How do you convert samples to Amiga?

[SKOLMAN_MWS]

best hq resampler http://cd.textfiles.com/amigaplus/so...s/AudioConvert

in PAL using standard 24000 Hz

sample audio in adp4 16bit 24kHz for old songplayer
http://skolman-mws.w.interia.pl/amiga/adp4/

[h0ffman]

Originally Posted by Trapscallion

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

You need to install .NET FrameWork v4

http://www.microsoft.com/en-gb/downl....aspx?id=17851

[pandy71]

Originally Posted by SKOLMAN_MWS

best hq resampler http://cd.textfiles.com/amigaplus/so...s/AudioConvert

Doubt on this (http://en.wikipedia.org/wiki/Sample_...ate_conversion - SRC is tricky and computationally intensive task) but feel free to prove it by method similar to: http://src.infinitewave.ca/

Originally Posted by SKOLMAN_MWS

in PAL using standard 24000 Hz

Are you referring to which 24000Hz? (see list bellow)

Period Sampling frequency (In Hz for PAL)

142 24978.133803
143 24803.461538
144 24631.215278
145 24461.344828
146 24293.801370
147 24128.537415
148 23965.506757

[SKOLMAN_MWS]

You do not understand. Most of converters downsampling/upsampling give the poor quality of the output samples.

Of course, if someone is deaf has weak hardware HiFi, will not hear the difference.

[pandy71]

Originally Posted by SKOLMAN_MWS

You do not understand. Most of converters downsampling/upsampling give the poor quality of the output samples.

Of course, if someone is deaf has weak hardware HiFi, will not hear the difference.

I give you link to objective comparison for different SRC, some of them use very complex methods (for example long FIR filters) to create very high quality SRC (both in objective and subjective area). You can't pretend that you hear something bellow -170dBFS - arguing that someone is deaf looks like Godwin's law http://en.wikipedia.org/wiki/Godwin%27s_law
SoX is well proven and appreciated by audiophiles with not only ears but also with expert knowledge.
So i understand it quite well and this was my point - if you referring to something as "best hq resampler" then i nicely ask about proof for such claim.

I am struggling to get this working. Opened up the script in a drawer on the Amiga, it told me what to do. Sox12 and WAV2AMIGA.Script are in a drawer with the .wav file, Sox12 wont execute and the script just opens as a text file. Tried then creating a drawer and putting the WAV sample inside by itself and clicked both SOx and Script in the Drawer above, still no luck!

Is there an easier way to do this on a Mac? Like a batch convert to the correct settings for the Amiga then just transfer them via PCMIA and then just start using them in Octamed?

[alkis]

Originally Posted by Hagotae

Is there an easier way to do this on a Mac? Like a batch convert to the correct settings for the Amiga then just transfer them via PCMIA and then just start using them in Octamed?

Recently, doing an Angband (cross)recompile I came across the sound files being mp3. This is what I used to convert them all.

for f in *.mp3 ; do sox -v `sox "$f" -n stat -v 2>&1` "$f" -D -r16574 -b8 `basename "$f" .mp3`.8svx remix 1; done

The nested sox call (inside the backticks) is to find the "scale" factor that would allow me to turn the volume up without clipping (like normalise), you may ommit it if you don't want to make samples louder.

for f in *.mp3 ; do sox "$f" -D -r16574 -b8 `basename "$f" .mp3`.8svx remix 1; done

I know you said Mac and this is for linux, but this is in hope it helps.

I am still a little unsure though how i use the converter, so i have to put both sox and wav2amiga.script files into the same drawer of audio that i want to convert and double click the script file, which just opens it up as a document, do i have to put these codes in there? Opening the SOX file on the Amiga just says it was unable to find files

[Daedalus]

It may or may not be suitable, but I've used SampleManager in the past for converting samples for use on the Amiga and it has done a good job. It's an Amiga program and available on Aminet: http://aminet.net/package/mus/edit/Samplemanager and is written by the same guy who made HD-Rec.

[pandy71]

Guys - i don't understand why you insist to use Amiga for sample rate conversion - on modern PC (even on Atom CPU) this will be faster and also probably significantly higher quality.

@SET PATH=%PATH%;Disk:\path_to\sox

@rem PAL=28375160
@rem NTSC=28636363

@SET PERIOD=124

@SET CLK=28375160
@SET /a AFREQ=((%CLK%+8)/(8*%PERIOD%)) 
@SET /a APREC=((%CLK%+8)/(32*%PERIOD%))

@ECHO Samplerate=%AFREQ%
@ECHO Precompensation=%APREC%

@md 8SVX
@for %%a in ("*.wav") do "sox.exe" --multi-threaded --buffer 524288 -S -V -D "%%a" "8SVX\%%~na.8svx" remix - gain -n -6.0206 treble 6.0206 %APREC% rate -v -s -I %AFREQ% gain -n -0.034264 dither -S -p 8 stats -b 8 stat 
@for %%a in ("*.mp3") do "sox.exe" --multi-threaded --buffer 524288 -S -V -D "%%a" "8SVX\%%~na.8svx" remix - gain -n -6.0206 treble 6.0206 %APREC% rate -v -s -I %AFREQ% gain -n -0.034264 dither -S -p 8 stats -b 8 stat 
@for %%a in ("*.flac") do "sox.exe" --multi-threaded --buffer 524288 -S -V -D "%%a" "8SVX\%%~na.8svx" remix - gain -n -6.0206 treble 6.0206 %APREC% rate -v -s -I %AFREQ% gain -n -0.034264 dither -S -p 8 stats -b 8 stat 

@ECHO Done!!!
@pause

Copy above one to dos batch (text file with extension 'bat' or 'cmd'), run in folder where wav, mp3, flac - Amiga period (in this case maximum legal value i.e. 124 ) is provided inside script so audio will be perfectly resampled to match Amiga hardware expectations.
Feel free to modify (this is mono version - all input formats are downmixed to mono, side to this i apply precompensation (gentle highpass - treble filter to deal with filter and DAC rolloff, dither is applied - HP TPDF - can be changed to any supported by SoX or disabled, samples are normalized to maximum possible by Amiga hardware - this is not optimal but... well it will improve SNR).
Converted files are stored in 8SVX folder. SoX accept limited number of formats - if this is necessary maybe pipe trough ffmpeg can be added (as ffmpeg is not capable to write 8SVX/IFF).

[Daedalus]

Perhaps because not everyone has the will or skill to come up with a batch file like that? (It is nicely done by the way - thanks!) I know I lack the first and the second would at least slow me down greatly. Personally I use Audacity under Linux to resample to different bitrates/resolutions and then SampleManager to do the straight format conversion with no reprocessing, and I'm happy enough with the results that any difference in quality that I might be missing is offset by the amount of time I saved by not messing around with batch files or bash scripts on a PC.

Yeah i have no idea about scripting... I have started using Sony's convrt to batch process audio with ozone 6 plugins down to 16 bit Wav 22.050k ( No IFF choice). It is just the next step of getting the samples prepped for Octamed SS was the issue i was having.

[Thorham]

To pandy71:

That script is nice, but the dithering makes everything sound very noisy. People who want nice 8bit samples for use in Amiga trackers aren't going to like that.

[pandy71]

Originally Posted by Thorham

People who want nice 8bit samples for use in Amiga trackers aren't going to like that.

You can modify script - removing dither is simple - just delete:
"dither -S -p 8 " and dither is no longer used (but beware that for regular listening dither is considered as absolute must to remove/decorelate quantization errors).
This is another example for what i try to say - Amiga is to weak to support correct signal processing (both offline and online) - normally dither shall be added at the end of processing chain where signal is requantized after mixing which is usually performed with sufficient precision (24 - 32 bits at least).

Btw not sure about Amiga audio output quality but i would expect that dither should be not perceived - not audible on a real Amiga HW if reconstruction filter slope and own system noise floor are present - it can be audible on PC equipped with 16 bit audio card (so trough Amiga emulation or by using 8 bit samples on PC tracker).

Btw 1 try to listen audio trough lowpass filter with characteristic similar to Amiga filter and apply roll-of related to sin(x)/x - also added precompensation may help - just adjust graphic equalizer knob to suppress high frequencies by few dB.
I can add Dynamic Audio Normalizer (gentle dynamics compressor - AFAIK LordMulder added this to SoX and not so long ago to ffmpeg - this can close processing chain and squeeze Amiga sound HW to normal maximum - there is always option for Copper driven audio where aggressive noise shaping can be applied and with high sample rate more than 16 bit perceived audio quality relatively easily achieved).

[Thorham]

Originally Posted by pandy71

but beware that for regular listening dither is considered as absolute must to remove/decorelate quantization errors

At the cost of considerable white noise. For listening to CD style audio, 14bit 28khz stereo is best.

Originally Posted by pandy71

Btw not sure about Amiga audio output quality but i would expect that dither should be not perceived

It's very audible on my A1200.

Basically, for samples that are to be used in modules, get the loudest, good sounding samples you can find, and convert those to 8bit at an appropriate sample rate WITHOUT dithering.

[Amiga1992]

Originally Posted by pandy71

Btw 1 try to listen audio trough lowpass filter with characteristic similar to Amiga filter and apply roll-of related to sin(x)/x

Nobody in its right mind would turn that awful filter on to make tracks.

[pandy71]

Originally Posted by Thorham

At the cost of considerable white noise. For listening to CD style audio, 14bit 28khz stereo is best.

Nope if you use noise shaped dither (8 bit system with sufficient oversampling - 4 times 44.1kHz is capable to be comparable to SACD)- but Amiga is to slow (even 68060) to do correctly things like this in real time in background.

Originally Posted by Thorham

It's very audible on my A1200.

Basically, for samples that are to be used in modules, get the loudest, good sounding samples you can find, and convert those to 8bit at an appropriate sample rate WITHOUT dithering.

Original filter? (i mean reconstruction - not one activated trough CIA).

Originally Posted by Akira

Nobody in its right mind would turn that awful filter on to make tracks.

Amiga have 2 filters, one is fixed and used to reconstruct 56kHz sampled audio (unless full level ON i.e. PWM volume control disabled).

[Amiga1992]

Originally Posted by pandy71

Amiga have 2 filters, one is fixed and used to reconstruct 56kHz sampled audio (unless full level ON i.e. PWM volume control disabled).

Not mine, I destroyed two filters on purpose on my A600s to get cleaner audio

[ReadOnlyCat]

Originally Posted by pandy71

(but beware that for regular listening dither is considered as absolute must to remove/decorelate quantization errors).

What do you mean by "regular listening"?
I know only one way of listening but I'm not sure if it is the "regular" one you mention.

Originally Posted by pandy71

there is always option for Copper driven audio where aggressive noise shaping can be applied and with high sample rate more than 16 bit perceived audio quality relatively easily achieved).

I have planned to test this (Copper audio) since I read that fascinating discussion about Paula's Pulse Width Modulation about six months ago. And I must admit I am a bit surprised no one has tried yet if it could be used mid scanline to output at 56kHz. The blitter is fast enough to fill the copperlist in real time so there really are no technical obstacles to experimenting it.
I even wonder if it could be used as a form of temporal multiplexing to allow 8 28kHz channels? I plan to give it a try eventually but I have more important projects for now.

Originally Posted by pandy71

Nope if you use noise shaped dither (8 bit system with sufficient oversampling - 4 times 44.1kHz is capable to be comparable to SACD)- but Amiga is to slow (even 68060) to do correctly things like this in real time in background.

Which is why 14 bits stereo at 28kHz is the current practical best on a real Amiga.
Unless copper audio can beat it of course!

(Apologies for the off-topic.)

[Thorham]

Originally Posted by ReadOnlyCat

Which is why 14 bits stereo at 28kHz is the current practical best on a real Amiga.

Indeed. Sounds quite good, actually, if down sampled with SOX.