44.1 kHz vs. 48 kHz - why not use the higher?

[hnoormohamed]

I am ready to master my CD album and all my final mixes are done on 48 KHZ which is default in PT, will there be a problem if I master at 48 KHZ or do I need to make it 44.1 KHZ?

Please let me know.

Thanks,

Hanif

[Drew Mazurek]

You can master at 48K, but ultimately they need to be converted to 16/44.1 for cd replication.

[albee1952]

Originally Posted by hnoormohamed

I am ready to master my CD album and all my final mixes are done on 48 KHZ which is default in PT, will there be a problem if I master at 48 KHZ or do I need to make it 44.1 KHZ?

Please let me know.

Thanks,

Hanif

Are you mastering, or is it going to a mastering house? If its going out, ask the mastering house what they prefer(when I send out for mastering, I bounce to 96K, even though my sessions are usually 48K). If you are "self-mastering(which I don't recommend) then I would bounce at the session sample rate and master in a new session. Why don't I recommend self-mastering? The 2 most important aspects of mastering are 1-a fresh set of ears(which the engineer can't have), and 2-superior monitoring(which none of us have if we didn't spend the price of a house on speakers, converters, room treatment, etc)

Why do I bounce to 96K? I want to give the ME a high-resolution product and many plugins work better at higher sample rates

Last 2 cents; if you are self-mastering, make the mix as perfect as you can without trying to master at the same time(just my preference, YMMV)

[JFreak]

None of this matters; the ME can upsample with higher precision gear so the ONE AND ONLY thing you need to consider is HEADROOM of your mixes. Do not ever touch the top 6 dB no matter what bit depth your session is. If you can do that, just bounce whatever bit depth and sample rate your session is originally and ME can do the upsampling of needed.

[Carl Kolchak]

Originally Posted by JFreak

:<.... snip....

just bounce whatever bit depth and sample rate your session is originally

... >: snip

I'd disagree in some respects :

If the session is at 16bits, (or even in some cases, 24bits) the mixer summing and effects still happen in a 32bit floating point / 48bit fixed / 64bit floating point world.

Why introduce truncation, or dither, and loose so much?

@ OP :

I'd ask the M.E. If they want 24bit fixed point, or 32bit floating point files.

But yeah, as for sample rate, just leave it at whatever the session was tracked / mixed at - leave any sample rate conversion to the M.E. (Not that it's in any way a lossy process, so long as you aren't reducing the sample rate below what it was originally recorded at).

But if you are self mastering, a red book CD has to be 16bit / 44.1kHz, so you'll need to dither, and SRC.



Sent from my iPhone using Tapatalk

[JFreak]

Well if the software is able to export the 32-bit or 64-bit float mix that it does, it would be great. But usually if you have a "native" session (32/64bit) you can only bounce 16/24-bit fixed.

[Bob Olhsson]

I'd recommend the original sample rate 32 float. All modern mastering systems can handle that. The only exception is an old Sonic Solutions system from the '90s. which some people still use.

I agree with peaking to -6 or even -10. The reason is monitoring headroom while mixing. A lot of common converters get wonky at higher levels and you don't want to be compensating for that wonkiness in your mix.

[jamesleonard]

Hello,

Although 44.1kHz has become the de facto standard in all work stations, many producers today will operate their sequencers at a much higher sampling frequency of 88,200Hz or 96,000Hz. Although these sampling rates are far beyond the frequency response of human hearing, these sampling rates can reduce unwanted side effects such as *Frequency Cramping*

Frequency Cramping often occurs when a processor or effect is employed in an audio workstation at frequencies that are close to half the sampling rate. For example, if the current working project were set at 44.1kHz and the engineer were to boost a wide range of frequencies at 18kHz - the boost occurring on the the higher frequency side of 18kHz could extend well beyond 22kHz. This is more than the projects sampling rate and, in accordance with the Nyquist theorem, this would result in aliasing and Frequency Cramping, whereby the boost reduces sharply resulting in an uneven balance.

The attached thumbnails shows this effect in action - and sonically it appears harsh, reducing both presence and spatial resolution of the sound, whats more, this effect can be much more noticeable when using processors and effects that emulate analogue characteristics - such as: analogue modelled EQ's or distortion units - since these often introduce further harmonics either side of the frequency range being processed.

The diagram shows an unbalanced EQ curve (pink) compared to the original curve (grey)... The 2nd diagram illustrates this phenomenon.

*(Please substitute 10kHz for 18kHz - because I could not be bothered to plot a logarithmic graph for this example - but this diagram will suffice )

Also.. the boost at 1k serves no purpose in THIS example - so please disregard it - as it has no relevance to the diagram - so forget it's even there