put all the samples (they can be in drawers too).

PS: Sorry if my english is poor... :P

[Amiga1992]

Cool shit ADR!!!

[Trapscallion]

Thanks for the help guys ^_^

[adrdesign]

Yeah! Im working on converting few sample libraries Akira, check for future posts. Im amazed how the amiga can sound nowadays...

[Apollo]

I did use SoX in the past for downsampling and normalising. Its command line only, so for batch processing some scripting knowledge could be handy.

Before I switched to MacOS for audio I did use 'Wavelab' and its batch processing facility.

[h0ffman]

Can you use Sox to batch convert pre-processed 8-bit wavs to iff?

[h0ffman]

i suppose that script from ADR is knida, it just need to take out the resampling?

[pandy71]

I would advise to use PC and SoX (modern not old) for resampling also separate tool for noise shaping ( http://www.hydrogenaudio.org/forums/...dpost&p=680736 )

Why such complications? With noise shaping dynamics of the 16 - 20 bit system can be achieved on Amiga (Paula - 8bit thus 4 channels) especially when sampling rates are over 40 - 47kHz but decent 13 - 15 bit dynamic is also available for sampling frequency 29073Hz.

PS
for variable rate playout TPDF dither is recommended (SoX option -S)

[adrdesign]

Originally Posted by h0ffman

Can you use Sox to batch convert pre-processed 8-bit wavs to iff?

Yeah, it's very simple:

for iff-8svx
sox (infile).wav (outfile).8svx <- note that 8svx implies 8 bit, so if the source sample is in 16bit it will be converted

for amiga raw

sox (infile).wav (outfile).sb <- this converts to signed integer pcm in 8 bit.

if you want to use with the script, change line 25 to

list >t:batch all sub #?.wav lformat="echo *"Converting %n -> %p%m.8svx*"*nsox12 *"%p%n*" *"CONVERTED:%p%m.8svx*""

(or .sb if u want raw)

@hoffman: Thanks for your work, I have learned a lot from your recent songs. You're a genius.

It Would be very difficult to make a proper gui to get all the sox parameters out? I'm thinking some kind of Soundforge for amiga, we only need the graphical interface. Some ideas, where to start?

[h0ffman]

Cheers ADR

just this alone is going to save me SO much time. I've been converting them to IFF using SoundBox! Open Wav, click 6 times to IFF, save, repeat for 10 samples!!

Might put together a drag n drop app so you can just dump all the files onto a window to create the amiga versions of them.

[h0ffman]

I've just put together a little app for converting samples using Sox under windows called Wav2Amiga.

Simply open the app, drag the wav files onto the window and hit DO IT!

It will then dump a .8SVX version in the same folder.

If people want this expanding, like selecting a destination folder or something, let me know.

[adrdesign]

Thank you!! Let's try

[adrdesign]

It works faster than light! How does it works? It converts only or does some upsampling? It would be nice to set the output rate, type of resampling and some compander options (with different kinds of curves), btw it is very useful as this. Thanks again!

[h0ffman]

It's just a basic drag n drop which then runs the sox.exe in the folder against each file.

When I get some time, I'll make it so you can have multiple configs which you can add yourself, basically just allowing you to configure the arguments for the command.

Might even add a default set for each note in the ProTracker range

Will be a little while before I get back on it though as Akira has just landed me all the final bits for PT-1210!

[Trapscallion]

Hey all sorry for the long awaited reply.

I have tried to open both of the WAV2AMIGA and Wav2Amiga zips posted above by hoffman and ADR and neither programs will open via terminal on mac or command prompt on Windows.

Any ideas?

[Trapscallion]

here is a screen shot of the error message i am getting trying to open the Wav2Amiga file

[adrdesign]

The script only works for amiga. I've been using sox in windows and it has lots of newer features.

---------- COPY AFTER THIS LINE ---------
cd %~dp0
mkdir converted
FOR /f "tokens=*" %%A IN ('dir /s /b *.wav') DO sox --multi-threaded --buffer 131072 -S -V -D -r 22050 "%%A" ".\converted\%%~nxA.sb" gain -n -0.348878 dither -f shibata -p 8 stats -b 8
pause
------------- COPY BEFORE THIS LINE -----------


paste to notepad and save as convert.bat
use this script within sox and zlib.dll in the same folder. Add some samples in (subfolders allowed). It will maximize and convert to 8 bit 22050 sound with noise shaping. Note you can replace the "22050" value after -r in the script with the sample rate you want.

hope it helps

[pandy71]

Originally Posted by adrdesign

The script only works for amiga. I've been using sox in windows and it has lots of newer features.

It will maximize and convert to 8 bit 22050 sound with noise shaping. Note you can replace the "22050" value after -r in the script with the sample rate you want.

hope it helps

With one remark - Shibata filter (noiseshaping based on psychoacoustics human ear characteristic) will work only in limited frequency ranges, some of SoX filters (noiseshaping) are tuned only to 44100 frequency.
general point of view from SoX developers perspective is that for low sampling rates (subNyquist from human point of view i.e. samplerate less than 32kHz - 40kHz there is no objective optimal noishaping method).
Usually in such cases (were samplerate is lower than allowed by noiseshaping filter) SoX fall to HighPass TPDF dither which seem to be good balance between perceived noise level and quantization errors however personally (subjective) i still prefer noiseshaped sound.
To use Shibata in such cases (for example for 28 - 29kHz samplerate which is reasonable close to 32kHz) i would recommend or pipe or use temporary file and do conversion based on principle - resample to required samplerate without dither/noiseshaping then change value for sample rate to higher (but lowest acceptable by your noiseshape filter) do requantization then write file with resampled sample rate. so for example source is 48kHz 16 bit, do resampling to 28604Hz or to highest supported (for 15kHz HSync modes) 29073Hz at 24 bit at least (natively used by SoX) then change signaled sample rate (but do not resample) to for example 30400Hz (seem to be lowest samplerate supported by Shibata) then do requantization (perform conversion from 24 bits to 8) while keeping 30400Hz sample rate, such temporary file need to be writen to 8svx (or any required file format) with correct sample rate information (i.e. for example 28604 or 29073Hz).
Amiga with video mode where HSync is higher than 15kHz can use higher audio sample rates (up to Sysclk/512 when level control is enabled or even higher when audio level control is turned off).
I dont know how Amiga system behave when HSync rate will be lower than 15KHz.