But wouldn't this mean eac3to transcode from 16 -> 24 -> 16?

(2021-08-24, 01:28 AM)bendermac Wrote: But wouldn't this mean eac3to transcode from 16 -> 24 -> 16?

No. Internally, eac3to works with 64 bits. Any DAW will similarly require dithering to output a 16-bit file after changing its pitch/tempo.

As for the 'best' method, it may well be verging into imperceptible/placebo territory but iZotope's Radius tool probably outperforms everything else.

iZotope's Radius is no longer available. Dead for years.

Thought I'd share some of the knowledge that pipefan passed onto me previously. This is how you do speed changes in izotope.

First of all the calculation:

(new speed/existing speed) x sample rate

So for example with Cinema DTS it would be

(24000/1001)/24) x 44100 = 44056

You process your aud as normal in foobar, then load your wav into izotope. Select resample and enter 44056 but tick change tag only. Then resample again and enter the sample rate you want the track to be (i.e. 44100, or 48000 if it's going on a disc etc) but this time do not tick change tag only. That will give you a new wav with the correct speed. Then use the gain effect on your surround and LFE channels. Then finally use the dither effect and ensure you choose auto blanking to avoid dithering any periods of silence.

You can substitute the figures above for any conversion you want.

(2023-07-27, 07:49 PM)alleycat Wrote: You process your aud as normal in foobar, then load your wav into izotope. Select resample and enter 44056 but tick change tag only. Then resample again and enter the sample rate you want the track to be (i.e. 44100, or 48000 if it's going on a disc etc) but this time do not tick change tag only. That will give you a new wav with the correct speed. Then use the gain effect on your surround and LFE channels. Then finally use the dither effect and ensure you choose auto blanking to avoid dithering any periods of silence.

So you have to resample twice? What is the advantage to doing this in Izotope vs a dedicated audio editor like Audacity? With VHS rips, I often have to do numerous, repeated speed changes until I can get the timing close enough since they are often off by something small, like 0.005% and this can be dependent on VHS player. Then once the speed is adjusted, the sync can be done.

No you only resample once because with the first process all you change is the tag. In terms of benefits as izotope is a professional program it's supposedly better at conversions that a free program like audacity, but I can't back that up with actual examples. I'm partially deaf so I can't pick up on differences but pipefan used to say he could tell the difference and that audacity sounded (to his ears) like garbage.

But if audacity works for you then by all means continue. I suspect in your case it might be harder to calculate the initial tag change as you won't be changing from one set speed to another like one would when processing Cinema DTS. I'm just putting this here as it's not as obvious how to do these changes in izotope now. I gather back in the day it was simpler using the radius tool discussed above but it's no longer part of the program.

i appreciate the input. i just wasn't sure if there was some major reason to do it in izotope that i was unaware of.

i've used audacity for all my syncs and comparisons. i've a/b compared dozens of PAL dvds slowed down to match US DVDs/blu-rays and if the two tracks are the same identical mix, i've never heard anything wrong with audacity's slow down. i guess i can just throw something into izotope and test it out but i'd be surprised if audacity sounded any different, let alone "like garbage". if there is a difference i'll be the first one to make sure everyone knows, lol.

Hey So I have been trying the MPV method of tempo/pitch change. and it works pretty good.
For example PAL to NTSC would be
Out put channel 2.0 by default.

For higher audio channels one could add

(2023-07-27, 07:49 PM)alleycat Wrote: Thought I'd share some of the knowledge that pipefan passed onto me previously. This is how you do speed changes in izotope.

First of all the calculation:

(new speed/existing speed) x sample rate

So for example with Cinema DTS it would be

(24000/1001)/24) x 44100 = 44056

You process your aud as normal in foobar, then load your wav into izotope. Select resample and enter 44056 but tick change tag only. Then resample again and enter the sample rate you want the track to be (i.e. 44100, or 48000 if it's going on a disc etc) but this time do not tick change tag only. That will give you a new wav with the correct speed. Then use the gain effect on your surround and LFE channels. Then finally use the dither effect and ensure you choose auto blanking to avoid dithering any periods of silence.

You can substitute the figures above for any conversion you want.

Having used this method a lot, I would simply say that izotope's resampler is almost perfect, but eac3to's is maybe 99.9% worth it.
The simplest for me is to use it, especially since the technique of changing the rate of samples on izotope, does not give a time as precise as EAC3to since the values are rounded when it is done on izotope (even if the difference final is only ≈100ms).
For my part, I mainly capture PAL laserdiscs, so I do this:
-slowdown (or for cinema DTS : -24.000 -changeTo23.976 ) -resampleTo4800.

I leave the output file in 24 bit, because the dither is not audible that way.
EAC3to uses TPDF which I prefer over Mbit+ with noise shaping.

I would also advise always exporting as 24 bit at the end.
But considering the size, it can be interesting to release the 5.1 tracks in 20 bit.
In 20 bits, the container will show 24 bits, but inside it contains 20 bit with 4 superfluous 0 bit.
These 4 superfluous bits weigh nothing in DTS-HD MA / THD / Flac. But in PCM you will have the size of a 24 bit, so there is no point in exporting in 20 bit if you plan to deliver the audio in PCM.

Comparison of a resampling exported in 24 / 20 / 16 bits :
https://slow.pics/c/ONYVFlaI