Please stop show 'Bit depth' info in lossy audio formats.
Only lossless audio formats (Flac, TrueHD, DTS-MA, ...) can have 'Bit depth'.
Lossy formats (AC3, standard DTS, AAC, ...) can't show this info because you are confusing people.
this is not so easy, i.e. lossy compression does not mean that there is no bit depth. This is classic to know the bit depth for e.g. lossy video formats.
I agree it may be discussed for audio.
Actually, there is already no bit depth for MPEG Audio and AAC.
I have more doubts about DTS and AC-3: DTS: there is a field "Quantization resolution of source PCM samples" (16/20/24) so I think it is good to keep it. AC-3: there is no obvious information about bit depth, but I never saw encoder or decoder accepting something different than 16-bit. Do you have an AC-3 file with 24 bit content in output with last bits useful as a prrof that AC-3 can do more than 16-bit?
Dolby E: I see bit depth info (16/20/24) in docs, so I guess it is used same as output bitdepth.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Is really easy. Only lossless formats can guaranty a bit depth precission.
When you ask me: "Do you have an AC-3 file with 24 bit content in output with last bits useful as a prrof that AC-3 can do more than 16-bit?" seems than you don't understand the lossy audio compression.
The samples, in time domain, are converted in frequency domain samples (using float maths), and only the more important components than fit in the bitrate requested are preserved and stored.
The decoder can't even guaranty recover 16 bits from source, sometimes yes (if all the frequency domain samples fit in the bitrate), sometimes not (if some components must be sacrified). For that include a Bit depth (16, 24 or any) only confuse the users than think recover this precission.
The DTS header have a field to store the bit depth of the source. But is useless because the precission can't be recovered, and because the info can be wrong (some encoders puts always 24 no mather the bit depth of the source).
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Only lossless formats can guaranty a bit depth precission
This is the definition of "lossy", this is not the discussion: in lossy compression, bit depth information is obviously the maximum bit depth.
The discussion is about knowing if it is possible to have information about more than x bits. it is a bit more obvious with video e.g. HEVC has an explicit 10-bit profile which does not mean that this is lossless and that the bit depth precission is guaranted.
consider that "Compression_Mode" == "Lossy" means that Bit depth is the "maximum" possible bit depth.
So let focus on "maximum" bit depth information:
AC-3: I checked the official decoder from Dolby (I would have prefered to check the encoder, but I don't have it), looks like it can decode up to 32-bit (if 64-bit output, precision is said to be maximum 32-bit) and there is no information about bit depth information in the stream, so I am OK to remove bit depth information from MediaInfo. Patch.
DTS: I have no access to official encoder or decoder from DTS, so I can not look at it. For the moment, due to the presence of the "Quantization resolution of source PCM samples" field in the spec, it looks like there is some limitations to the possibilities with the encoder. No change.
The DTS header have a field to store the bit depth of the source. But is useless because the precission can't be recovered,
Up to the value in the field, as for any lossy format.
and because the info can be wrong (some encoders puts always 24 no mather the bit depth of the source).
MediaInfo shows information from the file, no more. Handling bugs from encoders is out of the scope of MediaInfo (I accept to remove such values only if most major decoders have a problem). Please report the bug to the encoder maintainer.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
this is not so easy, i.e. lossy compression does not mean that there is no bit depth. This is classic to know the bit depth for e.g. lossy video formats.
I agree it may be discussed for audio.
Actually, there is already no bit depth for MPEG Audio and AAC.
I have more doubts about DTS and AC-3:
DTS: there is a field "Quantization resolution of source PCM samples" (16/20/24) so I think it is good to keep it.
AC-3: there is no obvious information about bit depth, but I never saw encoder or decoder accepting something different than 16-bit. Do you have an AC-3 file with 24 bit content in output with last bits useful as a prrof that AC-3 can do more than 16-bit?
Dolby E: I see bit depth info (16/20/24) in docs, so I guess it is used same as output bitdepth.
Is really easy. Only lossless formats can guaranty a bit depth precission.
When you ask me: "Do you have an AC-3 file with 24 bit content in output with last bits useful as a prrof that AC-3 can do more than 16-bit?" seems than you don't understand the lossy audio compression.
The samples, in time domain, are converted in frequency domain samples (using float maths), and only the more important components than fit in the bitrate requested are preserved and stored.
The decoder can't even guaranty recover 16 bits from source, sometimes yes (if all the frequency domain samples fit in the bitrate), sometimes not (if some components must be sacrified). For that include a Bit depth (16, 24 or any) only confuse the users than think recover this precission.
The DTS header have a field to store the bit depth of the source. But is useless because the precission can't be recovered, and because the info can be wrong (some encoders puts always 24 no mather the bit depth of the source).
This is the definition of "lossy", this is not the discussion: in lossy compression, bit depth information is obviously the maximum bit depth.
The discussion is about knowing if it is possible to have information about more than x bits. it is a bit more obvious with video e.g. HEVC has an explicit 10-bit profile which does not mean that this is lossless and that the bit depth precission is guaranted.
consider that "Compression_Mode" == "Lossy" means that Bit depth is the "maximum" possible bit depth.
So let focus on "maximum" bit depth information:
AC-3: I checked the official decoder from Dolby (I would have prefered to check the encoder, but I don't have it), looks like it can decode up to 32-bit (if 64-bit output, precision is said to be maximum 32-bit) and there is no information about bit depth information in the stream, so I am OK to remove bit depth information from MediaInfo. Patch.
DTS: I have no access to official encoder or decoder from DTS, so I can not look at it. For the moment, due to the presence of the "Quantization resolution of source PCM samples" field in the spec, it looks like there is some limitations to the possibilities with the encoder. No change.
Up to the value in the field, as for any lossy format.
MediaInfo shows information from the file, no more. Handling bugs from encoders is out of the scope of MediaInfo (I accept to remove such values only if most major decoders have a problem). Please report the bug to the encoder maintainer.