Dithering to Vdat

[YISH313z]

I'm in the process of tracking a song out to Vdat from the direct outs of the 24/48 mixer with two purposes in mind:

1. To Mix-down in Scope from scratch for Client to listen and decide on using.
2. To have the individual tracks separated to deliver to the Client to have a real engineer mix them(Or me if they so choose )

So...
I have 20 stereo tracks and and a few Mono tracks namely Drums(Kicks n Snares n Hats), I have loaded 6 VDAT tapes to handle a one pass run, All instruments are from the Scope environment except a Two Track A-Capella, and some last minute sound effects which are sent by asio out to the STM mixer, so I'm within the 32bit environment for the most part.



My dilemma is if I record this as 32 bit, it will of course not be able to be imported into Pro-Tools with out reducing the Bit rate to 24 bit, and thus losing some of the fidelity.

So should I:


1. Just record the Tracks at 24 bit in VDAT(No Dither)

or

2. Record the Tracks at 32 bit and reduce with a standalone program (Wave Editor) at a later time


Most of the Instruments are 16 bit sample libraries coming from STS sampler, also used was 1 voice of Minimax for bass and Prisma for a lead. The record is an Urban RnB Song


Any suggestions?



Keep in mind, I'm still using 4.0

[astroman]

you don't loose any fidelity by reducing the bit depth - choose whatever is most convenient.
imo whatever is perceived as 'quality' in VDAT within the Scope environment is not related to numeric figures.
It's rather the undisturbed, perfect time alignment of the audio data stream that makes the difference.
At least I could hear the more 'tight' more 'pristine' effect on 16bit data just as well...
I consider myself fairly objective in this context, as it's not the most convenient appreciation
(making things more complicated than desired)
and of course there's no scientific proof - I just notice it from time to time
hypothesis:
a CPU doesn't perfectly align the data due to load and task/thread switching, cache and buffer handling and whatever may come along in this context...
I'm writing about realtime audio (monitoring), not mixdown rendering

cheers, Tom

[Nestor]

Professional engineers prefer, in most cases, to record at the highest possible bit rate and then if there is no another chance, dither the wav files to whatever bit rate you need for the mixing stage, in that case, to get the files into Pro-Tools. It is said to sound better, and I am talking about the final mix, rather than monitoring like Astroman did.

[garyb]

actually, for the mix engines Samplitude/Sequoia are the sequencer's of choice for the top shelf engineers that are only concerned with audio quality, but we have Scope, so the sequencer isn't that important where sound quality is concerned(if you mix in Scope).

just use 24bit.

worrying about something that most customers won't notice is pointless. there's nothing wrong with making it as good as you possibly can, but the convienience of being able to finish the product in our own lifetime can't be overlooked. your converters are only 24bit. recording at 32bit does give a bit more headroom, which is nice, but it's not the difference between a good track and a bad one. the real problem is "is the music great?". if the music is great, people will like it. if the music is bad, a cleaner recording won't change that.

if you record to vdat, that's great, it sounds great, but you lose some editing convienience.
you can edit in any sequencer you like. playback from the sequencer(pretty much any sequencer) will not be the audio equal of vdat playback. vdat recording is very simple and straight-forward vs. the sequencer can include metronome and bar/beat markers making editing really nice and midi is included.

what you need to do for any project is to take stock of what the job is, and what tools will be best to achieve those purposes. it sounds like you know what you want to do, so go ahead and do it. don't stress the minutae, though. if working at 24bit makes it easier to make an awesome production in the end, you're not giving up too much. you'll never get the full resolution heard in the studio to the finished product anyway. a cd is 16bit and a lot of music is listened to in a lossy format like mp3. it IS good to have the best raw material to mix as possible, but the extra 8 bits of 24bit(compared to the industry standard of 16bit) is plenty of headroom for a top quality production. if you were going to edit in Samplitude, i'd say use 32bit...but i don't like editing in Samplitude(although there are many who love it, and they're not wrong either). Audition and Reaper can handle 32bit files, i'm pretty sure, and i'd even bet that Logic can. any good studio would have a sequencer that could play the 32bit files, but for an HD session, the PT sequencer would be nice. 24bit will open anywhere. if you don't know where the files are going, just make them 44.1k 24bit, so that everyone can use them.

[tgstgs]

the noise floor of the algos or samples stored should be less than the final truncation_

good vibes

[garyb]

the noise floor of the algos or samples stored should be less than the final truncation_

good vibes

yes, exactly. pop music doesn't even need to be dithered. truncating sounds just as good for music with the limited dynamic range of most pop music.

[Nestor]

What an interesting topi in here....

I realize that curiously, right now that we have the best equipment and possibilities to end up with very high end quality outputs, we tend to listen at music through the internet most of all, which gives us a poor quality production compared with those HD outputs we are talking about.