What is the universally prefered mastering format?

[Big 3rd]

What should I send my mastering engineer?
24 bit/96k/aiff file....16bit/48k/ wav....what?

And why?

[lucey]

Whatever bit and sample rate you're using to track and mix. Not up or down. Check with the ME first, but all should work. There is no 'standard' for mixes, that's your call. I've heard 16/44.1 mixes that trump all. 24/44.1 WAV is the most common for me. AIFF maybe 15% of the time.

[drakewire]

Whatever you recorded at, if you recorded at 24 bit 96 then send that if record to DAT send that. Dithering is something that should be done by an M.E... The most important thing to remember every time you dither you are losing information, so its best to leave that info for an M.E to decide

My two Cents

[Adam Dempsey]

Standard interchange format (AES) is actually broadcast wav (.bwf) but .wav suffices.
Recommended: 24 bit and at whatever your recording/mixing sample rate.

Dithering is something that should be done by an M.E...

To 16 bit for CD, yes. Not wanting to confuse things but it doesn't hurt to dither from higher DAW word lengths (eg, 32 bit) to 24bit (again, AES standard). Mathematically, this will minimize digital losses, especially if applied at every 24 bit post-dsp bottleneck in the signal chain, and as performed anyway by many processors & plug ins.

Any reputable mastering facility will have their preferred formats stated on their website, or via a phone call...

[masteringhouse]

In addition to the above stereo interleaved files are usually preferred over dual mono to help insure that there are no issues matching phase between tracks.

[Lagerfeldt]

Interleaved WAV 24 bit 44.1 kHz

I just mastered an album with split files in 24 bit/88.2 kHz, and lo and behold - one set of files was not correct, but no phase issue just some kind of bug in the end.

I used Barbabatch to make all files interleaved before importing them.

[J.P.]

If the audio is being mixed outside the box, why not record the two track to DSD ? How does this format factor in for the modern mastering engineer?

[Riccardo]

Originally Posted by J.P.

If the audio is being mixed outside the box, why not record the two track to DSD ? How does this format factor in for the modern mastering engineer?

This is a production choice, not a mastering one.

Some people will tell you DSD sounds is better, others that PCM is the best choice, at the end of the day whatever camp you are in the choice is yours.


p.s In doubt you could print either 1/4" fifteen ips or 1/2" thirty ips

[Misja]

Send in whatever you are mixing in. Tape reals are so passé but cool!!

[jayfrigo]

As others said, whatever the native mix is. The most common formats are wav, aif, or broadcast wav format, anywhere from 44.1 kHz to 96 kHz (192 is OK too, but quite rare). 24 bits is preferable and most common, but anything from 16 to 32 (float) is OK. I probably see more 24-bit / 48 kHz wav than any other flavor, but I do see a pretty healthy variety across the board. Analog tape is less plentiful every year, and DSD is pretty specific to individual clients and specific projects, and not widely seen.

[editronmaximon]

I would just add the following if you are working at 44.1kHz or 48kHz, and then doing an analog mix and capturing that analog mix to a second daw:

When you capture the 2-mix off the analog console, you would do well to capture at one of the higher sample rates, and at 24 bit.

This may have a couple of advantages. It may better capture some of the "stuff" generated by the analog mixdown. It may also be helpful to the mastering engineer, since higher sample rate and higher bit depth files may take further digital processing better.

[Misja]

Since we are working at least several years with that format that is kinda of a no brainer no pun intendet. The master engineer can also decide to upscale that so te definition of what he does after the mix is well needed or not. That's why deliver it in the format you are mixing is all about.

24bit is oke when you are on an really outdated daw.

32bit float is better or like in others even higher. But you can't really go wrong with 24bits.

It remains deliver the format you are mixing in ... providing it's not clipping.

[editronmaximon]

Again, if you are doing an analog mix, I would suggest printing it at the higher sample rate. That is better than the mastering engineer having to upsample it.

32 bit floating point has a disadvantage in that you cannot really properly dither floating point audio [if that matters to you]. However, it seems that a lot of the daws that mastering engineers use are floating point, so that's kind of a drag. 64 bit float is better than 32 bit float.

If you can determine that your mastering engineer's daw is fixed point, then I would definitely send 24 bit fixed.

[Misja]

Originally Posted by editronmaximon

Again, if you are doing an analog mix, I would suggest printing it at the higher sample rate. That is better than the mastering engineer having to upsample it.

32 bit floating point has a disadvantage in that you cannot really properly dither floating point audio [if that matters to you]. However, it seems that a lot of the daws that mastering engineers use are floating point, so that's kind of a drag. 64 bit float is better than 32 bit float.

If you can determine that your mastering engineer's daw is fixed point, then I would definitely send 24 bit fixed.

Really? Sorry this is going off topic.

32Bit float is 24 bit fixed but added 8 bits for convenience really hard to clip if you can add a bit when needed. With a headroom 24bit fixed near about -144Db it's hard to venture for more. Not that i say it's not better but we really need better hearing for that. 64 for bits floating in my understanding is really 48 bitls fixed added with 16 bit's to be used below the -144 Db headroom or added if you wish so. That makes it 288 Db of headroom? Damn that's much. Oh wait i used up just one more bit. Vitually there is no real difference. Practical use is never use the extra floating bits. Agreed with the doubled fixed points. But there is always. If it sounds good it is good

[jayfrigo]

Just to be clear (thread lurkers and such), the requirements of processing and delivery are different. While 64-bit floating point may have some advantage in processing, 24-bit fixed point is plenty for delivery. Once it's out of the box, real world analog signals can't even use the full 144 dB dynamic range. 24-bit audio also will be perfectly happy to be subsequently processed at 32-bit float, 48-bit fixed, or 64-bit float.

Saving in a 32-bit float format is useful if you plan to do further digital-domain processing within the DAW environment, such as tracks in an on-going project. However, it will need to become fixed point eventually if you plan to let it out into the real world, whether by transmission, such as by an AES line, or to be output from a DAC.

[24-96 Mastering]

If you can determine that your mastering engineer's daw is fixed point, then I would definitely send 24 bit fixed.

From my perspective, at this point, I'd generally rather receive 24 bit fixed than 32 bit floating.

The headroom advantage 32bit FP offers can become a problem when utilized (as a delivery medium): When receiving a 32FP file that exceeds 0fs, I don't know whether the client listened to the mix clipped or had the signal stream attenuated somewhere down the line before conversion. When receiving a 24 bit file otoh, I can be reasonably certain I'm listening to what the client listened to.