How to know if a FLAC is a real FLAC?

Discussion in 'Audio Hardware' started by mickael57280, Aug 7, 2017.

  1. mickael57280

    mickael57280 Member Thread Starter

    Location:
    France
    Hi everybody, I have lots of FLAC files and to be sure it's not an upscaled mp3 I open the files into a spectral visualiser.

    I use a program called Spectro, useful because it's an option that show you the frequency cut off, and that's the confusion for me.

    I have some files who the cut off is at around 22khz (+ or - 0,5) but I have some files at around 20 khz but whit frequent short up to 22khz, and I have a track from a live album that say the frequency cut off is at 11,5 khz but with massive up to 22khz.

    For test I have download a 128 mp3 and the frequency cut of at 16khz but nothing never go up to that.

    So how to be sure my files are real flac?

    Thank you very much
     
    SOONERFAN likes this.
  2. Rolltide

    Rolltide Forum Resident

    Location:
    Vallejo, CA
    Wouldn't the file size be the giveaway?
     
  3. MrRom92

    MrRom92 Forum Supermodel

    Location:
    Long Island, NY
    If you trust the source it came from then there's probably not much more to it than that


    For downloads, you're always operating on trust


    If there's the possibility of getting a CD rip then at least you have some method of verification available to you
     
  4. Sid Hartha

    Sid Hartha Well-Known Member

    Location:
    The Midwest
    If it sounds lossless, it is a real flac.

    If you can't tell the difference, you're wasting money and drive space.
     
  5. c-eling

    c-eling Forum Resident

    I have a few old compact disc's where there is around a 20 cut-off, some old Depeche Mode for example
    Traders Little Helper is a good little tool that was suggested to me by @Ham Sandwich
    Confirmed a Tidal download of what was supposed to be a lossless FLAC was just upsampled, apart from using Spek
     
  6. davmar77

    davmar77 I'd rather be drummin'...

    Location:
    clifton park,ny
    best way to know for sure is to check the frequency with cool edit/adobe audition, audacity or a similar program. I've done it many times. you really can't go by your ears or the file size in a lot of cases.
     
    russk likes this.
  7. skimminstones

    skimminstones Forum Resident

    Location:
    Bexley, UK
    I use a free program called audiochecker. Its simple enough, you add the files to it and it scans them. Tells you if it was originally a FLAC or not. You cant rely just on file size as you can re-encode an MP3 to FLAC and the file size will change drastically but at its base it is still an MP3.
     
    Grant, SOONERFAN, andrewskyDE and 5 others like this.
  8. tuttle

    tuttle Forum Resident

    Location:
    Toronto, Canada
    You cannot tell from size, because someone may have upsampled a lossy file, "converting" it to lossless. I've found this after buying music from some download sites.

    Audiochecker is worth using. I use it as the "first cut" test. If it says the files are CD quality (lossless), then they are. But, it sometimes gets fooled and gives false negatives, falsely identifying lossless files as lossy. It tends to get fooled by old, noisy recordings or by musical distortion in certain frequencies, or recordings (often 1980s) where musical data basically cuts off in treble frequencies..

    So, if it says a file is good, it is. If it says it fails, then you need to check spectrals to know for sure.
     
    Last edited: Aug 7, 2017
  9. Rolltide

    Rolltide Forum Resident

    Location:
    Vallejo, CA
    I'm curious to learn more about the use cases at work here. How do people find themselves with the problem of needing to make sure a FLAC wasn't sourced from an MP3?
     
    eric777 and Kyhl like this.
  10. c-eling

    c-eling Forum Resident

    Last time I had itunes installed many moons ago I grabbed the Men Without Hats EP Folk of the 80's. I found out last month tidal offered a losseless version along with the mp3, it was quite pricey. So I paid a premium price for an up-sample is what my case was about. I always check DL's. Of course Tidal had no clue what I was talking about when I requested a refund a few times. Never got one, and Tidal will never get my business again.
     
  11. mickael57280

    mickael57280 Member Thread Starter

    Location:
    France
    Thank you everybody, using Audio Checker now and it's save my life, just one strange thing with a track, said it's CDDA but just with 43%.
     
  12. RiCat

    RiCat Forum Resident

    Location:
    CT, USA
    Don't know what you use for AV but ours would not permit Audiochecker as it reported it attempted system security violations.
     
  13. mickael57280

    mickael57280 Member Thread Starter

    Location:
    France
    For for me, but I think it's not reliable
     
  14. mickael57280

    mickael57280 Member Thread Starter

    Location:
    France
    So for make it simple, is there any reliable solution?

    I have an album of Thelonious Monk and John Coltrane Live At Carnegie Hall, Audio Checker said only one file is false, when I open it with spectro, the frequency cut off is at 11,7 khz, is it possible?
     
  15. c-eling

    c-eling Forum Resident

    Could be a very low bit rate mp3. My cassettes usually go to about 16
     
  16. goodiesguy

    goodiesguy Real Wild Child

    Location:
    New Zealand
    Lossy sourced FLAC's seem to go up to about 15khz, then you see only spatterings of frequency information above that, instead of a nice full frequency or a mild taper off.
     
  17. RiCat

    RiCat Forum Resident

    Location:
    CT, USA
    I use the spectrograph in Audacity. The only time I use it is when purchasing downloads. I have had a few instances where the flac. provided was clearly an up-sampled original file. I have sent screen shots and always gotten refunds. What I have learned is depending on the selling service the files are not owned by the seller. There are services that are aggregators providing a sort of front end to purchase from. They buy from providers and then resell to you and I. Sites like HD Tracks have never let me down. Usually it is those that offer a selection at what seems like a discount best buy that are not always what was advertised. Re-sellers do not check each offering in every format for correctness. I have tried a few supposed software that say they properly check. Before you take the results as fact I suggest you rip a few of your cd's to flac and check the flac files. My experience is that false positives and false negatives are so common as to make me question the value of the checker.
     
  18. Keep in mind that these tools have false positives and false negatives.
    That means that some lossless files will get identified as being from a lossy source. And some files from a lossy source will get identified as being lossless quality. Use the results from these tools as a guide, not as the final word.

    I've had classical CD recordings get falsely identified as being a lossy source. Especially older classical recordings. One reason is that classical recordings don't have the high frequency content of typical rock recordings (lots of cymbals and snare drum). Not all classical recordings will put a mic up close to the violins to capture as much high frequency sounds from the violins. Taken together, some classical recordings don't have the obvious high frequency content and can get identified as lossy because of that.

    Trader's Little Helper comes with a tool called auCDtect that's a bout 9 years old from: True Audio Checker - Project Description - Tau Projects
    Lossless Audio Checker is a newer tool from Lossless Audio Checker
    I haven't actually tried Lossless Audio Checker yet. I've just been using the version of auCDtect that comes with Trader's Little Helper.
     
    Grant and shaboo like this.
  19. brimuchmuze

    brimuchmuze Forum Resident

    Yes. I have this album on CD, and the segments where Monk is playing solo have little content above that frequency. This is explained first by the frequency range of the piano, and also by the nature of the recording .

    So I think you can stop worrying and just enjoy the music. We are lucky to have it :)
     
    Ham Sandwich likes this.
  20. I've found a couple downloads from Bandcamp and CD Baby that have been lossy source even though the downloads were FLAC. In those cases it has been because the artist or their label gave Bandcamp or CD Baby a lossy source. The CD version of those albums have been from a lossless source and sounded better than the lossy source download. It happens. I haven't found any commercial CDs that have been from a lossy source. But I haven't really been checking for that. I'll check a CD if I hear something that makes me suspect it might be a lossy source. But that's very few of my CDs that I've suspected and tested.

    There was a problem with audience recordings and soundboard recordings for the Grateful Dead and similar bands would end up with a lossy source accidentally getting used in a lossless archive. Those lossy sources have largely all been identified and weeded out by now. That's the reason that Trader's Little Helper added in the detection for lossy source.
     
  21. Cockroach

    Cockroach Member

    Location:
    Visalia, CA
    I think the only real for-sure way to tell is to use your ears. If it sounds crisp and clean, it's probably lossless. If you can hear any kind of artifacts, then the source was probably lossy. With some older recordings, it may be too hard to tell either with ears or measuring equipment/software. You just kind of have to go on blind faith with those. And bear in mind that many newer albums were mastered with so much dynamic range compression (to make them "louder") that it'll be impossible to tell either way.
     
  22. harby

    harby Forum Resident

    Location:
    Portland, OR, USA
    One piece of evidence you can investigate on a brickwalled recording is to see if it is digitally clipped. Mastering brickwalls by using digital limiters to get levels right up to 0dB, but if it stays in the original digital form, there won't be clipping (or there will be a very rare one or two samples per file that are at max value.)

    When the audio is losslessly encoded and decoded, it is put through several banks of frequency filters; any alteration to the original waveform may remove audio bands, and the resulting filtered waveform can exceed 0dB. If there are runs of several samples being clipped throughout, it is likely a CD that's been put through a lossy encoder with bad decoding.

    Lossy files can have audio indistinguishable to the ear, and also can be encoded with high frequency spectrum info (although this is usually wasteful as you only can see it, not hear it). Besides MP3 where the lowpass settings have been overridden, MP4/AAC uses spectral band replication to simulate very high frequencies, and this does not need to be bandwidth limited.

    Here's a CD rip that I cut a segment out of, AAC encoded and decoded, and then FLAC encoded. I'll bet most any auto-detector gives the same answer for before and for the after:

    rushin.wav (5.1MB)
    rushout.flac (3.3MB)

    Spectral views of both are barely different, but you could only tell if switching between them with identical views.
     
    Last edited: Aug 8, 2017
  23. gregorya

    gregorya Forum Resident

    Oh great, so now I need to find a tool to make sure that my FLAC-checker tools are working properly.

    A "FLAC checker checker"... :):)

    I'm sure I will learn that the FLAC checker checker gives false positives and negatives as well, so of course, the answer is a FLAC-checker checker checker.

    Geez Louise, music is complicated... :)
     
    Ham Sandwich likes this.
  24. head_unit

    head_unit Forum Resident

    Location:
    Los Angeles CA USA
    FLAC is just a lossless encoding format. If it shows as a FLAC file and plays, it's a FLAC file. You could encode something great perfectly, and you could also encode vinyl rips from a child's portable turntable.

    I think you mean to ask "how do I know if it is a WELL MADE file?" the answer is: you can't know for sure. If it sounds good, enjoy! If it doesn't sound good, maybe someone made a bad copy of something and then encoded it to FLAC out of ignorance or stupidity. And, there are a bunch of FLACs of stuff like 1970s bootleg live recordings which are perfect lossless copies of sometimes horrible sounding originals.

    I always both get a bit annoyed and laugh when I see someone posting "DO NOT MAKE AN MP3 FROM MY FLAC" of, like, the MC5 with a terrible PA in an acoustically horrible venue, totally overloading the input microphone of the cassette recorder. Yeah, when I turn that into a 256k MP3 I'm gonna kill the sound quality :rolleyes:
     
  25. Tim Müller

    Tim Müller Forum Resident

    Location:
    Germany
    Hello,

    one indicator could be the appearance of the spectrogamm (like, in Audacity or the like).
    Lossless music has a continuous spectrogramm.
    MP3 encoded music has a somewhat grainy appearance (you must zoom in to see it). I guess, that's because of the filtering through filter banks when encoding MP3.
    Furthermore, some lossy formats tend to brickwall the spectrum at around 15kHz or so.

    Those visual clues can be indicators of lossy sources.

    Best regards
     

Share This Page