General Audio Waveform in DICOM

[Valerio Bamberga]

Hi!

I'm trying to add a MP3 file to a DICOM.

This should be possible according to the DICOM specifications ("General Audio Waveform" in part 3).

Has anyone ever tried this using Fellow Oak?

I don't see a DicomTransferSyntax about Mp3 or Wav and I'm not sure how to proceed.

Thanks for any idea :)

[Reinhard Gruber]

Hi,

there is no TransferSyntax for Wav. Use some default ImplicitLittleEndian or ExplicitLittleEndian transfersyntax.

The magic of waveform is not in Transfersyntax, but in AbstractSyntax. So instead of "CTImageStorage" or "MRImageStorage" or so you then will have "GeneralAudioWaveformStorage" (see here: https://dicom.nema.org/dicom/2013/output/chtml/part04/sect_i.4.html).

You then will NOT store data into the very well known PixelData tag (because that is reserved for pixel data). You will have to add a Waveform Sequence (5400,0100), there you have to add all your metadata like sampling requency, channel number, bits stored, ... And of course within that Sequence you store your actual Data in WaveformData (5400,1010).

This all is described here: https://dicom.nema.org/dicom/2013/output/chtml/part03/sect_C.10.html#sect_C.10.9

of course you can use fo-dicom to build your dataset. But I have never done something else than pixeldata. So sorry, I cannot give you any sample how to deal with waveform.

But if you manage to do so, some contribution would be very great!