Hi,

Today hi-res audio formats such as 192kHz/24bit are being introduced, claiming to improve sound quality compared to CD-quality 44.1kHz/16bit audio.

But is that really true?

* Sampling frequency: According to the Nyquist theorem it should be sufficient to use a sampling frequency of 2x the max humanly detectable audio frequency. So a sampling frequency of 44.1 kHz should cover audio frequencies up to 22.05 kHz, which is above what humans can detect. Now that assumes a perfect low-pass filter, which does not exist in reality. But even giving room for a realistic 2-4 kHz range for the low-pass filter, it should be sufficient. So why use a higher sampling frequency, such as 192 kHz?

* Bit resolution:
a) 16 bits provides 2^16=65536 levels of amplitude. Can the human ear really detect more than that? And does the audio equipment (speakers, amplifiers) not introduce more coarse amplitude granularity than that anyway?
b) Of course, one could argue that the quantization noise is less with 24 bits compared to 16 bits. But is that really an issue in reality?

Please let me know if you have nay info or comments on this.

Blind test

There are several others. All similar results.

Been debated here a number of times. There is an AES paper as well but that combines a number of tests to show a small positive outcome even though none of the individual papers cited show an audible difference.

So, I will not worry about this nor will buy the so called hi-res, no need. I enjoy the music, not the nits that may or may not be there. My clock keeps on ticking; don't have time to waste.

The problem is most high res audio is poorly mastered. So hearing any differences will be minor. There are a few that properly master high res audio like AIX audio which will be the best way to hear any differences.

About the biggest advantage high res audio gives you is you don't need a brick wall filter to band limit the signal prior to conversion into the digital domain (to avoid aliasing). If you sample at 44.1kHz, that gives you a max frequency of 22.05kHz. If you want to limit the audio to 20kHz, it means your filter has 2.05kHz to go from pass to block. This can be tricky to design,especially one that avoids lots of phase changes and amplitude swings near the limit.

If you sample at 48kHz, you have 4kHz to design your filter. Use 96kHz and you have the entire range from 20-48kHz for your filter to go from passing to blocking, which is much easier to design. Remember these filters have to be done in the analog domain, where your resistors, capacitors and inductors are imperfect.

Once in the digital domain, resampling and band limiting is much easier - we can design filters that digitally filter the audio far better and far more perfectly since the effects are highly controllable and very repeatable and we're not dealing with imperfect components.

So with properly mastered hi-res audio, a difference can be heard?

If so, what is the technological explanation for that?
-That the low-pass filter is easier to make, so higher frequencies pass through compared to when sampled at 44.1 kHz (as you discussed in your previous post)? Most adult persons have problems even hearing frequencies that high anyway (i.e. in the ~20 kHz range).
-Or is it due to 24bit resolution instead of 16bit? If so, is it due to less quantization noise, or less coarse amplitude granularity?

The underlined pieces haven't been true since the 1980's. In actual practice the digital signal is first oversampled and then the anti-aliasing filter is implemented digitally. The analog filter is quite gentle and easy to implement since it filters at a much higher frequency. This Wikipedia entry provides a good explanation.

https://en.wikipedia.org/wiki/Oversampling

One thing I have noticed and its especially noticeable for me in my car but I can up the volume quite a bit on higher rez titles while suffering no distortion.

Kind of a brave post.

Recent tests have shown that, all things otherwise being equal, the difference between 16/44.1 and "hi-rez" is inaudible. In the recording process, where wiggle room is more important, 24-bit makes sense, even if it will be mixed down to 16-bit for the end product. But the same is not true for playback-only. There, 16-bit is just fine.

As for sampling rates, there is some indication that higher rates like 96 kHz and 384 kHz do more harm than good. Though this is not audible either.

The best seems to be either 16/48 or 24/48, which would have been good to know back when they were inventing CDs, but there you are.

Well, no. As stated elsewhere oversampling many times has been a standard from the days when they advertised how many times the player over samples.
Next excuse?

Uh, using 192khz for your sampling rate is a form of oversampling. You're using a higher sample rate than necessary, which if we assume is 20kHz for human hearing, is many times higher than you need.

You still need an antialiasing filter because any frequency higher than the Nyquist frequency will be aliased as a signal, and you cannot post filter it out. But using over sampling and a relaxed filter design is why you over sample in the first place - it's easier to design a good low pass filter that starts at 20khz and blocks completely at 96khz (using 192khz sampling). You still need the analog antialias filter because anything higher than 96khz, be it 96001hz or even MHz, will alias into the digital signal. No amount of digital filtering will get rid of an alias because you cannot differentiate between an alias and desired signal. It's like recording a flute and clarinet together and trying to remove the flute afterwards.

Once in the digital domain, you can band limit the signal from 96khz down to 22.05 or 24khz via a digital filter so you can down sample to 44.1 or 48khz. This will give generally better results because of the lowered analog filter design requirements.

The only question one has to ask re: supersampling is: when to do the downsampling? Do you want to reproduce higher frequencies than the ear can detect only to let the ear filter those frequencies out, at the expense of higher electronics costs, more bits in your DAC, and higher bandwidth? Or do you do that filtering at the studio, and ship 16 / 44 to the end user and end it. The latter is the only sane / logical approach, unless one is using the samples for some kind of processing in a studio. But again, not pertinent to end-users, who just want to listen.

I had the exact same argument at a VR video shoppe I was working at: the owner thought it was a good idea to try and send 6k or 8k video to VR headsets which could only replicate 3.5K resolution (11 pixels per degree, i.e. the Oculus Rift). It was an absurd situation. Instead of just pumping in more bits where they count: bitrate, they tried to get higher resolution through to the user's eyes than what the actual headset was capable of. It's just basic engineering: who does the filtering? And at what stage of the signal chain is it done at? If you do it anywhere post-studio, say, after you stream it, you're just wasting people's internet bandwidth or increasing download times for nothing. That data will never make it to the user's eyes if the headset can't deliver it.

Of course current VR helmets are nowhere near the limits of human visual acuity, unlike CD quality being more than what we need for our hearing, but the principle is the same: don't ship bits to the customer that they can't benefit from, instead spend that budget where it does matter such as at the exact right resolution but with more encoding bitrate (for video, that is). For audio things are MUCH simpler, and this is a solved issue. High-res audio is a complete and total sham and people pushing it are out to make $$$ from suckers who don't understand physics or sampling theory. CD quality is all you will ever need, period. Fact.

This is not complicated. Bits = dynamic range (how quiet the noise floor is relative to the maximum sound output). Sampling frequency = frequency response. Most humans can hear to about 20kHz when young and this generally goes down as you get older (not to mention people with hearing damage from say blazing away cheap ear buds at high volumes to overcome a noisy environment like an airplane without some type of noise cancellation and/or sound isolation (say in-the-ear phones) will lose it sooner.

Dynamic range is good...to a point. I personally do not want to listen to the quietest sound my ears can possibly hear and then suddenly switch to realistic levels of a jet engine nearby taking off (~120dB). In fact, I don't ever want to hear 120dB sounds PERIOD (we're talking instant hearing damage). In other words, just because you CAN do something (like say jump off a cliff), that doesn't mean you SHOULD.

It is very difficult in real world conditions and environments with real world microphones to get much more than 93dB of dynamic range. I've seen some recordings in very quiet concert halls achieve up to 98dB or so of dynamic range. 16-bit gets you 96dB. Certainly, 18-bit will cover just about any imaginable real world source and then some and with noise shaping, you can get that without true 18-bit playback. 24-bit is OVERKILL to the nth degree!

There is no reason to have it on the playback side of the chain, but every reason to have it on the recording side (for headroom so you don't have to have perfect input levels to achieve the best possible recording which you can later then cut back to 16-bit or 20-bit or whatever floats your boat.

Sampling. In the early days of digital audio, some early CD players used what we call "analog brick wall" filters. These sound bad as they typically screw your frequency and/or phase response far below the cut-off point. That's because analog filters aren't digital lines, but curves. To create a high enough order filter to brick wall just above 20kHz when CD response is cut off abruptly at 22.1kHz, let's just say it's not pretty. THIS is where all the high-end "audiophile" rags have pummeled CD Audio for ages.... There's just one problem. It's a load of HORSE CRAP!

Welcome to oversampling and digital filters! If you oversample an input signal (to say 96kHz or even 192kHz and by this, I mean up-convert a 44.1/48kHz type signal, not record at 96/192) and then apply a digital filter, that curve will nicely eliminate your problem frequencies with room to spare in the digital domain and when you convert back to an analog signal! High-end rags LOVE this idea for recording at 96kHz (ye old 24/96), but don't seem to comprehend the fact that it's not needed to achieve the same result in the DAC without the 96kHz signal! That's why you oversample! CD players came with 2x, 4x, 8x and even 16x internal oversampling). This eliminates the brick-wall filtering problem and leaves you with a nice phase-neutral cut-off at 22kHz. Problem SOLVED. 0-22kHz response with nary an issue!

Hey, welcome to even more crazy DAC designs! Oversampling worked, but why stop with one solution when you can use advertising to push other methods of achieving the same result! The most common one was 1-bit converters. These streamed all 16-bits in a CD player out at high speed (in serial port fashion, not dissimilar to modern day SATA hard drives versus PATA) and then reconstructed it, shoving all that filtering noise into the ultrasonic region where you can't hear it while putting the audible spectrum back where it came from. Isn't technology awesome? Audio rags would fuss about this and that, but darn it, they needed to sell advertising to people making $5000 DACs so NOTHING sounded as good as those $5000 models even though the frequency anaysis suggsted performance difference in the 0.05dB range or less (most humans have a hard time A/B telling 0.5dB differences, let alone 0.05! while most speakers are typically in the +/- 3dB range (that's up to 6dB differences) and room interactions at some frequencies can cause 8-11dB swings in some frequencies! Now YOU tell me where your $5000 should go! For that $5000 DAC or for $4500 speakers plus a $500 DIRAC room/speaker correction system? I know where I'd put my $5000.

Then there's the LP vs digital debate. It's a joke. LPs are more akin to 8-bit/32kHz digital if we're talking about USABLE frequencies plus vinyl loses its high frequency response over time as the needle drags through the grooves doing damage. Then there's clicks/pops.

YET, LPs CAN and sometimes DO sound better than some CDs. Why? The list of combination is long, but the most common reasons are mastering issues. People like LOUD music. So record companies force...ahem..."encourage" their mastering engineers to make music (especially of the "popular" variety) as LOUD AS POSSIBLE! That means compressing the dynamic range! So you take your 96dB a regular CD can do and you squeeze it down to maybe 10dB of dynamic range! You're now treading in the digital basement! 4-bit audio will do just fine! But records can only be compressed so much before the needle does crazy things like jump all over the record! So that Red Hot Chili Peppers album that has maybe 10dB dynamic range at most with square waves (clipping) all over the place in the REALLY loud spots (no dynamic range) might only be compressible down to say 25dB of dynamic range on an LP. Unbelievably, the record now has more dynamic range than the CD...a LOT more. No go back to older albums that actually used to TRY and get 60-70dB of dynamic range from a record and compare it to the new "remastered" CD that now gets 30dB of dynamic range and holy cow, that LP really sounds more like real life! This is why "digital" isn't cut and dried in practice, but all the problems are due to humans, not the format! Popular music is typically mastered for the least common denominator. Radio stations used compression boxes with records to make them louder for ages (you can hear it better on car radios!) and then with CDs and then CDs started getting made that were already loud and they'd compress them even more! Is it any wonder digital got a bad reputation when they were purposely making them sound like crap because that's what they think the average human WANTS to hear??? They COULD have just put a compression button on car stereos and left the albums alone, but hey, that would be a good idea! (you know something managers and bean counters know nothing about).

The gist is this. 24/96 is great on the recording side and means squat on the playback side. You still have iTunes selling compressed (as in lossy AAC, which is actually pretty good at 256kbps, but still technically a step DOWN from 1983 Red Book CD standards which is UNCOMPRESSED!) Instead of talking about 24/96 formats, maybe someone should just encourage the digital music sellers to make uncompressed 16/44 available online! I've got over 100mbps download speeds. The time difference between downloading an AAC file and a WAV/AIFF file is negligible and with 10TB of storage connected, so is the file size.

The next question, of course is whether "TrueHD" and "Master Audio" is audible or not. The answer is a probably at most times NO. DTS is pretty darn audibly transparent at home encoded rates. Even Dolby Digital at 640kbps sounds pretty darn good. But Blu-Rays have a lot of space to waste, so WHY NOT? But it's not something I'd worry about all night long whether I'm missing something or not with the DTS mix versus the DTS-HD Master Audio track when both are 5.1 (i.e. I don't bother when encoding my own and that saves me hard drive space; to get 7.1 to work correctly, I have to use Master Audio pass-through since programs like Handbrake won't encode in DD Plus 7.1 (it would be nice if it did) yet.

One other reason digital got a bad rap in the early days (pre-Loudness Wars) was due to simply not learning the lessons of a new format.

Analog recording required pre-emphasis: because analog recording rolled off the high frequencies, engineers compensated during the recording process by boosting the recording levels at the top end. It became as routine and as natural as breathing in the recording studio. When digital came out, engineers continued to artificially boost high frequencies for pre-emphasis (AKA the RIAA curve, in vinyl-land). Some early DACs even had "pre-emphasis" LEDs. Consequently, many early CDs had unbearable high frequency response, imparting the legendary "harsh" sound CDs were accused of.

But digital recording has no roll-off at high frequencies like analog does. Still, old habits die hard. Years after digital took over, I asked an engineer friend of mine if he was still boosting the high end when he recorded digitally. He said, "Of course." I asked why, and reminded him digital doesn't roll off at the top end. It was like he'd never even heard such a thing. And this guy was no dummy, he'd recorded a number of popular Chicago bands, including Smashing Pumpkins, and he still automatically recorded the way he always had with analog tape.

AES BLIND Test in a CONTROLLED ENVIRONMENT also found no significant AUDIBLE improvement in 16 vs 24-bit Audio. But that only compared TECHNOLOGY limitations audibility and did NOT compare the ENTIRE Recording/Mastering/Printing Process.

I ALWAYS prefer SACD/DSD and 24-bit PCM Audio Recordings on DVD-Audio, Blu-Ray, etc. to 16-bit CD's [many of which I have dupes]. And lower compression DTS5.1 to DD5.1, which I prefer vs CD's. PART of the improvement is that 16-bit CD's STILL vary quite a bit as to what was used in the Recording Chain....24-bit Mastering (incl. Mixing, Filtering + Signal Processing) is used much more frequently today....but that may or may NOT be true for the actual RECORDING Equipment....only on SACD/DVD-Audio Discs will I see that they are 24-bit THRU-OUT the Recording/Mastering process....look for the "DDD" Logo. And in older recordings, even when re-Mastered, it is likely you will hear some Tape Hiss [even if processed with an Adaptive Bandwidth Filter] and also minor Wow/Flutter defects.

Only those us of who remember the DIRECT-TO-DISC Vinyl Recordings will fully appreciate just how BAD the Recording/Mastering systems were back in the heyday of Vinyl....yes, even compared to Ampex 24-Track Recording Systems. Lincoln Mayorga and others ELIMINATED as much of the Filters, Compressors and other "Stuff" that was messing up the Audio.

A 16-bit Recording has just BARELY enough Dynamic Range....but ONLY if the Mastering Engineer is carefully Tweaking the Gain Control up and down....or adds a COMPRESSOR (may be only rolling off the peaks). With 24-bit Dynamic Range available, the Recording Engineer is less likely to "play" with the Dynamic Range and would be much less likely to add a COMPRESSOR/Peak-Limiter. Fortunately (Unfortunately???) only SOME people can TELL the difference between the Compressed "Fake" they have been hearing all their life and the much more "Live" and Impactful UNCOMPRESSED Music.

FYI: When Lincoln Mayorga visited our Local AES (Audio Engineering Society) Meeting back in the mid-70's, when his Direct-to-Disc Vinyl Recordings were being released, he claimed that he could HEAR whether the Initial Kick Drum OVER-Pressure Transient was being correctly reproduced...and whether the sound "SUCKED" due to one too many Signal Inversions in the Record/Reproduction System...which is VERY EASY to do, since most Amplifiers, Filters, et.al. INVERT the signal....if it DIDN'T Invert, the higher Output signal could leak into it's Input, causing Oscillation. Clearly there are SOME people (esp. professional musicians) who can hear subtleties NOT so audible to J. Q. Public.

More recently, the Transient and Envelope Delay Response of Loudspeaker Systems has been recognized as being one of these less obvious audible issues, explaining why Electrostatic and Planar-Magnetic Speakers have long been recognized as being head-and-shoulders "better" than most any other technology. BTW: Transient Response and Envelope Delay must ALSO be considered when comparing 16-bit Oversampling Filters....it is NOT just about Frequency Response. A similar argument also applies to Vacuum Tube Amps [which have low Transient Inter-modulation Distortion due to minimal Feedback] vs Transistor Amps with LOTS of Feedback [and have ever HIGHER Power so for OSHA Safe Levels they NEVER clip the waveform...and more efficient multi-driver Speakers also help] vs Tube Amps [which frequently ROUND-OVER rather than Clip the Peaks to provide that "Sweet"...but FAKE "old-timey" Sound].

Even today, there is less ATTENTION TO DETAIL when making 16-bit CD's, compared to Hi-Rez Audio releases....unfortunately recording engineers don't take the TIME to make sure the signals NEVER go into Clipping.....and as discussed above, they DO spend a lot of time playing with the record levels and/or use a COMPRESSOR to MAXIMIZE THE IMPACT....yes it sells....but we all know many are CRAP Recordings.

NOW as to the OTHER REASON why 24-Bit Audio is BETTER than 16-Bit CD's.....it's the ONLY way you will be able to hear Discrete SURROUND Recordings....which Maximize Impact in an entirely different way....feel free to either join the bandwagon....or keep listening to antique STEREO....perhaps extolling the virtues of your "realistic" Vinyl sound with plenty of Wow, Flutter, Clicks, Pops, Surface Noise and even more limited Dynamic Range....

I also wouldn't be surprised to learn that there are perhaps a HANDFUL of people out there who bought a COMPRESSOR Box so they could listen to the STEREO Track on recent DVD/BR Movies so they have more of that CD or Vinyl "QUALITY" to the sound......

And technically speaking, it shouldn't given the limitations of real world environments and microphones. Personally, I'd never want more than 100dB of dynamic range period (given even quiet rooms, you'd have to have peak volumes over 110dB to achieve even that much in practice and that's too darn loud for anything but a brief cannon in the 1812 overture or something).

You realize that's because they're actually TRYING to get good quality sound, right? Take one of those recordings (as long as its only stereo) and convert it to 16/44.1 CD and it'll sound identical. The same is true of these "fantastic vinyl LPs" that are supposedly out there. Record them to CD. They'll sound identical. Personally, I run them through iZotope RX after recording from LP. Not only can it remove all click/pops with nary any negative effect on the music, but with its learning noise removal features, I've actually gotten LPs to sound quieter with less master tape noise than the CD versions in some cases (e.g. Tori Amos' Little Earthquakes and Under The Pink albums).

24-bit mastering doesn't really mean a darn thing. It's purely for recording headroom to make the mastering engineer's life easier (one take and you're done instead of risking clipping). Once you're done, a downmix to 16/44 is fine.

Some of my favorite recordings are analog (e.g. Most of the Pink Floyd catalog like Dark Side of the Moon, Wish You Were Here, The Wall, etc.) and they sound GREAT (I have vinyl versions of most of their albums as well and a high-end turntable setup as well in my Carver ribbon speaker room). A little tape hiss doesn't ruin life. Even guitar amplifiers have their own noise levels. When I made my last album I avoided all external guitar pedals, amps, etc. and input the Fender Strat straight into Firewire interface box and let Logic Pro do all the guitar processing. The only analog input I had was the vocals and a microphone for acoustic guitar. Everything else was digital piano/synth/drum kits. I made it at my house and it sounds like Pink Floyd quality sound at a tiny fraction of the price to make.

I don't know that it's a matter of taking the time to do it. It's EASY to avoid clipping entirely. In 24-bit mastering, you have tons of headroom and when you're done, you can just let the computer "normalize" the track and it will make the loudest sound the maximum amplitude (which you can offset however much you want below 0dB) and if there's some loud noise in an instrument causing issues, it's easy to go in with an editor like iZotope and repair/snip the waveform as long as it's small.

My point is that it's not hard to avoid clipping, but some of these guys DON'T CARE (or their bosses order them to make it louder to the point it's almost unavoidable). The worst offender I can think of that I've heard commercially is The Red Hot Chili Pepper's Californication album. It hits the rails so often I think I'm going to vomit, but sure enough in a car environment, much of it disappears due to road noise, etc. The sad thing is someone released a pre-master version and it's 85% improved, but at least SOME of the clipping was the band's raw recordings not being properly handled to begin with, not at the mastering stage. You can't fix a bad input gain setting after the fact very well. That album will never be a high-end show piece for that reason.

But take Tori Amos' 1996 Boys For Pele album that was almost entirely recorded straight to DAT tape at 16-bit 48kHz (save the drums and possibly some backup vocals which were recorded on analog tape). That album was my #1 pick for a long time to test speaker systems for realism because it sounds absolutely REAL on many of the tracks on quality speakers. I use dipole ribbons in my high-end living room setup and her voice floats in three dimensional space to the point where with my eyes closed, I'd swear she was in the room. There'll never be a 24/96 version of that album since it was recorded at 16/48 on DAT. But other than two brief moments where it briefly tags the rails for a split second on "Caught a Lite Sneeze" there's no other clipping on the album (an alternate take of that track I have on a box set release doesn't have the clipping, but isn't quite as emotionally impacting at that point either; I'd rather deal with the tiny clip). But the album is stunning and proves 24/96 means nothing.

What band wagon? Surround sound music is mostly dead (even more so than 3D IMO) as most people just aren't that interested in it. The reasons should be mostly obvious.

Most people don't even have 5.1 setups and even those that do usually don't set it up properly. And even then how many people want to sit there at home and listen to music? Most people listen to music while they're on a computer or walking with earphones on a treadmill, etc. And even those that do listen regularly at home on a quality system might prefer stereo since out of the few dozen surround recordings I have, only about 10% are "really good". Many do very little with the surround channels (sometimes Dolby's Logic IIx Music mode sounds better with the stereo track extracted to surround than what the engineers did with the surround remix). Now Pink Floyd's DSOTM and WYWH are excellent in surround as is Alan Parson's ON AIR. Sheryl Crow's The Globe Sessions (in 6.1) is pretty good as is Dire Strait's Brothers In Arms and The Eagle's Hotel California mixes. But something like Billy Joel's 52nd Street or The Police's Every Breath You Take weren't very good surround mixes, IMO.

But again, it's not 24/96 that means a darn thing. It's the extra surround channels. DTS CDs were typically recorded in 16/44.1 and they sounded great. You don't need SACD or DVD-Audio to achieve excellent surround. Any of them can be converted to something like FLAC at any bit/frequency setting you like.

I also throughly enjoy playing my 70's Music....esp. if it's been Re-Mastered. But with a Noisy, Distorted Fender, how can you TELL if the Recording Chain is adding Distortion????? OTOH, it's much easier to tell listening to say John Williams playing his Classical Guitar....ditto Kimo West or Sonny Chillingsworth playing Hawaiian Slack Key Guitar....

In today's world, clipping of the Microphone...and the Mic's Preamp is STILL a problem....GIGO applies....and some Tech is a LOT better than older Tech....

Surround Music is still available...but it moved to FLAC Downloads and Blu-Ray....esp. Concerts and Movie Sound Tracks....

Very interesting comments!

MagnumX :
"Most people don't even have 5.1 setups and even those that do usually don't set it up properly. And even then how many people want to sit there at home and listen to music? Most people listen to music while they're on a computer or walking with earphones on a treadmill, etc. And even those that do listen regularly at home on a quality system might prefer stereo since out of the few dozen surround recordings I have, only about 10% are "really good"."
So true. I have a home quality system and it took me 2 years to understand a little bit about frequency response in small room, speakers placement, bass trap and first reflexion treatment etc. It is not perfect at all but man ! how fun it is to take time to listen music carefully, eyes closed !

Well, if I can't tell (don't notice a problem), I don't worry about it. I can just listen to and enjoy the music.

Well, I'm sure there are still SACD classical releases as well, but personally I'd just be happier if music quality from the industry improved in general. I don't really care about the formats (name brands, company competition such as SACD vs DVD-Audio, etc.) so much. I think we'd be better off if they just standardized something good and actually used it. I will say I've heard "compressed" (in terms of dynamic range) that still sounds nice and clear/clean and is fine for many types of rock music. The digital tools today can achieve these effects far more cleanly than they used to. Whether it's a good artistic choice or not is another matter.

When capturing/recording audio tracks it is benifical to use greater bit depth. It's not that you have more headroom...headroom above reference is still ~20 dB.

When one captures at higher bit depths, the DAW processing (edit/mixing) is effectively more transparent. The therotical dynamic range is irrelevant , a 24-bit signal properly dithered to 16-bit for release format is perceptually equivalent. 20-bit @ 48 kHz is a very robust signal to process/mix with.

http://lavryengineering.com/pdfs/lav...ing-theory.pdf

This is a excellent article debunking the "digital steps" myth.