LP>CDR, 24bit/96kHz vs. 16bit/44.1kHz, downsampling, HELP!!!

[Kuja]

Hello,

I want to transfer a huge LP collection to CD-R for occasional listening, ...in my office, car etc.

I bought external M-Audio Audiophile USB ADC/DAC

It can record up to 24bit/96kHz

My TT rig is:
Thorens TD-126 MkI turntable,
SME 3009 S2imp arm,
Shure V15VxMR cartridge,
Graham Slee Gram Amp 2 SE phono preamp,

My SME was rewired with Van den Hul MCS - 150 M cable.
Cable from tone arm to preamp is Van den Hul D - 502 HYBRID.

Shure V15VxMR has this microridge stylus, which is great for archiving old LPs - much less distortion from worn out grooves.

The downside is that the treble is rolled off and the sound is too warm for my taste.

(yes, I played a lot with cables and capacitance, but this just seem to be the character of this cartridge)

Using test LP sweep tones, I created a custom graphic EQ preset for Adobe Audition 1.5 (CoolEdit originally before Adobe bought it).

Now I get ruler flat frequency response from my cartridge/cable/preamp combination and the sound is much crisper now

The big question is:

do I record at 24bit/96kHz, apply the graphic EQ and then downsample to 16bit/44.1kHz?

At first I thought that this is a good idea.

But then I started reading numerous web pages dedicated to downsampling...

My brain got swollen from all this dithering, triangular noise shaping, custom curves, bit mapping, etc.

I came to conclusion that I might get some distortion, additional noise, lowered dynamic, digital artefacts, etc. when I downsample.

On the other hand, 24bit has much more information to work with...

I only plan to use the graphic EQ with the adjustments no bigger than +3.8 dB.

Nothing drastic.

Do I get better sound by applying moderate EQ in 16bit/44.1kHz,
or doing it in 24bit/96kHz (or better a 2x multiple of 44.1kHz) and then downsampling to 16bit/44.1kHz

If downsampling is better, what settings would you recommend for Audition/CoolEdit?

Thanks a lot,

Aleksandar

 

[magicplastic]

Downsampling should always be the last stage in the process as any digital eq is subject to rounding errors which will introduce distortion - this will be far less noticable at 24bit than 16bit.

 

[Kuja]

Downsampling should always be the last stage in the process as any digital eq is subject to rounding errors which will introduce distortion - this will be far less noticable at 24bit than 16bit.

Thanks!

So you are saying that distortion introduced by downsampling from 24 to 16 bit is less, compared to applying moderate EQ in 16bit?

 

[magicplastic]

No, Im saying that if you are recording in 24bit then you should eq 24bit, just because you can.

24 bit will always capture your input more accurately, and you sound like the sort of person who would notice the difference

 

[Kuja]

OK, I'm not sure if you understand my dillema.

I won't use 24bit only because I can.

I know that 24bit contains more information and that theoretically I should do all editing in 24bit.

But I'm not staying in 24bit, because I will burn 16bit CDs.

Downsampling from 24bit to 16 bit seems to be a very complex thing and it can introduce distortion and many artefacts on its own.

Since my final output is 16bit CD, I'm wondering which is going to do more harm to my music:
a) EQ in 24bit followed by downsampling to 16bit,
b) or EQ in 16bit without resampling.

 

[dangus]

Why not just leave it at 24/96 and burn them as audio DVDs, since DVD-R blanks are now cheaper than CDR. 16/44 is sinking into the tar pits of history.

See this page: http://www.eximius.nl/dvdaudio.php

 

[Kuja]

Why not just leave it at 24/96 and burn them as audio DVDs, since DVD-R blanks are now cheaper than CDR. 16/44 is sinking into the tar pits of history.

See this page: http://www.eximius.nl/dvdaudio.php

I don't have a DVD-A player.
My Panasonic DVD player has 24/192 DAC, but its analogue section is crap and it sounds awful compared to my Rotel RCD1072 CD player.

Also I can play my CDs anywhere, not so with DVD-A

 

[macboy]

OK, I'm not sure if you understand my dillema.

I won't use 24bit only because I can.

I know that 24bit contains more information and that theoretically I should do all editing in 24bit.

But I'm not staying in 24bit, because I will burn 16bit CDs.

Downsampling from 24bit to 16 bit seems to be a very complex thing and it can introduce distortion and many artefacts on its own.

Decimating from 24 to 16 bit is not a problem. In fact, with a good quality noise-shaping algorithm, you can get better perceived S/N ratio and Dyncamic range than from a direct 16 bit sample. Changing sampling rate (e.g. 96 kHz to 44.1 kHz) has the potential to cause more problems. In this particular case, the difference in sampling rates is > 2, which means you will get better results than say from 48 kHz to 44.1 kHz (which I would recommend against).

Since my final output is 16bit CD, I'm wondering which is going to do more harm to my music:
a) EQ in 24bit followed by downsampling to 16bit,
b) or EQ in 16bit without resampling.

Definitely sample and EQ in in the highest resolution possible, then decimate down to 16 bits (with dithering/noise shaping, as desired) at the very latest possible point, i.e. after all processing of all kinds is completed. The only real decision to make is whether to record at 44.1 or 96 kHz.

Some DSPs maintain the audio samples at a very high resolution internally: 32 bits is very common, and 48 bits is used as well. At the final output, the audio is decimated to the required number of bits. Since most EQ is done using IIR filters, maintaining high precision is extremely important; rounding errors can lead to unstable filters since the output becomes part of the input. 16 bits is generally considered to be insufficient for IIR filters in hi-fi audio. You can bet that even if CoolEdit/Audition is using 16 bit input and output files, it is maintaining at least 24 significant bits during processing, decimating back to 16 bits after processing (and hopefully only when saving the file).

 

[magicplastic]

Sorry, I didn't make myself clear enough

Recording and eq'ing at 24bit 96k then resampling ( dithering) down to 16bit 44.1k will sound better ( in theory, depending on your ears and equipment ) than recording and eq'ing at 16bit.

There will be a loss of quality going from 24bit to 16bit as you are discarding data

I would agree that you would be better using dvd-r and making audio dvd's, but that might require some new equipment to play back on.

I am not an expert on dithering algorithms, so , if it were me I would stick to the default settings in whatever program I were using - that whole area is very complicated....but its better to have started with more accurate (24bit) representation of the audio, especially with dynamic material.

Its easy to get bogged down in sound conversion techniques ,but the differences are normally marginal and far outweighed by the fact that your data(music) is more accurate to begin with.

I would try recording 10 seconds of dynamic audio at 24 bit and the same material in 16bit - dither one down and see if you can hear the difference - set it up so you can do an easy a-b comparison - thats a good way of auditioning different dithering techniques too

If you keep the original raw 24 bit files archived, it will make it easier to redo the cd's if you find a dithering technique that works better for you, rather than starting the whole process again

Hope that helps...

 

[jcx]

96 to 44.1 math works fine, finding an implementation you can be confident in is more difficult

you may be able to record @ 88.2 Ks/s 24bits which renders the decimation trivial, you still need to find competent bandlimiting filter and rounding/dither code

surely these are available as plugins somewhere

maybe audioasylum's PC Audio forum could help

 

[magicplastic]

I think I posted at the same time as macboy - but we are both saying roughly the same thing...

 

[tbrowne]

24/96 -> 16/44.1 downconversion

You might find this useful... I wrote it for this very purpose...

FinalCD

It's nowhere near as fast as CoolEdit in "HQ" mode due to the way it was written, but the quality is plenty good enough, IMOO.

- Tom.

 

[wa2ise]

I won't use 24bit only because I can.

I know that 24bit contains more information and that theoretically I should do all editing in 24bit.

But I'm not staying in 24bit, because I will burn 16bit CDs.

Another advantage of using 24 bits to sample your vinyl is that you can have lots of headroom without sacraficing resolution. More specifically you could set the audio amplitude to half level. Which would normally yield 23 bit quality. But if there is a transient spike of loud music it won't get clipped. So in "post processing" you could normalize the digitized waveform so the loud spike is maintained and still adjust for 16 bit audio for CDs. In other words, you don't have to be as careful with analog level setting than if you were doing 16 bits directly.

 

[Werner]

Record in 24b (or rather '24' bit) and 88.2kHz.

Keep an eye on recording levels, because the true ADC accuracy of that M-Audio is slightly more than 16 bits. Definitely not 24.

Configure Audition to work in 32 bit mode, and disable default
dithering after every process step (dithering should only be applied
once, if at all!).

Then edit the clicks out.

Then apply your EQ together with gain change (if so desired).

When all is finished downsample to 24b/44.1kHz, and
only then requantize to 16b, using dither. (There is a chance you
don't want dither, as the LP input contains plenty of noise to be
self-dithering. Same for noise shaping.)


The reasons for the above procedure are that

1) the EQ works better if it is further removed
from the system's frequency limit and

2) the downsampling in Audition
is massively better than the decimation filter of that cheap ADC in the M-Audio. That ADC aliases like there is no tomorrow, and this will be triggered aplenty by the ultrasonic crap and distortion spuriae coming from every cartridge.

Since you are EQ'ing you may want to include a soft and smooth roll-off beyond 16kHz or so. Nothing drastic, just to -3dB at 22kHz or so.