Tracking and mixing at 88.2/24bit...anyone doing it?

[darthjujuu]

This question is as old and monotonous as any, and dates back to the origins of teh interwebz ...

but I'm curious as to what you all are CURRENTLY doing and have experimented with.

completely a waste to track/mix at 88.2 and then dither down to 44.1 at mastering phase?? or maybe not even dither down at all, since.. these days, most of us DIY dudes are deploying mostly online and alot of our stuff goes straight to our audience's iTunes and never even sees a CD. so i feel like higher resolutions might be becoming increasingly more and more relevant.

i say 88.2 instead of 96 because it is a somewhat known fact that dithering down from 88.2 to 44.1 makes more mathematical sense as they're divisible by 2.

also, while we're on the subject: dithering techniques? some better than others? practical applications? wtf does dither even mean anyway?


DISCUSS.

 

[Klosure]

I don't bother with anything over 44.1, mainly cos I think that my Apogee duet itself added so much more with its converters over my previous interface (and that was an RME Fireface).

I think that the 24bit is the important area here, there are benefits to 88.2 but in the processor usage department, it makes more sense to stick to lower sample rates as the gain is pretty insignificant compaired to the amount of plugins and the quality I can run them at.

 

[HatchetDrummer]

I think it depends heavily on the signal chain capturing the source in the first place. The better the analog chain that has the ability to correctly capture the upper harmonics is the one thats going to allow you to really hear the difference between the two. If your analog chain is lower quality then your more often then not just capturing more crap then you really need. JMO.

IMO when tracking lighter music, i.e. anything with acoustic guitars and simple mic setups on drums that will never see sample replacement, I've noticed that the higher sample rates, I've only messed with 88k personally, have a more tape like quality in them. Not the compression aspect but the overall picture just sounds a little bigger.

I've never directly compared tracks that were done on tape vs. a higher sample rate, but from my experience with 2" machines the closest your ever going to get to one is to track at the higher sample rates, but there is still a lot of that mojo missing, and no sample rate or converter is ever going to be the same, again JMO.

Also a lot of it will have to do with your converters and how the different sample rates effect them. Most people overlook the fact that certain AD/DA chips have different sounds at different sample rates. Not only the chip but the analog stages as well can have a profound effect on how the converter performs at said sample rate. I've mixed with a number of converters that all have their own sounds and have noticed that some converters just like being at 44k, or like both of my 3g's like being run at 96k. (never recorded at this sample rate but I've watched a few blu-rays in the studio before and OMFG!) I also think that its worth mentioning that ITB compression, at least to my ears, sounds more natural and more like a real analog compressor at higher sample rates. I never really got into the Waves API 2500 until I used it on acoustic guitars in a project that I recorded the entire album at 88. But at that sample rate, for that purpose, I would have been scared to replace the plug-in for the real deal because it sounded soo damn good.

I personally record at 44k mainly for the fact that i have 2 8 channel ad/da adat converter boxes that I run into my two 3g's. So in order to get all 24 channels in and out of the computer and to the console I need to run at 44 or 48. I could mix stems into the console and use 16 channels in order to run my system at higher sample rates, but until can afford two more 3g's I'm good with 44 or 48k recording. I love the sound of analog summing enough that I try to do whatever it takes to do as little summing as possible ITB.

Not a mastering engineer, but if I do need to add dither to a mix, its always apogee UV22hr or Sonnox Oxford TPDF. The first mastering session I ever sat in on was with John Cuniberti and after talking to him about it TPDF is what I use 99% of the time.

my 2 cents

 

[Metaltastic]

One thing to clarify Julian - dithering masks quantization errors that occur when going from 24-bit to 16-bit, it doesn't have any effect on downsampling artifacts The main reason 88.2 is preferable is because it's a straight 2:1 ratio with 44.1, so there are far fewer (if any) artifacts compared with going from 96 or even 48 down to 44.1. And yeah, I guess if you've got the HDD space and processor power to handle tracking at 88.2/24, you might as well! (I mention processor power since having all your files being twice as big tends to slow down the sessions, though maybe that's just more of a RAM or bottlenecking issue)

 

[greyskull]

Nah

 

[dani]

you've also got to consider your plugins have to work twice as fast for 88.2, as they need to run twice the number of samples, etc. so, you'll instantly double the CPU cost of any plugin!

also, your buffer settings might need to be doubled, as 128 samples latency suddenly is half the length of time it used to be..

44.1khz 24bit seems to be the norm around these parts!

thanks,

 

[Wolfeman]

I do everything at 96k 24bit.

Nah just kidding, I much prefer 44/24 for some reason.

 

[KeithTidd]

i record at 88.2 then mix in 44.1 - I'd mix in 88.2 but my setup can't handle it. Try at least tracking at 88/96, i think it makes a noticable difference in the sound. If you have cheap convertors, running them at higher rates will make them sound alot better.

 

[BLUElightCory]

I've done some tests using different sample rates, and I (much to my disappointment) prefer the sound of 88.2 or 96 kHz, it just seems to have a nicer depth and high end that I can't quite get at lower sample rates. However, I track at 44.1 and/or 48 because I don't have the resources to pull off using 96kHz unless it's a simple acoustic/piano/etc. recording without a lot of tracks.

It is also not necessarily true that (with modern conversion) 88.2kHz downsamples "better" than 96kHz - the "divided by 2" theory is unimportant when dealing with the actual math of how most modern sample-rate conversion works.

Record at the highest sample rate you can get away with and you'll be just fine.

 

[Metaltastic]

i record at 88.2 then mix in 44.1 - I'd mix in 88.2 but my setup can't handle it. Try at least tracking at 88/96, i think it makes a noticable difference in the sound. If you have cheap convertors, running them at higher rates will make them sound alot better.

Hmm, how exactly do you do this dude? Do you edit at 88.2 as well, and then export at 44.1 once you're ready to consolidate?

 

[John_C]

If you have cheap convertors, running them at higher rates will make them sound alot better.

I'll second that *in theory*

i'm not saying it makes any audible difference because i don't know, but in theory it makes sense that you'd hear more of an advantage upping the sample rate with cheap converters than you would with expensive ones

 

[RCW]

On the theory of running convertors at a higher sample rate to get a better sound.... I have a MOTU unit. When I run that shit at 96 khz it gets extremely airy and has this quality about it that I personally don't associate with warmth or very good sound at all TBH.

That's my take on it. It might just be my setup though (or my ears)

 

[waltz mastering]

i say 88.2 instead of 96 because it is a somewhat known fact that dithering down from 88.2 to 44.1 makes more mathematical sense as they're divisible by 2.
DISCUSS.

It is a myth that 88.2 will yield superior results when converting to 44.1 based on the idea the conversion is simple math.

Depending on which sample rate converter you are using, they can all sound different whether downsampling from 88.2 or 96k based on the algorithms and filters used.

Here is a comparison of different src's and how the differ from each other going from 96 to 44.1
http://src.infinitewave.ca/

Whether 88.2 kHz or 96 kHz sounds better often comes down to the converter your using and the sound you prefer.

also, while we're on the subject: dithering techniques? some better than others? practical applications? wtf does dither even mean anyway?

Dither is low level noise that is added to the signal to disguise or cover up quantization errors or truncations that are caused when changing bit depth.

Some people prefer the sound of different dithers and noise shaping, but this can only be heard at very low levels and sometimes can be insignificant when it comes to making a recording sound good in the first place, but shouldn't be over looked when your going for optimum quality, cause it all adds up.

 

[darthjujuu]

a+ thread. some mixed opinions as usual...but interesting. maybe something to think about for us who've never ventured out of the 44.1 realm, ey?

 

[Harry Hughes]

It is a myth that 88.2 will yield superior results when converting to 44.1 based on the idea the conversion is simple math.

Depending on which sample rate converter you are using, they can all sound different whether downsampling from 88.2 or 96k based on the algorithms and filters used.

Here is a comparison of different src's and how the differ from each other going from 96 to 44.1
http://src.infinitewave.ca/

Whether 88.2 kHz or 96 kHz sounds better often comes down to the converter your using and the sound you prefer.

Definitely.
I'm honestly still surprised people are still believing that 'simple math of conversion" myth.
I'm pretty sure it was debunked years and years ago

 

[John_C]

wow, that website gives some interesting results. Izotope and core audio leopard seem to outperform almost everything

 

[darthjujuu]

http://src.infinitewave.ca/

holy bannanas... i always suspected cubase of sounding a bit "brighter" than reaper. compare the 1khz tone of reaper to the 1khz tone of cubase o_o looks like it wasn't just in my head after all

 

[Machinated]

http://src.infinitewave.ca/

holy bannanas... i always suspected cubase of sounding a bit "brighter" than reaper. compare the 1khz tone of reaper to the 1khz tone of cubase o_o looks like it wasn't just in my head after all

erm dude, you sure you understand this right?

so you are tracking in both at 96kHz and then downsampling?

 

[darthjujuu]

erm dude, you sure you understand this right?

so you are tracking in both at 96kHz and then downsampling?

alright, now i get it.

 

[HandsOfDespair]

Tracking at higher than 44.1 would make a difference if you record stuff like jazz and classical music, not rock or metal.