Topic: strange flac size issue ?? now wavelab montage issue (or user error!)

[Gordon]

never ran into this before.  I was about to seed my 24 bit matrix of the peruvian pedicurists from loki and noticed something strange.


all source were recorded at 24/48.  the aud was done on my tascam p2, the sbd on daves R4.  everything was flaced with traders little helper at level 8.



the 24 bit flacs of the matrix are 369mb.  the 16 bit are 350mb!  the 24 bit of just the aud are 705mb and the 16 are 346mb.  the raw matrix (un tracked) is 369 as well. 


how in the world could the 24 bit only be 16mb bigger than the 16 bit?  how can there be 336mb difference between the matrix and the aud?


I checked and non of the files are corrupt or incomplete etc.

[JasonSobel]

what program did you use to mix the sources?  My guess is that you had the internal dither set to 16 bit, so even though you think it's a 24 bit file, it's actually just 16 bits of data and then 8 zero-bytes on every sample.

FLAC would see that and would compress the "24 bit" file very much.

[Gordon]

what program did you use to mix the sources?  My guess is that you had the internal dither set to 16 bit, so even though you think it's a 24 bit file, it's actually just 16 bits of data and then 8 zero-bytes on every sample.

FLAC would see that and would compress the "24 bit" file very much.

I used the montage in wavelab.  I'll take a look at the settings but I know I have wavelab set to 24 or 32 bit float.  the raw matrix wave file reads as 24/48 and is 1 gig for 1 hour.


ps  thanks again for the card!  finnaly got everything off them and it'll go out priority monday.  sorry for the delay.

[JasonSobel]

I used the montage in wavelab.  I'll take a look at the settings but I know I have wavelab set to 24 or 32 bit float.  the raw matrix wave file reads as 24/48 and is 1 gig for 1 hour.


ps  thanks again for the card!  finnaly got everything off them and it'll go out priority monday.  sorry for the delay.

is the bit meter showing 24 bits after its mixed?

(and no problem with the card - no worries on any "delay").

[Gordon]

damn it!  so I decoded the "24 bit" file and open it in wavelab and sure enough the bit meter only shows 16 bits.  so at what point did this happen?  in the rendering?  in montage I don't see any settings that would change it and I confirmed that I am in 24 bit float.

[JasonSobel]

yeah, my guess is that it happened in the rendering.  while you chose to render the file to 24 bit, perhaps you had the UV22HR (or whatever other dither you used, but if you're using WaveLab, my guess is that you're using UV22HR) set to 16 bit.  that would cause the file to be dithered to a 16 bit file, yet stored in a 24 bit file.  but it's hard to say exactly what happened.

[Gordon]

yeah, my guess is that it happened in the rendering.  while you chose to render the file to 24 bit, perhaps you had the UV22HR (or whatever other dither you used, but if you're using WaveLab, my guess is that you're using UV22HR) set to 16 bit.  that would cause the file to be dithered to a 16 bit file, yet stored in a 24 bit file.  but it's hard to say exactly what happened.

I use izotope to dither.  anyway my guess is that I rendered in place (not creating a specific file) and that had something to do with it.  sucks to do it all over again but I am.  just for the 24 bit and for my personal 16bit.  no need to confuse the leechers on etree of why they need to re download the 16 bit.

[sygdwm]

funny. everytime i would "process in place" when dithering, it wouldnt do shit. you had to "create a specific file" to got 24>16.

[Gordon]

funny. everytime i would "process in place" when dithering, it wouldnt do shit. you had to "create a specific file" to got 24>16.

for dithering you do.  but for some other rendering you can process in place.  for resampling with resampler 192 for example you can render in place.

[sygdwm]

oic. i dont resample.