Login / Register
 
A PRACTICAL comparison of bit depth
New Reply
Subscribe
BrianT
Thread Starter
#1
15th April 2003
Old 15th April 2003
  #1
Lives for gear
 
Joined: Jul 2002
Location: Nashville
Posts: 1,816

Thread Starter
BrianT is offline
A PRACTICAL comparison of bit depth

I just realized there is a record I mixed that might provide the ability to compare 16 vs 24 bits on a purely practical level.

I mixed an album for Mercury on an artist named Anthony Smith about a year ago. It was recorded 100% directly to Pro Tools Mix over about a 2 year period. It was mixed 100% within Paris, except for a TC6000 connected digitally, and printed to a 1/2" ATR102.

The reason I bring it up is that because of the way the album came together with most songs having initially begun as demos, it is a collection of Pro Tools recordings at 44.1/16, 48/16, 44.1/24 and 48/24, all recorded by one person, all mixed by another person.

My point is, I would challenge anyone to determine which songs are which bit depth and sample rate by using only their ears. Now 1 year later, I can't tell which is which, without looking it back up in the docs.

So I would offer the thought that most of the discussion of this subject is merely academic, at least when auditioned on what I believe is a fairly good sounding commercial CD through high quality D/A.

Which means that maybe if we spent more time worrying about the music, and less time worrying about the specs and numbers (and doesn't digital audio give us lots of numbers to masticate?), people might be buying more CDs.


Regards,
Brian T
#2
15th April 2003
Old 15th April 2003
  #2
Gear Guru
 
thethrillfactor's Avatar
 
Joined: Jun 2002
Location: New York City
Posts: 14,175

thethrillfactor is online now
Brian,

Nice thoughts and I've mixed songs at times lately( started at 16 bit 44.1K) where i wondered the same thing.

My problem is more with what we have to work with at the end.

I have a problem with busting my A** to get something to gel well and then have it bottle necked at the end(16 bit 44.1K).

I've been listening to both SACD's and DVD's lately and the higher bit depth does make a substantial difference(as well as the higher sampling rates).

I haven't seen as many arguments lately about bit depths though when it comes to tracking. I guess people are actually starting to focus more on the music.

By the way, i do hear a difference when mixing with digital effect units(M6000 for eg.) that are being dithered down on the inputs and outputs.

It kinda defeats the purpose pf paying $15K for the unit(which does all its processing at either 48 bit or 64 bits) and then is dithered down to please the interface.
BrianT
Thread Starter
#3
15th April 2003
Old 15th April 2003
  #3
Lives for gear
 
Joined: Jul 2002
Location: Nashville
Posts: 1,816

Thread Starter
BrianT is offline
I would agree about the final mix bit depth. But multitrack, I'm far less sure about there being any worthwhile difference.

Paris is an ideal platform for bit depth comparison, since it's able to accomodate both 16 and 24 bit files in the same session without sonic penalty (unlike Vegas).

Not only have I done back to back tracking passes at both 16 and 24 bits, here's another interesting test I've done.

My Paris system will render tracks to disk contiguously very quickly and en masse. Maybe 6 seconds for a 4 minute file. And since files can rendered at both 16 and 24 bits at will, then using Undo reversed nearly instantly, there is an easy comparison to be made.

I've finished mixes I thought sounded good using 24 bit multitrack files, then simply rendered all of the multitrack files to 16 bits and reprinted the identical mix.

Results? I can't tell which is which on a modern production. Not even. This usually includes some serious plugin use, which is where the supposedly noticeable difference occurs.

Maybe I should do that exact test and find a way to post it online for comparison. The same mix twice at the same mix bit depth, but using 16 vs 24 bits for the multitrack files only. Since I use only Paris for mixing, maybe this is not true in all DAWS. Maybe some other DAWs have a definite bit depth preference for some reason I don't know. Mine does not.


Regards,
Brian T
Jax
#4
15th April 2003
Old 15th April 2003
  #4
Jax
Lives for gear
 
Joined: Jul 2002
Location: San Francisco
Posts: 4,782

Jax is offline
Quote:
Originally posted by thethrillfactor
It kinda defeats the purpose pf paying $15K for the unit(which does all its processing at either 48 bit or 64 bits) and then is dithered down to please the interface.
In a way not directly related to the sys6k, I do feel that one. I know I've said this before, but this is why I'm recording at 16/44.1 more than I ever did before. The audible sacrifice of coming down from 24/88.2 or even 24/44.1 isn't worth it, and if no one else notices, then it actually doesn't matter what resolution we started at.

Maybe it's the quality of my converters (which I have great faith in, so....), maybe it's the dithering, or it could be the way I mix. On the final disc, if you can't tell and you don't have to think about it, it doesn't matter.
#5
15th April 2003
Old 15th April 2003
  #5
Gear Head
 
heylow's Avatar
 
Joined: Jun 2002
Location: Nowhere, Man
Posts: 70

heylow is offline
Interesting thoughts..........

You may have to pardon my ignorance, but I dont see how anyone can hear the difference in bit depth. Am I wrong? It seems to be that, all things being equal, 16 bit is the same as 24 bit except that you dont have bit crunch with 24 bits to get decent resolution, resulting in more headroom without compression at a good resolution.

I'm not saying I have compared directly, because, truth be told, I have only heard 16 bit stuff done with LOTS O compression at tracking whereas the 24 bit stuff is often done without so this alone makes 24 bit "sound better" to me but its not actually the bitrate we are talking about as much as it is the compression in that scenario.

Do you guys actually hear bitrate differences?



As far as sample rate....I havent heard much in the way of comparisons ecxept some internet stuff in which i didnt hear a big fat difference. I could often MAKE myself hear it but in the interest of keeping it real, I would concur that I do not hear what many others claim to. I can do 44.1 and 48 and I dont hear it there.

I guess, in a way, I dont care much. I dont believe it makes a hell of a lot of difference in rock music.....I say rock music because thats what I know...it may be different for classical, I dont know. All I know is I dont believe the hype fo the most part.

Of course, that said, i have to admit that I track digitally and mix analog and find myself tracking at 48k "just in case".



heylow
Jax
#6
15th April 2003
Old 15th April 2003
  #6
Jax
Lives for gear
 
Joined: Jul 2002
Location: San Francisco
Posts: 4,782

Jax is offline
Don't know, heylow. Before it all gets dumped the shallow, narrow pool of 16/44.1, I can hear the difference between 48 and 44.1, especially with cymbals and reverb trails. 88.2 and 96 I can hear that same difference once-ish over again, in a way. The top is more open although to my ears, the bottom seems to thin out a bit, as if some kind of distortion or masking in the lower sampling rates gives the illusion of more robust bass. I've speculated that maybe the low frequenices are being captured more accurately, and that the noise/resonance/digital filter artifacts causing the illusion have been moved that much farther away from what can be heard by all humans except the $6 million man and Bernie Grundman. Bernie must be able to hear all that stuff, right? Or else why would people go to him so often to make things perfect? Must be his level of experience...

it ain't really about his ears or your's or mine as much
#7
15th April 2003
Old 15th April 2003
  #7
Lives for gear
 
RKrizman's Avatar
 
Joined: Jun 2002
Location: Santa Monica, CA
Posts: 7,774

RKrizman is offline
Re: A PRACTICAL comparison of bit depth

Quote:
Originally posted by BrianT


My point is, I would challenge anyone to determine which songs are which bit depth and sample rate by using only their ears. Now 1 year later, I can't tell which is which, without looking it back up in the docs.

So I would offer the thought that most of the discussion of this subject is merely academic, at least when auditioned on what I believe is a fairly good sounding commercial CD through high quality D/A.

I don't even need to use my ears to know that they are all 16/44.1.

-R
#8
15th April 2003
Old 15th April 2003
  #8
Lives for gear
 
doug_hti's Avatar
 
Joined: Jul 2002
Location: Calabasas, California
Posts: 1,142

doug_hti is offline
Quote:
Originally posted by Jax
In a way not directly related to the sys6k, I do feel that one. I know I've said this before, but this is why I'm recording at 16/44.1 more than I ever did before.
Now something that came to my attention lately is PT TDM sessions. I read somewhere, I think in the manual, that you are supposed to dither again on 16bit sessions as well...This obviously makes sense with the 48bit processing.

So my question as well, is doing a session at 16bit even a "16bit" session?

As far as the whole bit depth, sample rate deal.
I think if you are doing the whole DIY approach with a DAW from start to finish, I think it's worth it, but IMO where it's potentially flawed is the software SRC, as well as most individuals not being able to monitor the final mix at the intended delivery rate and bit depth. Mixers don't usually have to deal with that issue, because they turn it into the mastering engineer at the highest rate they are working at, then the ME can listen to it through superior D/A conversion into analog gear and recapture it at the intended sample rate...so they are making adjustments to the song from what it's going to be at 44.1k, not at what it was at 48 or 96, etc.

However, on an actual album last year, half was tracked to logic at 44.1, 16 bit with not the greatest A/D and half was to a mix system 48k, 24bit with Apogee A/D.
The roughs prior to mixing, I could hear the differences in the quality quite considerably. But when they were mixed by a great mixer with 48 outs on to a SSL with additional outboard gear/efx with PT being used for some processing, but mostly as a tape machine, I couldn't tell nearly as much.

My point, if there is a descent amount of outboard and analog gear, I don't think you are going to hear the differences of bit depth or sample rate nearly as much. But with that being said, I think there is a cummualatlve difference over the process when things are done at higher bit and sample rates.
__________________
doug
Rab
#9
16th April 2003
Old 16th April 2003
  #9
Rab
KMR Audio
 
Rab's Avatar
 
Joined: Jan 2003
Location: London
Posts: 534

Rab is offline
For digital recording it has to be the case that, all other factors being equal, the higher the resolution the better the recording. Any DAW is essentially a sampler, and although there's validity in people going all gooey about 12-bit samplers 'back in the day', this doesn't really hold true for a "crunchy" multitrack recording.

Bit depth is a major issue for recordings with a wide dynamic range (eg. classical or acoustic projects) where there is more resolution available at lower levels (because there are more bits down there) so it sounds better. If a track is brickwalled throughout, you're not going to notice it so much because there's not much being asked of the low-value bits. Also bear in mind that even though you're working nominally at 24 bits, you may well be getting only 21-22 genuinely useful bits depending on the quality of the convertor.

Bottom line has to be with CD as the main delivery at present, 16/44 is ultimately the listening experience anyway. But when a 24/96 format is finally here, it's guaranteed to sound better. How could it be otherwise?

At the risk of being seriously flamed, does anyone still find their old 70's rock recordings (ie. AAD) just sound so much better than current DDD recordings? Or is this just because my ears got "trained" that way all those years ago?
__________________
slut first, pimp second...
#10
16th April 2003
Old 16th April 2003
  #10
Lives for gear
 
Renie's Avatar
 
Joined: Jun 2002
Location: London UK
Posts: 1,777

Renie is offline
Quote:
Originally posted by Rab


Bottom line has to be with CD as the main delivery at present, 16/44 is ultimately the listening experience anyway.

This conversation is going in this direction, so MP3 comes to mind.
And radio/TV.

The listening experience for the majority is probably well below 16/44 then? true?
Rab
#11
16th April 2003
Old 16th April 2003
  #11
Rab
KMR Audio
 
Rab's Avatar
 
Joined: Jan 2003
Location: London
Posts: 534

Rab is offline
Quote:
Originally posted by Renie
This conversation is going in this direction, so MP3 comes to mind.
And radio/TV.

The listening experience for the majority is probably well below 16/44 then? true?
Up to a point... but when you're kicking back on the sofa of an evening for a night of quiet immersion in your favourite Brotherhood Of Man album , I'm guessing you'll be listening to CD rather than your iPod? My point is that SACD (and other jostling formats) aside, CD is the best we've got at the moment.
#12
16th April 2003
Old 16th April 2003
  #12
Lives for gear
 
Renie's Avatar
 
Joined: Jun 2002
Location: London UK
Posts: 1,777

Renie is offline
Right, but I always have to dance to that album..
#13
16th April 2003
Old 16th April 2003
  #13
Lives for gear
 
Ted Nightshade's Avatar
 
Joined: Mar 2003
Location: state of jefferson
Posts: 1,328

Ted Nightshade is offline
I am doing some mixdowns from analog to digital, using a Cranesong HEDD.

We tried mixing to 1/4" analog tape, mixing to 24:96, mixing to 24:44.1, mixing to 16:44.1.

What sounded best, by far, to our ears was 16:44.1. I'm thinking the HEDD must be really optomised for 44.1, and the dither is an important part of the sound. The dither is rather reminiscent of some of the nicest things about the 1/4 machine we were using, one with germanium transistors.

Everything just blended best and had the most musical result at 16:44.1. On other material, we might have gone with the analog machine, but we didn't want the compression and noise for the stuff we're doing now (acoustic instruments & percussion), and it's a joy to mix to exactly what will be heard on the CD.

When 24:96 is commonly available as a playback medium, as it may be some day, I hope it will work well to upsample from the 16:44.1 without changing anything.
BrianT
Thread Starter
#14
16th April 2003
Old 16th April 2003
  #14
Lives for gear
 
Joined: Jul 2002
Location: Nashville
Posts: 1,816

Thread Starter
BrianT is offline
Quote:
Originally posted by Rab
For digital recording it has to be the case that, all other factors being equal, the higher the resolution the better the recording
I disagree with that logic. It's only a better recording to me if I can notice the difference, and then if I prefer that difference. And I have demonstrated to my own satisfaction that I cannot hear some of the differences that people assume "should" be there, based only upon a number.

If I can notice, and I like the difference, great. Otherwise, it's just a bigger number.

So my question is, if I play you two different recordings back to back, and you cannot tell which is which by listening, does the recording with the bigger number still have to be a better recording?

I think that if it sounds better, then it is better. But if I cannot hear a difference, it's just some numbers as far as I'm concerned.

I suppose the reason I'm hammering away at this concept is that it underlines a fundamental principle of making music that I believe may be getting lost in the fascination with specs. None of the numbers matters to me if I can't hear the difference, and more importantly, if I don't prefer the difference. Regardless of how badly a manufacturer would like to sell me some new stuff.

The post above this one is a perfect example of what I mean. If Ted had assumed the bigger numbers just had to sound better, he would never have bothered to listen to the other options and make up his own mind.

Always make up your own mind. Because when was the last time a dealer or manufacturer said to you, "Nah, what you've got is great.....you don't need to buy this thing I'm selling. Just work on improving yourself instead." ?

grudge


Regards,
Brian T
#15
17th April 2003
Old 17th April 2003
  #15
Gear addict
 
ExistanceMusic's Avatar
 
Joined: Mar 2003
Location: Melbourne, Australia
Posts: 476

ExistanceMusic is offline
Just a point, bus isn't it true that the differences in bit depth and sample rate really only become apparent at the lowest points in the dynamics for bit depth, and the highest frequencies for sample rate?

In which case, for alot of music, debate over bit depth and sample rate is nearly moot, particulary with rock and pop and the current trend for everything to be as LOUD as possible (debate for another day).
Bit depth degraging as signal level decreases is simple not an issue (provided it's decent to begin with, IE 44.1/16) because there just isn't much quiet stuff going on in alot of music, IMO.

At the other end of the scale away from the loud stuff, when recording/mixing music with large dynamic range like classical or jazz, bit depth/sample rate would be incredibly important, otherwise we'd be listening to nice sounding loud parts, and cheesy lo-fi quiet bits, yes? (I know, it's an exageration, just making a point :D )

So I would think it depends on the program as to what format you choose, am I right?

__________________
Jesse Mahoney

ExistanceMusic :at: hotmail dot com
Rab
#16
17th April 2003
Old 17th April 2003
  #16
Rab
KMR Audio
 
Rab's Avatar
 
Joined: Jan 2003
Location: London
Posts: 534

Rab is offline
Quote:
Originally posted by BrianT
I disagree with that logic. It's only a better recording to me if I can notice the difference, and then if I prefer that difference. And I have demonstrated to my own satisfaction that I cannot hear some of the differences that people assume "should" be there, based only upon a number.
I'm not challenging you on a creative or artistic level - purely on a technical one. This is equivalent to digital photography - the more pixels you have in the frame, the better a representation of reality it is. End of story.

I'm also not disputing whether you can hear the difference or not... for most recordings, I know that I couldn't tell the difference either. But where I know I can tell the difference is in material with a wide dynamic range - ie. recordings with a lot of high-energy and very low level material. This has to make sense... convertor information is less likely to be accurate at the low-level bits - the more bits and higher resolution you have down there, the greater the chance of your recording reflecting the analogue signal that fed your convertor.

I just wonder whether 24/96 recordings will sound inherently better when a successor to CD arrives compared to 16/44? And isn't it worth considering the implications of how the recording will be heard in the (not too distant?) future when printing material now?

Ultimately, I'm guessing we'd both agree that a great performance competently recorded and engineered at "low resolution" (whatever you judge that to be) is a far more rewarding experience for everyone than a pile of pap recorded at X-bits and Y-kHz.

New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+ 
 
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
audiobob / Post Production forum!
1
musaee / So much gear, so little time!
2
DNR / So much gear, so little time!
6
Azarael / So much gear, so little time!
1
EsherOne / Music Computers
1

Forum Jump

SEO by vBSEO ©2011, Crawlability, Inc.