"Lossless Audio Checker"?

[mindstormer]

I was trying to find an application that detects whether a lossless file is truly lossless because I have some reasonable doubts that some of my "lossless" files are not really lossless. Upon googling, I came across Lossless Audio Checker (http://losslessaudiochecker.com/). I was wondering if this tool is reliable given its application description and whether the method actually works.

I've used it and it confirmed that the lossless file I was suspicious of actually isn't lossless and files that I thought were lossless were detected as lossless. Is it accurate though? Is this application reputable? Is there a better way of checking whether lossless files are actually lossless? This is important because the point of a FLAC file is to sacrifice size for quality.

Thanks!

P.S. I was thinking of getting a pure FLAC collection (with each file checked so that they are actually 100% lossless) and then convert them to AAC or MP3 based on my standards. I want them to get the best qualityize ratio but with much greater emphasis on quality... I heard that high quality MP3/AAC files are significantly smaller than FLAC files but with negligible impact on quality. What application do I use do to this? LAME?

[poisondeathray]

The only way to be certain if something is lossless is if you have the original file to compare it to . You can do a md5 checksum of the decoded output and compare , or similar methods like compare waveforms

If this application only checks 1 file , not 2 files , it's not necessarily accurate

by detecting cutoff frequencies. It can show you the full frequency range for each file.

The problem with this method is clipping and cutoff can occur in the "original" file (e.g. a concert recording) , so the presence of this alone does not determine if it is "lossless" . By definition the only way you can call something lossless is by comparing it to the original file

Some FLAC encoders might apply dithering, so it's not necessarily lossless if the processing pathway applies other processing (e.g. audacity does this by default unless you disable it)

[jagabo]

You can open two audio tracks in an audio editor, invert one of them, and mix the two together. If they are identical the result will be complete silence.

[poisondeathray]

P.S. I was thinking of getting a pure FLAC collection (with each file checked so that they are actually 100% lossless) and then convert them to AAC or MP3 based on my standards. I want them to get the best qualityize ratio but with much greater emphasis on quality... I heard that high quality MP3/AAC files are significantly smaller than FLAC files but with negligible impact on quality. What application do I use do to this? LAME?

It depends how good your hearing is, your equipment, source quality

Audiophiles will cringe at suggesting AAC or MP3 for anything besides the garbage can . These lossy compressed formats will cut off high frequencies regardless of bitrate used. You can see this on a spectral audio graph in any audio editor. But whether or not you can tell by listening is another matter

Personally I can't tell at high bitates - it's good enough for me . You have to do your own tests to see what works for you

[mindstormer]

Cool. Thanks for the info, guys!

[jagabo]

I heard that high quality MP3/AAC files are significantly smaller than FLAC files but with negligible impact on quality.

It depends how good your hearing is, your equipment, source quality

I have never seen any reliable reports of anyone being able to consistently pass double blind ABX testing of high bitrate MP3 and the original uncompressed PCM. Of course, every audiophile says 320 kbps MP3 sounds like trash. But then they can't pass the test. Most will refuse to take it.

[Mephesto]

I heard that high quality MP3/AAC files are significantly smaller than FLAC files but with negligible impact on quality.

It depends how good your hearing is, your equipment, source quality

I have never seen any reliable reports of anyone being able to consistently pass double blind ABX testing of high bitrate MP3 and the original uncompressed PCM. Of course, every audiophile says 320 kbps MP3 sounds like trash. But then they can't pass the test. Most will refuse to take it.

I have done a rigorous test on a sample I was very familiar with and have heard since kindergarten. I could ABX MP3 at any CBR bitrate even above 320 kb/s. VBR MP3 became transparent at 256 kb/s but I correctly discerned 192 kb/s 35 out of 50 times.

AAC I could ABX up to 144 kb/s but couldn't do 160.

The difference that I ABX'd at these high bitrates was one isolated section of the audio where you hear a slight click when a heavy note is struck and gets more pronounced on the encoded MP3/MP4. At 160 kb/s AAC even that click was perfectly retained like the original as well as on 256 kb/s VBR MP3.

Audiophiles are straight-up toaster-******* bullshitters if they say 320 kb/s MP3 is trash.

[jagabo]

The difference that I ABX'd at these high bitrates was one isolated section of the audio where you hear a slight click when a heavy note is struck and gets more pronounced on the encoded MP3/MP4.

Yes. I'm sure it's possible to design waveforms where it's possible for many people to detect a difference between high bitrate MP3 and the original audio. Just like it's possible to design video test patterns that show the weaknesses of video filtering, compression, etc.

[Mephesto]

The difference that I ABX'd at these high bitrates was one isolated section of the audio where you hear a slight click when a heavy note is struck and gets more pronounced on the encoded MP3/MP4.

Yes. I'm sure it's possible to design waveforms where it's possible for many people to detect a difference between high bitrate MP3 and the original audio. Just like it's possible to design video test patterns that show the weaknesses of video filtering, compression, etc.

No, this was a legit song. I didn't design it nor foresee what part of the song will be a weakness for the encoders. I picked it because it was short and I was very familiar with it. It turned out a particular part of it was the only part left that I could discern a difference as the rest of the song became transparent as the bitrate crept up. Flanging and smearing artifacts ceased, top-shelf was no longer gone so dullness ceased, pre-echo stopped being noticeable except on that one note where it accentuated the little click that was already a bit audible. At high enough bitrates, even that was gone.

[deadrats]

i don't know what you guys are talking about, i can definitely tell the difference between a high quality flac and a high bit rate mp3 but then again i use head phone when i want to listen to music.

i bought a load of music back in the good old allofmp3.com days and at first i was buying high bit rate mp3's until one day i decided to download some of the songs i had already downloaded only this time in lossless compressed format and the difference is pronounced.

try it yourselves, if you have a cd of some piece of music with lots of range, something like a Mozart or Beethoven or one of the star wars sequences with lots of violin or even something like a rock opera like maybe meat loaf's bat out of hell or similar genre, rip the cd to both flac and 320kps mp3 and listen to both of them with some decent head phones plugged directly into a decent sound card, you'll hear the difference.

[Mephesto]

i don't know what you guys are talking about, i can definitely tell the difference between a high quality flac and a high bit rate mp3 but then again i use head phone when i want to listen to music.

i bought a load of music back in the good old allofmp3.com days and at first i was buying high bit rate mp3's until one day i decided to download some of the songs i had already downloaded only this time in lossless compressed format and the difference is pronounced.

try it yourselves, if you have a cd of some piece of music with lots of range, something like a Mozart or Beethoven or one of the star wars sequences with lots of violin or even something like a rock opera like maybe meat loaf's bat out of hell or similar genre, rip the cd to both flac and 320kps mp3 and listen to both of them with some decent head phones plugged directly into a decent sound card, you'll hear the difference.

I used decent headphones in the testing as well. And music on a CD goes thru a few generations of conversions (which pro-analog audiophiles complain about) to get on that CD in the first place. My piece was from a game so it was testing against the real original source.

Do an ABX double-blind test, even 320 kb/s MP3 at CBR is extremely difficult to discern let alone VBR. I guarantee you will fail the test badly.

[jagabo]

i bought a load of music back in the good old allofmp3.com days and at first i was buying high bit rate mp3's until one day i decided to download some of the songs i had already downloaded only this time in lossless compressed format and the difference is pronounced.

That's meaningless as you have no way of knowing how poorly those MP3 files were made. They may have intentionally been degraded so those buying flac files would think they got something for the addition cost.

And really, allofmp3.com? The most notorious mp3 ripoff site on the net? What would you expect from them? I wouldn't be surprised if their flac files were defective too!

[jagabo]

No, this was a legit song.

Yes, I understood. I was pointing out that there are specific sounds that won't be handled well by a lossy codec. And that any random waveform segment may appear occasionally in real music. So you may hear a difference in MP3 files on rare occasions. But if that difference is only audible as 1 click in 1 out of 1000 recordings it's not important.

[hech54]

Audiophiles are straight-up toaster-******* bullshitters if they say 320 kb/s MP3 is trash.

This is going to be my new sig.

[Mephesto]

No, this was a legit song.

Yes, I understood. I was pointing out that there are specific sounds that won't be handled well by a lossy codec. And that any random waveform segment may appear occasionally in real music. So you may hear a difference in MP3 files on rare occasions. But if that difference is only audible as 1 click in 1 out of 1000 recordings it's not important.

I agree the loss is not important, just pointing out it isn't transparent and can be detected by human ears, especially the notoriously crappy CBR of MP3 that fails at every bitrate. For majority of cases, 160 kb/s MP3 and 112 kb/s AAC is HQ enough.

This is going to be my new sig.

In that case, let me...

Audiophiles are straight-up toaster-******* bullshitters in general, especially Hydrogenaudio assgoblins who think they are somehow different.

There, fixed.

[Ethlred]

Hear him!
Hear him!

Toaster-******* bullshitters is right up there with 'Take a flying **** at at rolling doughnut' for over the top invective.

Now if only Yahoo would allow that or even Assgoblins I would use it over Troll. I love the smell of burning troll in the morning that that assgoblin smell it smells like ... like victory.

Copied to my Repost collection for that special moment in the future on sites that allow the artistically potty mouthed posts and flamewars. Baldrick does not like flamewars. I do but I respect his desires. Mostly.

[Hoser Rob]

Audiophiles are straight-up toaster-******* bullshitters if they say 320 kb/s MP3 is trash.

This is going to be my new sig.

Actually, with old crappy recordings ... and that's not all old recordings by any means, you can't get much better than Mercury Living Presence or Dark Side of the Moon ... mp3 320K is fine.

And with most recordings made in the last 10 years or so, they're so compressed (maybe 1-2dB of dynamic range) you may as well use mp3 too. Look up "loudness wars". It stinks. I don't expect everything to be audiophile quality but there are good songs I wouldn't buy because they sound so bad.

Dithering the signal like Audacity does may not be bit perfect but it's a subtle effect. I think maybe the OP dl'd some files that someone converted from a lossy codec to flac, probably misguidedly thinking that would improve the sound. It won't.

[raffie]

Any lossy audio compression ditches a whole lot of information from the original file and while it may sound fine on it's own, the second you listen to the original (on descent speakers, no 4 ohm crap for starters) you'll hear things you didn't hear before.

That's what for example MP3 does, it throws stuff away, detail, to be exact. If you only have the MP3 there's no way you'll miss what you don't know should be there and you think you heard everything there is to hear.
You've been tricked. Simple as that.

[Cornucopia]

Yes, we know it's a trick. We know things are lost. We are all willing to go along with the trick because of the great space/time/cost savings and because SOME/MOST of the time we don't notice the difference, though difference there is. Once again, though, it's relative. 320+kbps mp3, or similarly quality AAC/AC3/DTS, etc., can be "transparent" to a great majority of listeners, even with an ABX listening test using Hi-quality sound system.
That's the WHOLE POINT of using them. Heck, you can get ~1/20th the filesize of a stereo 16bit, 48kHz LPCM WAV file original by using a mono, 8bit, 12kHz ADPCM WAV file. But people DON'T use those, because the loss of quality is obvious to everyone who has ears.

Have YOU taken an ABX test?

Now, how about getting back On-Topic?

Since flac can include in its command line switches a "verify" task to automatically compare with the source, why don't you just use that and be assured that all flac copies are truly identical? (to the source, mind you - of course the source could have been a bad mp3, but you can't get around that).

Scott

[noemi7]

Back to topic: I tested Lossless Audio Checker and seems that it only checks if there is signal cut below 20 kHz. Hence any FM recording will be flagged as lossy when among audio community it's considered to be lossless [I expect it may rise a discussion but it's worthless ].

Another checker, auCDtect is more detailed and includes spectral frequency graph, albeit not so nice like Audition has: http://y-soft.org/English/products/auCDtect-Task-Manager/ However, it's much slower than other audio checkers. Nice option is that it is possible to add manually external audio codes/decoders.

[jagabo]

You can use Spek to get a quick spectrum graph. But again, just because there's high frequency content doesn't mean the audio is lossless.

[Cornucopia]

Yes, the assumption that a filtered frequency response correlates to whether something is lossy or not is already a WRONG assumption!
There are:
1. Lossless files that have filtered bandwidth
2. Lossless files that have unfiltered bandwidth (well, at least to the limit of the nyquist frequency).
3. Lossy files that have filtered bandwidth
4. Lossy files that have similarly unfiltered bandwidth (same exception).

Given a Decoded-to-Uncompressed source file and a losslessly compressed child of that source, you can verify via audio differencing the accuracy of the lossless compression. And that is the ONLY real way to show it's losslessness (because you have to decode both first). But you can NEVER establish the provenance of the lossless file that way. That source file might be 8th generation, etc. It seems the OP is really wanting to do that - establish provenance - and there is NO good way to do that with ANY media filetype.

Scott

[Mephesto]

Hear him!
Hear him!

Toaster-******* bullshitters is right up there with 'Take a flying **** at at rolling doughnut' for over the top invective.

Now if only Yahoo would allow that or even Assgoblins I would use it over Troll. I love the smell of burning troll in the morning that that assgoblin smell it smells like ... like victory.

Copied to my Repost collection for that special moment in the future on sites that allow the artistically potty mouthed posts and flamewars. Baldrick does not like flamewars. I do but I respect his desires. Mostly.

If I were you I wouldn't be a copy-pasta lamer regurgitating epithets randomly without a clue of the intricacies surrounding their use. You'll sound like an inept tool and end up trolling yourself before the guy who hit a nerve with you.

Of course I exaggerate by boxing the audiophile crowd with the Technophiliacs of the now-defunct TOTSE.com who sooner rushed to prove with photos than deny the ravaging toaster burns on their cocks like any even subnormal freak would do. But I took it upon myself to denounce all technodork clubs toasterfuckers, and they're all crying about it because they know it's true.

while it may sound fine on it's own, the second you listen to the original (on descent speakers, no 4 ohm crap for starters) you'll hear things you didn't hear before

Seek psychological help.

Now to digress onto topical matters, despite what Cornucopia said, a good quality spectrograph can indicate evidence of DCT compression which manifests itself not only with an upper freuency shelf cut. No truly original, lossless source selectively shit-cans portions of quiet audio which you can see on the spectrograph way below the upper frequencies.

They become more visually visible upon subtracting the channels. This is really hard if not impossible to detect on very high bitrates though. 320 VBR for all intents and purposes is lossless.

If you post a sample here (preferably the whole thing, your problem if you break copyright laws) I'll take a look but if it is indeed sourced from a lossy MP3 file, I likely wouldn't be able to tell if it was a high bitrate.

[Cornucopia]

Certain kinds of audio effects ALSO do the same thing (gating), so those are ALSO not a clear indicator of lossy compression.
And again, if somewhere further back in the chain a file was lossily compressed, even if it is now uncompressed, that will give a false positive to a losslessly-compressed child of it's immediate source.

@Mephesto, maybe you weren't around back in 2006 when I created an online CD WAV vs. MP3 ABX shootout (in response to ROF's incessant BS). The first iteration revealed the filtering involved in LAME's more heavily compressed options; in a later iteration, I pre-filtered everything (including the CD WAV), making the comparison (between it & the high bitrate mp3) infinitely more difficult.

Scott

[Mephesto]

Certain kinds of audio effects ALSO do the same thing (gating), so those are ALSO not a clear indicator of lossy compression.

I doubt it. Unless those audio effects purposelly apply DCT compression to achieve that "sizzly" artifacts that makes teens today prefer MP3 over WAV.

And again, if somewhere further back in the chain a file was lossily compressed, even if it is now uncompressed, that will give a false positive to a losslessly-compressed child of it's immediate source.

If it endured lossy compression at any stage then it was... lossily compressed and will look such in the spectrograph. Your point?

@Mephesto, maybe you weren't around back in 2006 when I created an online CD WAV vs. MP3 ABX shootout (in response to ROF's incessant BS). The first iteration revealed the filtering involved in LAME's more heavily compressed options; in a later iteration, I pre-filtered everything (including the CD WAV), making the comparison (between it & the high bitrate mp3) infinitely more difficult.

Scott

I've joined in 2009 so nope. Explain what kind of "filtering" we're talking about here.

[Ethlred]

If I were you I wouldn't be a copy-pasta lamer regurgitating epithets randomly without a clue of the intricacies surrounding their use

If was me I would not do that either.

You'll sound like an inept tool and end up trolling yourself before the guy who hit a nerve with you.

You seem to have issues that are not related to anything I said. As an experienced Troll Slayer, twelve years and many arms collected, I usually know what and when to use specific tools. In this case your words are not limited to audiophiles.

Toaster-******* bullshitters

Simply has no mention of audio. It could be directed at old Amiga fans or people overly enamored with toasters of the cooking kind. The real universal gem though was this:

Assgoblins

Hear him!
Hear him!

That was freely stolen from the latest Aubry Maturin novel I read. I thought it was a better fit than the Back Bencher's Hear Hear. Much more pompous in what I used.

Do you have a problem with praise for good work? If so I do apologize.

Now back to your regularly scheduled technophilia where I usually find Cornupia to be worth reading as well.

[Mephesto]

I just encoded to 320 kb/s MP3 on all 3 modes though the highest Q setting on VBR only gave me about 260. All of them were too easy to spot on a spectrograph. Entire top shelf was cut off except spikes of drums that had reasonable amplitude.

With AAC 320, it was much harder. It took out only pockets of low-dB noise in the ceiling shelf (20-22khz) but otherwise looked identical to the WAV. I can still spot this without the original to compare it to but there's only one way to find out.

But at this point, the lossy audio is half the size of the lossless so I question if it matters anymore.
Losing everything above 16khz like MP3 does matter to people who might be remixing the song at slow speed and lose the brightness of upper bands that would be audible at half the speed.

So if looking at a high-quality spectrograph can't tell you whether the song is really lossless, it's lossless enough.

If was me I would not do that either.

Good good.

You seem to have issues that are not related to anything I said. As an experienced Troll Slayer, twelve years and many arms collected, I usually know what and when to use specific tools. In this case your words are not limited to audiophiles.

Troll slayer eh? What dunes of the intrawebz have thou fought the good fight on, comrade? It appears we have lost the war as we have been unable to curb the mass immigration of tards from that other world.:/

Simply has no mention of audio. It could be directed at old Amiga fans or people overly enamored with toasters of the cooking kind.

what u talkin about brah they all the same breed.

The real universal gem though was this:

Assgoblins

They are! Dont believe me? Look at their latest botched up listening tests or any thread for that matter.

Do you have a problem with praise for good work? If so I do apologize.

As a misdiagnosed schizophrenic I am indifferent to praise or criticism.
But for real, you said something along the lines of saving my description of most audiophiles to your collection so I simply warned against copy-paste lamer tactics if you intended to troll pedoph--I mean audiophile communities. 'Don't want anyone validating their stupid-asses.

[hogger129]

The only reason I can see to use lossy compression is if you have limited space on your portable player. Buy the CD and back it up to FLAC or ALAC. I realize most music is the victim of the loudness war but if you look at secondhand music stores, you can find older CD's that are mastered right that sound a lot better than a 320 MP3.

[Mephesto]

Lossless backups should always be kept but casual sharing of music in FLAC format is a hell of a waste of time and resources. Average internet speed is still way too slow and I really hate in 2013 still waiting half a minute to download a simple audio clip.

[Annakkin]

Interesting discussion. Think that many of participants never have proper speakers ( min 150W RMS with price tag above 1000$/ par) nor appropriate amplifier in line with real source of music, which is still Vynil LP. These days we see introduction of SACD (studio Audio CD) od DSD that can provide audio reproduction up to 44 kHz. My pick up on turntable have bandwith up to 60 kHz, and cost more than best laptop even today. Just because doctors test our hearings with simple sin wave and detect that we do not react on tone above 20 kHz (most of us) someone decide to put limit on CD format bandwidth. In line with same logic, they decide that if you use PC speakers that hardly reach 16 kHz you cant notice difference if they cut off all sound information's above 16 kHz. So how to find difference. Not simple, and there is one way only.

With my friends on university we did it long time ago, when CD is introduce as new ultimate quality format for sound.


One of my friends play violin, and we recorded on 4 channel (2 x stereo on same tape) on TASCAM studio tape recorder on max speed with 48 kHz bandwidth. On 2 tracks we recorded without any filters. On other 2 tracks we recorded from same microphones over digital filter set on 22 kHz. And than we play these recordings over appropriate speakers ( hand made with up to 60 kHz electrostatic tweeters).


And 10 of us notice in first 3 sec of replay dramatic difference between these two recordings.
It is about accurate reproduction, because if you remove some harmonic that fall somewhere above 30 kHz you change sound spectrum and it is not as original, nor it sound as original. And you can hear difference.


So if you do not have audio system that is above a 5-8K$ you cant compare sound quality. And yes on My Lap top I have difficulties to differentiate MP3 on 320Kb/s from FLAC 16bit/44kHz (stadard CD spectrum) , but I can notice even on Laptop difference between MP3 and FLAC 24 bit 96kHz. If i Play on my sound system all of you will notice dramatic difference between all of them.
And yes , for headphones, try to check differences but use some Sennheiser (HD700 or IE800) or AKG for test (what you think, why all new phones can play FLAC and WAV beside MP3).


After you make similar tests, as I did long time ago, you will not complain that average album in FLAC 24 bit 96 kHz is 1-1.5Gb with 10 min song about 200 Mbit and 9Mb/s rate on your NAS drive, and you will forget for MP3.