Dithering in VS 3 plugin "External Insert"

[SoundKlang]

For a few days I am thinking about the necessity to apply dithering when using a track insert in the Pyramix mixer. I am not sure about the following. If there is something wrong in my comprehension of this topic, please say so.

My first question is wether dithering of the Pyramix's mixer strip signal with it's internal bitdepth to the bitdepth of the used DA converter is taking place implicitly when using the VS 3 plugin "External Insert"?

If the answer is "yes", my next questions would be, does this dithering happen after the Send volume control of the VS 3 plugin "External Insert" and is it possible to configure this dithering?

If the answer is "no", I wonder if such a dithering is necessary? Even though an external analog chain will add it's own noise to the signal, still the signal coming from the DA to feed the analog chain would be the result of an undithered reduction in bitdepth and thus would show a lesser resolution than it would seemingly have if it was dithered. Is that true or is there something missing in my comprehension?

Searching for a way to add dither under these circumstances, I found that it is possible to use a third party plugin for dithering. In the signal flow a dithering plugin must be placed in front of the "External Insert" plugin to ensure adding the dither noise happens as the last processing stage before the DA conversion. This requires to leave the Send volume control of the following "External Insert" plugin at 0 db, because after the dithering no more altering of the digital signal may take place. If in addition the I/O interface's mixer does not apply dithering after it's own level fader for the digital output signal, then also this I/O interface's level fader would need to remain at 0 dB. There would be no possibility to control the send level to the DA, which may be useful for level-matching between the DA's output and following analog outboard gear or when driving it.

If this is correct so far, then another possibility to achieve a control over the send-level of the "External Insert" could be to place the third party dithering plugin and the "External Insert" plugin post fader in the Pyramix mixer strip. Is that possible with the Pyramix mixer?

Otherwise, would it be possible to include a dithering option in the VS 3 plugin "External Insert"?


(@Merging: Regarding the previous emails, please excuse if this looks like insisting on the topic. After a few more days thinking about this, I continue to wonder if this could be something and if maybe my emails were unclear or the wrong channel to place this question.)

[fl]

Even if all the conversions you describe are done with simple truncation, wouldn't all of the resulting quantization error be swamped by the dither applied at the final mix stage?

[allemano bernard@neuf.fr]

Hi,

Normally, it is not necessary to add dithering. It is the work of CD manufacturer. It is he who decides if this is necessary or not. This is usually not the job of a recording studio. Indeed, adding dithering, you may alter the quality of audio files.

So, don't worry, be happy

Best regards,

bernard

[Graemme]

Not Exactly...

In general, dithering should be used anytime a reduction in wordlength is required.

Exceptions?

In the case of the 32bit floating point to 24 bit fixed point conversion required for the insert plug-in: If your signal stays below 0dBfs, I doubt you'll have *any* issues at all. This conversion does not have the same requirements as e.g., a 24 to 16 bit conversion in the fixed point domain.

Graemme

Hi,

Normally, it is not necessary to add dithering. It is the work of CD manufacturer. It is he who decides if this is necessary or not. This is usually not the job of a recording studio. Indeed, adding dithering, you may alter the quality of audio files.

So, don't worry, be happy

Best regards,

bernard

[SoundKlang]

Thank you for the replies. It is true, this is not really an issue. It is not "wrong" without dithering, only I believe to hear a subtle difference in depth, localization and tonality with dithering.

I found a workaround: It is possible to send the signal to an Aux Send Bus (In-Place Panning activated) and set the Outputs of the Aux Send Bus to the DA. I was pleased with result with the Aux Send's Dither option activated (MT-R, Triangular, Acoustic Noise Shaping).

[Graemme]

Yep, that's what I do...except with flat, TPDF dither at 32fp to 24i...

I do hear a subtle difference using noised-shaped dither in this context but much less (if any) of a change using plain-old, flat tpdf dither.

To each, their own :->

gb

Thank you for the replies. It is true, this is not really an issue. It is not "wrong" without dithering, only I believe to hear a subtle difference in depth, localization and tonality with dithering.

I found a workaround: It is possible to send the signal to an Aux Send Bus (In-Place Panning activated) and set the Outputs of the Aux Send Bus to the DA. I was pleased with result with the Aux Send's Dither option activated (MT-R, Triangular, Acoustic Noise Shaping).