IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Is 24bit/44.1 kHz ALAC a good hi-res archive compromise, or inherently
MilMascaras
post Oct 5 2010, 22:15
Post #1





Group: Members
Posts: 5
Joined: 5-October 10
From: L.A.
Member No.: 84368



Is 24bit/44.1 kHz ALAC a good hi-res archive compromise, or inherently flawed?

Hello:
In 2008/2009, I abandoned .mp3 and .m4a for archiving my personal music collection.
I have converted exclusively to lossless for archival purposes, and have now rebuilt an amazing 20K+ song album library, favoring the best masters I could muster.

(For the past few years I have learned everything I could about EAC, Red Book, WavPack, WAV, AIFF, SHN, APE, FLAC, and ALAC. I have settled on FLAC and mostly ALAC (tag-ability). My primary tools on Macintosh are: Max, xACT, and XLD. Gratefully, I am pretty well versed on these concepts…)
====
BUT, now I am seriously working on my High Resolution archive, including my (vinyl) needle drop collection, with some primo rips, captured as 24bit/96kHz FLAC's.

I have been experimenting with converting this source material to 24bit/44.1 kHz ALAC, because of its portability and universality. The Beatles USB 2009 Box Set was done in 24bit/44.1 kHz FLAC, and the Neil Young Archives at 24bit/96kHz, and these were my inspirations for this compromise 'experiment'.

I am taking a 24-bit / 96 kHz FLAC, then converting them with XLD, choosing ALAC at original 24 bit bitrate, but choosing the sample rate at 44.1 kHz.

I end up with ALAC's at about 1600+ kbps. I don't necessarily want to enforce Red Book, I want a higher resolution file that will play on just about on anything; Not all my devices can play my FLACs..

They play PERFECTLY and awesomely on my iPhone, my music server to pioneer elite AV receiver via TOSLINK, and all my computers, and most every portable device I have thrown them on so far. I do realize that some devices only pass-thru 16bit though.
====
So I guess I get to the real question:
Is there any inherent flaw in my methodology that I am overlooking?
I realize that the 24/96's (and 24/192's) are better/purer to the source, but what are the OTHER downsides of saving high resolution files at 24bit/44.1 kHz?

Hoping for some last-minute insights before I commit to this project…
Thank You.

This post has been edited by MilMascaras: Oct 5 2010, 22:16
Go to the top of the page
+Quote Post
greynol
post Oct 5 2010, 22:22
Post #2





Group: Super Moderator
Posts: 10241
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Provided your peaks aren't 30dB down from full-scale, you'll be hard-pressed demonstrating any tangible benefit of 24-bit over 16-bit as a delivery format for vinyl.

BTW, I moved this from the CD HW/SW forum since this topic is not about ripping.

This post has been edited by greynol: Oct 5 2010, 22:26


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
MilMascaras
post Oct 5 2010, 22:56
Post #3





Group: Members
Posts: 5
Joined: 5-October 10
From: L.A.
Member No.: 84368



Hi & Thank You.
Not limited to vinyl, rather I am talking about High Resolution SOURCE material that is ALREADY meticulously captured at 24bit/96 kHz;

I am wondering what I might be overlooking (besides obvious kbps reduction) in converting this type of source matierial to 24bit/44.1 kHz ALAC for permanent archival purposes.
Trying to find the optimal "compromise" that will play on just about anything, especially on playback devices that may not natively support FLAC and/or 96 kHz + files.

I appreciate any and all insights...

This post has been edited by MilMascaras: Oct 5 2010, 22:58
Go to the top of the page
+Quote Post
Brand
post Oct 5 2010, 23:11
Post #4





Group: Members
Posts: 321
Joined: 27-November 09
Member No.: 75355



One thing that I thought of is you might want to do the downsampling properly. Not all software does this in the same way, some are better than others.

Here's one comparison page of some of them: http://src.infinitewave.ca/

Even the free version of r8brain does a better job than some expensive programs, for example..
Go to the top of the page
+Quote Post
greynol
post Oct 5 2010, 23:13
Post #5





Group: Super Moderator
Posts: 10241
Joined: 1-April 04
From: San Francisco
Member No.: 13167



It's up to the hardware you choose, I suppose.

I apologize for steering this towards a debate about whether so-called "benefits" of hi-res audio delivery formats are actually tangible; unless you wish to go there, of course.

EDIT: I had not seen Brand's excellent response prior to posting this. AFAICT, this is the only reason why anyone has been able to differentiate CDDA that makes use of all 16 bits from hi-res with music played under normal conditions (read: full scale does not cause damage to your hearing).

This post has been edited by greynol: Oct 5 2010, 23:20


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
DVDdoug
post Oct 5 2010, 23:27
Post #6





Group: Members
Posts: 2672
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
I am wondering what I might be overlooking (besides obvious kbps reduction) in converting this type of source matierial to 24bit/44.1 kHz ALAC for permanent archival purposes.
Trying to find the optimal "compromise" that will play on just about anything, especially on playback devices that may not natively support FLAC and/or 96 kHz + files.
"Archiving" typically means that you want to preserve each and every original byte, and you can't do that while going from 96 to 44.1.

Some people keep a high-quality lossless archive plus an "everyday" portable copy in a more-compressed, and perhaps more compatible, format.

For listening & compatiblity you won't hear any difference. So, if you want to keep the downsampled copy as your permanent archive copy, that's up to you. Heck, you can create a high quality MP3 archive if you wish... You won't hear any difference with most songs. Once you decide to "compromise", you'll have to decide for yourself how much to compromise.

QUOTE
Trying to find the optimal "compromise" that will play on just about anything,
MP3 & AAC are the most universal formats that will play on anything, but of course they are lossy. The nice thing about lossless formats is that you can convert to a different lossless (or lossy) format at any time.... If you have an iPod today, you can use ALAC. If your iPod dies and you get a different player that only supports FLAC or MP3, you can convert it.
Go to the top of the page
+Quote Post
Brand
post Oct 5 2010, 23:41
Post #7





Group: Members
Posts: 321
Joined: 27-November 09
Member No.: 75355



QUOTE (greynol @ Oct 6 2010, 00:13) *
EDIT: I had not seen Brand's excellent response prior to posting this. AFAICT, this is the only reason why anyone has been able to differentiate CDDA that makes use of all 16 bits from hi-res with music played under normal conditions (read: full scale does not cause damage to your hearing).

Did you mean bit reduction? Because I was talking about downsampling (and possible SRC artifacts).
Go to the top of the page
+Quote Post
greynol
post Oct 6 2010, 00:16
Post #8





Group: Super Moderator
Posts: 10241
Joined: 1-April 04
From: San Francisco
Member No.: 13167



I was referring to both bit rate and sample rate, though I do see that you were only addressing sample rate.


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
MilMascaras
post Oct 6 2010, 01:26
Post #9





Group: Members
Posts: 5
Joined: 5-October 10
From: L.A.
Member No.: 84368



Thank You, that is very helpful. Yes, I guess it is about downsampling...

And I surmise XLD uses QuickTime/Core Audio to downsample/convert to 44.1 kHz.
(I notice people use 'dither' when converting/downsampling 24/96 to Red Book 16/44.1)

What I had been previously doing is keeping/maintaining my high resolution rips as 24/96's (and 24/192's) in FLAC as my permanent archive, and simultaneously making an ALAC in 16/44.1 for iPod/iPhone & iTunes library.
When I referred to archiving, I guess I am misused the term: I want to slim down the multiple copies of tracks, and make the most universal copy I can make, and shift my library primarily to ALAC.

Then the Beatles USB boxset came out in 24bit / 44.1 kHz FLAC, and this choice piqued my interest...

I then did multiple A/B's between the Beatles Abbey Road 2009 stereo remaster CD (16/44.1) and the Beatles USB remaster source converted to 24bit/44.1 kHz ALAC (converted from FLAC to ALAC with XLD).
Son of a gun, it sounded better to me, I always picked the 24bit!
And I also A/B'd 24/44.1 ALAC's against 24/96 *ALAC's*, and it was effectively indistinguishable to me. It felt like I was on to something...

So now, I am intrigued, and am considering converting some/most of my 24bit/96kHz FLAC's to 24bit/44.1kHz ALAC as my ONE Permanent copy in my iTunes library and music server, as just about every music device I am playing these high resolution rips seem to support these ALACs. Yes, I am downsampling, but still technically maintaining it as a high resolution format, s'il vous plaît.

So far, my results have been promising, and I now favoring converting all my 24bit/96kHz FLAC's to 24bit/44.1kHz ALAC's, in order to keep ONLY 1 high resolution format in my permanent library.
I hope I don't regret this choice in a couple years, I think I am going to take the plunge... (Fingers crossed...)
Go to the top of the page
+Quote Post
saratoga
post Oct 6 2010, 03:11
Post #10





Group: Members
Posts: 5116
Joined: 2-September 02
Member No.: 3264



QUOTE (MilMascaras @ Oct 5 2010, 20:26) *
Then the Beatles USB boxset came out in 24bit / 44.1 kHz FLAC, and this choice piqued my interest...


You may be interested in this thread:

http://www.hydrogenaudio.org/forums/lofive...php/t77630.html

In there it was determined that the 24 bit version and the 16 bit version are identical aside from a slight volume difference.

QUOTE (MilMascaras @ Oct 5 2010, 20:26) *
I then did multiple A/B's between the Beatles Abbey Road 2009 stereo remaster CD (16/44.1) and the Beatles USB remaster source converted to 24bit/44.1 kHz ALAC (converted from FLAC to ALAC with XLD).
Son of a gun, it sounded better to me, I always picked the 24bit!


A/B or ABX? A/B probably won't give you good results for this. See here if the former:

http://www.hydrogenaudio.org/forums/index....opic=16295&

If the latter, it may be due to some problem in XLD.
Go to the top of the page
+Quote Post
2Bdecided
post Oct 6 2010, 19:39
Post #11


ReplayGain developer


Group: Developer
Posts: 5251
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



QUOTE (MilMascaras @ Oct 6 2010, 01:26) *
Then the Beatles USB boxset came out in 24bit / 44.1 kHz FLAC, and this choice piqued my interest...

I then did multiple A/B's between the Beatles Abbey Road 2009 stereo remaster CD (16/44.1) and the Beatles USB remaster source converted to 24bit/44.1 kHz ALAC (converted from FLAC to ALAC with XLD).
Son of a gun, it sounded better to me, I always picked the 24bit!
I'm not surprised - it's louder.

http://www.hydrogenaudio.org/forums/index....st&p=679888

Cheers,
David.
Go to the top of the page
+Quote Post
2Bdecided
post Oct 6 2010, 19:44
Post #12


ReplayGain developer


Group: Developer
Posts: 5251
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



QUOTE (greynol @ Oct 5 2010, 23:13) *
I apologize for steering this towards a debate about whether so-called "benefits" of hi-res audio delivery formats are actually tangible; unless you wish to go there, of course.
You're getting mild in your older age greynol wink.gif

TBH, how can anyone give a sane answer to the question without considering the benefits or otherwise of increased sample rate and/or bitdepth?

That said, if I'd paid money to buy 24/96, that's how I'd keep it. I'd probably make a 16/44.1 copy (and burn it to a CD, and encode it to mp3) for regular use on other devices. 20/48 arguably theoretically makes some sense.

Cheers,
David.
Go to the top of the page
+Quote Post
Wombat
post Oct 6 2010, 20:31
Post #13





Group: Members
Posts: 1083
Joined: 7-October 01
Member No.: 235



Sorry if i donīt contribute much to that thread but i must mention one footnote made in the thread that was linked to.

Quote: 2Bdecided
P.S. I find it amusing and depressing that there's all this talk of how much better it does sound at 24-bits, and could sound at a higher sample rate - as if 44.1kHz 16-bits was the limiting factor on these recordings! For goodness sake...

I canīt count on how many places i get confused with such posts. As if 24bit is easily to hear better on everything like night and day. You canīt even argue well enough outside Hydrogen about this cause to many jumped that mass-paranoia and *believe*

Edit: Not seldom these people are not impressed by numbers. At least no numbers that tell them 16bit is enough. But only knowing they hear a 24bit version most likely does...

This post has been edited by Wombat: Oct 6 2010, 20:38
Go to the top of the page
+Quote Post
greynol
post Oct 6 2010, 20:35
Post #14





Group: Super Moderator
Posts: 10241
Joined: 1-April 04
From: San Francisco
Member No.: 13167



48kHz makes perfect sense for video applications or hardware that cannot play 44.1kHz and does a poor job converting to 48kHz. I don't see any theoretical advantage in going over 44.1 as a delivery format, however.


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
Northpack
post Oct 6 2010, 23:47
Post #15





Group: Members
Posts: 455
Joined: 16-December 01
Member No.: 664



QUOTE (greynol @ Oct 6 2010, 20:35) *
48kHz makes perfect sense for video applications or hardware that cannot play 44.1kHz and does a poor job converting to 48kHz. I don't see any theoretical advantage in going over 44.1 as a delivery format, however.

I think the step from 44.1khz to 48khz may be the only improvement which could actually make an audible difference in some (very rare) cases for some (very few) people. Some years ago a was able to hear up to 22khz, unfortunately I never did an ABX test then and now it has gone down to 19khz. I guess I'm getting older...

This post has been edited by Northpack: Oct 6 2010, 23:48
Go to the top of the page
+Quote Post
saratoga
post Oct 7 2010, 00:29
Post #16





Group: Members
Posts: 5116
Joined: 2-September 02
Member No.: 3264



QUOTE (Northpack @ Oct 6 2010, 18:47) *
QUOTE (greynol @ Oct 6 2010, 20:35) *
48kHz makes perfect sense for video applications or hardware that cannot play 44.1kHz and does a poor job converting to 48kHz. I don't see any theoretical advantage in going over 44.1 as a delivery format, however.

I think the step from 44.1khz to 48khz may be the only improvement which could actually make an audible difference in some (very rare) cases for some (very few) people. Some years ago a was able to hear up to 22khz, unfortunately I never did an ABX test then and now it has gone down to 19khz. I guess I'm getting older...


Unless you were very careful with that 22khz test, you might just have been hearing harmonics. I once found I could "hear" 25kHz, until I got a piezo microphone and realized that just because I was playing a pure 25khz tone didn't mean thats what was coming out of my speaker!

Edit: well I suppose technically it wasn't a harmonic, but rather some other kind of distortion.

This post has been edited by saratoga: Oct 7 2010, 00:29
Go to the top of the page
+Quote Post
Wombat
post Oct 7 2010, 00:32
Post #17





Group: Members
Posts: 1083
Joined: 7-October 01
Member No.: 235



QUOTE (Northpack @ Oct 7 2010, 00:47) *
QUOTE (greynol @ Oct 6 2010, 20:35) *
48kHz makes perfect sense for video applications or hardware that cannot play 44.1kHz and does a poor job converting to 48kHz. I don't see any theoretical advantage in going over 44.1 as a delivery format, however.

I think the step from 44.1khz to 48khz may be the only improvement which could actually make an audible difference in some (very rare) cases for some (very few) people. Some years ago a was able to hear up to 22khz, unfortunately I never did an ABX test then and now it has gone down to 19khz. I guess I'm getting older...


Donīt mix hearing a sweep up to 22kHz means you can hear it in music content. 10 years back when i also could hear very high with sinuid tones a lowpass used for mp3 much away from my highest tones i could hear didnīt let me miss anything in a normal music content. Surprisingly.
btw. did anyone remember "baddudex" smile.gif most likely he could

Back to topic. I just have read WD will bring its 3TB drives. Well, i doubt there is any reason not to keep a backup of the data how you purchased or ripped it these days.
For daily use 16/44.1 might be the most compliant on all platforms, even the most exotic.
Go to the top of the page
+Quote Post
greynol
post Oct 7 2010, 03:13
Post #18





Group: Super Moderator
Posts: 10241
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Regarding bit-depth, I think it is worthwhile to revisit this post:
http://www.hydrogenaudio.org/forums/index....st&p=428452

Regarding whether this is on-topic, the OP asked "what are the OTHER downsides of saving high resolution files at 24bit/44.1 kHz?" My reply is wasted space and I'm trying to provide information to help indicate why I feel this way.


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Oct 7 2010, 04:05
Post #19





Group: Members
Posts: 4222
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (Northpack @ Oct 6 2010, 18:47) *
Some years ago a was able to hear up to 22khz, unfortunately I never did an ABX test then and now it has gone down to 19khz. I guess I'm getting older...


One of the counterintuitive things about hearing is that being able to hear pure tones up to say 22 KHz has little bearing on whether or not you can hear frequncies at 22 KHz or above beign removed from music.

The reason is masking. Masking is a property of the human ear where the loudest sounds in what is known as a critical band wipe out the ability to hear softer sounds in the same critical band. For example, 1000 and 1100 Hz are both in the same critical band. If one tone in the same critical band is appreciably louder than another, your ability to hear the softer sound will be far less than if the tones were in two different critical bands. For example 1000 and 2000 Hz are in different critical bands.

The ar loses sensitivity very dramatically at very high frequences, such as those above 10 KHz. Most musical instruments when playing high notes create sounds where the higher harmonics are softer than the lower harmonics. There is only one critical band above 15 KHz or so. The lower frequencies in this critical band will tend to be louder than the higher frequencies and are likely to mask them. Once masked, you can't hear them. Therefore sounds you can hear as pure tones will be inaubile when portrayed naturally, such as how they are usually parts of music.


Go to the top of the page
+Quote Post
uzernaam
post Oct 7 2010, 22:14
Post #20





Group: Members
Posts: 3
Joined: 7-October 10
Member No.: 84405




Since you said you're going from 24/96 to 24/44.1 what I would be concerned about would be artifacts of the anti-aliasing when changing sampling frequency. Even the best methods out there aren't perfect, and somewhere deep in the sound they will still be lurking. An "archive" usually employs a storage method that is intent on minimizing any possible sources of distortion.

Sounds like your 24/96 is the archive. Anything you wish to do to copies of it is up to you. If the sound quality of your portable copies is still important, you could continue with the 24/44.1, but you also mentioned portability being a concern. For me, the #1 most portable audio format is mp3. Granted, it has its flaws but you've always got your archive to turn to if you need that perfect copy.

If on the other hand you're requirements are the best sound possible that plays on all your equipment, sounds like you found the right compromise. Just be aware that going 96>44.1 can do some funny things and that some of your equipment might truncate 8 of those bits.
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Oct 8 2010, 00:34
Post #21





Group: Members
Posts: 4222
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (uzernaam @ Oct 7 2010, 17:14) *
Since you said you're going from 24/96 to 24/44.1 what I would be concerned about would be artifacts of the anti-aliasing when changing sampling frequency.


Resampling audio isn't rocket science, and doing it right isn't magic. There are a goodly number of resamplers out there that do downsampling without generating aliases that are above their very low residual noise levels.

Here are some technical test results of common resamplers:

Infinite wave SRC tech tests


Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 21st November 2014 - 11:42