February 19, 2018
 by Paul McGowan

Here’s something to contemplate. If ripping a CD results in identical bit-perfect data to the original and, if you store that ripped data on a hards drive, how could one ripping process sound different than another?

The answer is simple. It cannot.

Yet, few among us would suggest one ripping process sounds the same as another. Thus, if all the evidence is true, that means the data has to be different. Error correction was used. Something other than jitter or timing changed. (We know this because hard drives do not store clock data).

I wonder if there have been any studies or examples of ripped vs. original using different programs. I am often lazy when ripping CVDs and use iTunes at its lowest copy rate. My friend Jason Serinus, who is also a reviewer at Stereophile, scolds me for this practice, suggesting the results are less than optimal. Bad, in fact.

Love to know what’s true and not.

Subscribe to Paul's Posts

55 comments on “Ripped”

  1. I ripped my entire CD collection about 6 years ago. It was a bit perfect rip using software licensed to several companies including Naim. The ripper server could play the CD, rip it to WAV, convert to FLAC and also did an MP3 version at the same time. The WAV was checked to the AMG Pro database. I can’t say I could ever hear any difference between CD, WAV or FLAC and I was assured there should be none. The WAV was a direct copy and FLAC was lossless compressed. I have used FLAC for CD rips ever since.

    Logic tells me that computer companies used to supply software alternatively on CD and download, and they are identical. I therefore assume that to suggest otherwise for audio data is being paranoid. Don’t tell me audio data is any different from software data or any other digital data.

    Moreover, when computer chips were expensive we used to thin down Fortran code to take out redundancy. It did the same job, worked the same. Lives would have been lost otherwise, given the nature of the programs. I therefore have no problem with the idea of compressing redundancy in WAV files to save disk space or streaming bandwidth.

    It is curious watching the Qobuz player, which streams 16/44 as FLAC, the data normally ranges from around 350kbs to 1000kbs, it never gets very close to 1441kbs. High res streams are always fixed bit rates, 24/192 is about 5600kbs, which I always think is a vast amount of data for negligible or inaudible difference in sound quality.

    1. Although both audio CD and CD Rom use Reed-Solomon error correction, the sectors on an audio CD are pure data, whereas on a CD Rom there is less data and some error detection and correction bits instead. This makes the CD Rom even more accurate.

      1. Chris, you make a good point about error correction and when it is present. I assumed all CD rips have error correction. On the occasion I’ve used dbpoweramp it had error correction checks.
        The other one I was reading about is ethernet signal packeting that has error correction so it can never be wrong, hence any ethernet cable will do. The same cannot be said of usb.
        I had digital noise form my system recently. Pops and clicks like vinyl, every 10 to 20 seconds so. It was apparently due to the sending steamer’s reclocking having some incompatibility with the receiving device’s digital card, that also does reclocking. I switched to a non-reclocking streamer and the problem disappeared. It struck me as the first time I remember actually hearing error coming through in the digital signal and wondered how often, if at all, anyone actually hear data errors.

        I was surfing in the bath, googled “CD v WAV”, realised this thread is about whether two 16/44 formats sound different. Most of the audio community are still stuck as to whether double-blind you can reliably tell the difference between 16 bit and 24 bit. More worryingly, those that can tend to prefer 16 bit because it sounds less clinical. Those that prefer high definition all seem to call it more “musical”, which I take to mean subjectively they want it to sound better but can’t explain what difference they think they can hear.

        1. Steven, I do not think you are quite right about Ethernet. The frames contain a 32-bit CRC which will pretty reliably indicate corruption, but there is no error correction. The protocol using Ethernet is expected to detect corruption and request a resend. Incidentally USB packets contain a 16-bit CRC, but I am not really familiar with the complex family of USB protocols.

          I think that I have written here before that when, some years ago, I did AB testing of 16/44.1 and 24/96 or 24/192 versions of the same music I was unable to tell the difference but that, despite this, I think that Hi-Res now sounds good. Since the time of the original test I have had modern crossovers installed in the speakers and use a better (more expensive, at any rate) all-in-one. I should really rerun the test, but AB testing is *very* tedious.

  2. I used dBPoweramp or rarely Exactaudiocopy as software, but never went into more specific ripping hardware than a normal CD/DVD drive. That’s something I chose not do dive in deeper.

    We also have no chance to care for such stuff on manufacturers side, how good or bad the one or other CD production process is, so no need for me to do it better at home 😉

  3. The ripping software is a later stage in the extraction process. Different pressings of the same CD can have subtly different offsets (not important for audio listening, but significant for data verification) and CD drives have different capabilities for reliable extraction (from memory the Plextor 230A is the best). When first I read your post, Paul, I thought immediately about the file hashes which are used to verify the integrity of file downloads. There are many programs which can generate them. Unfortunately for music rips there tends to be extraneous data like album art and metadata and, in the case of FLAC files, the bit pattern of the audio data will vary according to the compression level used, even though in each case the data expanded from it will be identical. File hashing is not reliable as a way of establishing the integrity of different rips,

    In the Windows world the most respected ripper is EAC (Exact Audio Copy). In secure mode this is completely anal about doing multiple reads until it finds consistency. The price you pay for this is time; a rip can take 15 minutes. I started off by using this when I first ‘discovered’ FLAC files but unfortunately on the PC I normally use it failed to release the CD drive, so that a re-boot was necessary if you wished to rip a second CD. This was a show stopper, and I switched to dBpoweramp which is also well respected and only takes a little over 2 minutes for a rip. It verifies accuracy by generating a CRC (a sort of hash) for each track and then comparing this with the AccurateRip database to ensure that your result is the same as most other people got. EAC can also use AccurateRip for verification.

    From hundreds of CD’s I think I have only had three which failed to rip correctly according to the database. I have listened very carefully and failed to detect any audible problems with them. This is probably not surprising since it only takes a 1-bit error to change a hash radically, and you have gigabits on a CD.

  4. I also use dBpoweramp as my tool of choice for ripping CDs (occaisionally Exact Audio Copy – for any really difficult to read discs). dBPoweramp does many things when ripping the disc in terms of checksums etc etc (which I do not pretend to fully understand – apparently different drives have different capabilities in terms of error checking and correction) but once the track is ripped it uses the accuraterip database to verify that the rip you have made is the same as other people who have ripped the same CD (with the same CDID number) – this gives some level of confidence that your rip is accurate (it is unlikely that someone will rip the same CD with the same errors). Itunes works a bit differently and just reads the files in a single pass and does not perform much in the way of error correction, I am sure that 9 times out of 10 the rip is accurate even without the checks. It is reassuring to know that the ripped version of the CD is the same as several other people have ripped – this means it should be accurate….
    …..I have not seen any physical evidence that supports that different rips of the same CD – verified against the accuraterip database sound different. I have heard people claim that these rips sound different but cannot understand how… given that files once ripped get moved from pc to pc or server to server I can not see how they can sound any different (assuming everything else stays the same)…. if this was the case I have no idea how you can guarantee the quality of any digital music as even before the CD masters are produced the data goes through allsorts of manipulation and gets copied multiple times….

  5. Don’t understand the post. If you make 2 rips of a CD with 2 different softwares and both match as identical/error free, then they do sound the same played back in an identical manner.

  6. For vinyl playback it’s all about wow&flutter. Even if the wow&flutter values are smaller than those of the cutting machine the values will never have an identical temporal distribution! Thus a correct reproduction is never possible. Same argument is true concerning the accuracy and temporal deviations of the master clock for the digital recording and the CD player or DAC! From a standpoint of reproduction techniques the latter clocks should show a better accuracy than the clock used for recording! Thus I always wonder which patterns are really optimised for getting the best sound in the stereo system!

  7. JRiver is my ripper of choice and I use it also for tag management, but then generally will use Roon for playback. Error correction can be a pain at times depending on the condition of the CD (I buy some used CDs) but I have a pro-style polisher which helps most of the time. Interestingly enough, even when there is a ripping error, I struggle to have ever claimed I could hear what it didn’t like. I am told that errors can result in small dropouts or noises in the playback stream but I’ve never heard it.

    I’m in the camp of those that feel there is no audible difference in the playback of various file types, in fact I fail to hear any differences in playback between the CD itself and a straight 320kb MP3 file. I’ve tried so many times, AB testing, listening over and over to music I know well, just am not hearing it. This is just me, of course, I know many will say their experiences are different.

    I just received Mark Waldrep’s book and in it is a Blu-Ray that is supposed to have tracks on it allowing a quick file change and format change component for ABCing tracks for comparison, I will be anxious to try that. More on that at another time.

    1. Using JRiver with import options set to analyze audio some older or used discs that are scratched occasionally need multiple reads to get good data. And sometimes there is never good data on a couple of tracks. So I simply make a copy of the disc and the duplicate will read fine. Strange.

      1. When I first got JRiver and did imports I recall that the settings were to analyze the audio and “normalize” it. I then would see in the tag settings how tracks were adjusted to make them all the same volume on playback. I was not sure that I liked that aspect of it and went and turned it off so that the files just get ripped with no further analysis, other than the secure rip which makes sure that the audio that got ripped matches what is on the disk. That is the part I have troubles with at times, in that I get errors saying a certain number of reads was required to get good data, or there was unreliable data at xxx point in the rip, but I never seem to hear anything in playback that corresponds with the “error”.

        I, too, have made copies and then gotten a good read off the copied disk! Strange indeed.

        1. I never use “normalize” either but I went to “Expert Options” under CD Ripping and set it to “analyze audio during ripping” I also have mine set to secure rip. Maybe that’s what causes problems?

          1. That’s the part that I turned off. With that turned on, from what I gathered in reading the help wiki, the program would analyze the file during ripping and then put some kind of a marker into the file so that it would play the songs from that album back at a level volume so you didn’t get all the ups and downs from the quiet and louder parts.

            If you open the tag manager and go to the “Volume Level (replay gain)” section (under playback if you are using the most recent JRiver with the new tag manager interface) it will probably show how many db up or down it will play the file upon playback.

            Having said that, I don’t know if it codes it for anything you play it through, or just JRiver. Not sure at all if that marker (or coding, not sure what to call it) would come through and control other playback software, if you played through Roon or other software.

            Anyway, it bugged me, I wanted it to play back the way it had been given to me from the vinyl, for instance (take the louder parts with the quieter) so I turned it off. I have done absolutely ZERO testing of whether I could hear a difference between one analyzed with the replay gain and a file ripped without that setting enabled, this was just a personal preference thing for me.

              1. Yes I would agree and mistyped what I was trying to say!!! I had to go and re-read the wiki and certainly do not want the program normalizing things for me LOL! I can adjust the volume control.

  8. I have been getting into interesting dialogues (arguments) on Facebook regarding the sound quality differences between 320kb MP3 files and FLAC files. My position is that a lossy format is, by definition, missing data and shouldn’t sound as good as a lossless format. The counter argument is that what is missing is inaudible and won’t affect the sound. Is that position accurate and ripping to a high res MP3 is as good as FLAC? Obviously, my understanding of this subject is rudimentary at best.

    1. My experience, during those AB evaluation tests I did, was that I could distinguish between 320kbps MP3 and 16/44.1 FLAC, although in some cases it was very difficult and I spent hours listening to some tracks to identify critical sections where MP3 encoding has known weaknesses. My general conclusion was that if I could only download an album as a 320 MP3 then I should not worry about it, because I would still be able to enjoy the music without its imperfections being distracting.

      1. I would pretty much agree with your assessment chris! I have adopted the “it’s not the best, but good enough” philosophy and have to tell you it has done so much better for my mental attitude ha ha much less fretting about things.

        I also have come to believe, subject to considerable discussion as was pointed out by lavine43 above, that a full 320kb MP3 file already exceeds the dynamic range of the original source material, primarily tape in most cases. So, if the format already holds all the information that it was possible to capture then what is exactly been lost?

        Here is where the measurement folks weigh in and talk about what cannot be measured that may be present and may be lost with the MP3 or any lossy format. I will not argue that point, as it is each to their own, but I certainly do not hear the difference, all the subtle things that are discussed about a music file I can clearly hear with my kit, so, again, it’s “good enough”!!

        I would absolutely agree with most of the posts that talk about modern music (rock, pop, country, or, as soundmind would say, “manufactured music”) already being so compressed dynamically that it already holds less information than any lossy file specifications no matter what the lossy file type. acuvox will point out that lossy gives him a headache and I may not argue that at times either, depending on what music I am listening to (but I am not sure it’s from the actual file type, ha ha, but from the crappily mastered and engineered music in general) The mastering engineers of course tell me that they have to master the music to be heard over the din of a department store music system, or the crowds in a public place or bar and so on. As one guy told me, “we don’t make any money unless we get heard and to get heard we have to push the music through the background”.

      2. I read this article this morning, from one of national newspapers, the music editor and “head of audio” (very surprised at that, they don’t write about audio), so basically two musical non-audiophiles spoke to Gilad Tiefenbrun and went to a Linn main dealer for a listening session.

        mp3 v, 16/44 generally noticeable, marginal differences 16/44 to high res.

        1. I had to smile a bit about this article, knew it was dated (2014) when they kept mentioning Neil You and Pono, which is dead in the water now and in the rear-view mirror of technology. Not really telling me anything new to my ears, but, an interesting read. Thanks for sharing.

          1. Linn was a market leader in digital and streaming (about 10 years ago they stopped making CD players and it saved the company) and also lead in some of the best digital releases through their own recording company. They also used to sell other labels’ digital downloads as there were so few outlets. So they were the right people to speak to.

            This article struck me with the thought that even in 2014 it was pretty clear that high res music was never going to be a mass market product because not enough people had the equipment, cared or could hear the difference. I remain of that view and hence that Tidal is sunk (just a matter of time) and Qobuz has a chance as it is going for low volume high value customers.

    2. The answer is in the ear of the creator and the beholder. Commercial music is so processed there is already massive reduction in coherent information and no reference for accuracy. Every knob in a studio generates distortion and there are typically hundreds of knobs in every audio production chain. Radio processing, cheap DACs, amps, speakers and bad rooms mangle the same information that is thrown away by lossy compression so the results are garbage either way. Listening to bad recordings (99.9% of releases) through bad audio one never develops the aural discrimination to hear the difference.

      I can’t identify the difference between CD and 320K files on current pop music, but streaming and MP3 always give me a headache because distortion of distortion sounds worse than distortion of signal. OTOH, with a live mic pair zero processing recording the difference is obvious – and the same is true with the upgrade from Redbook to HD and DSD.

      1. Thanks for your insights. I try to avoid modern music (aka noise with a beat) as much as possible. I will continue to rip CDs to FLAC or download FLAC vs. MP3 wherever I can. I use EAC to rip. Maybe these old ears can’t tell the difference between formats, and maybe I’m just stubborn. Storage is cheap. I’ll go with FLAC.

    3. I’ve got lots of mp3 and flac music files on disk. Playing them (USB) through an Oppo 203 and DS dac, I find that most of the time the flac files sound better, often considerably better, but that there are some really good mp3’s and some poorly done flac’s. Off facebook, I’ve done some MP4 to flac conversions (AVC) and while they are acceptable, there is a easily noticeable loss.

  9. When I first started to build a server, I extensively ripped a couple CDs with several different software, into several different formats, played back through several different players. I wanted to make sure I didn’t rip my collection, only to find out I needed to rip it again for better quality. I could not tell the difference between rips done with any of the software into any of the lossless uncompressed formats. I simply use iTunes to rip to AIFF on my Apple MacBook Pro.

    I compared PC ripped versions of the album as well. They were ripped by an audiophile friend and copied to a thumb drive. He uses WAV for his PC server setup. No difference. I had a different guy insist that the PC version was not better because a copy of it was made, and and that it lost something in the transfer from/to the thumb drive. Does it ever end?

  10. Paul, using iTunes to rip CDs is fine as long as the preference to “Use error correction when reading Audio CDs” is checked. If there are no glitches or dropouts, the rip is almost certainly as bit perfect as it can be. If your CD is damaged or defective, you might get different results using different ripping software or hardware.

    At an Audiophile Society dinner in New York some years ago, Steve Guttenberg and I had the chance to talk to Alan Silverman about this. Alan had mastered many of Chesky’s CDs. (David himself was there and did not contradict Alan.) He said he used iTunes rips as a quality-control check on the albums he mastered, and he told Steve it was wrong to think that other ripping software would sound different.

    Alan’s explanation went like this, if memory serves. When he got a CD back from the production plant for a QC, he ripped it in iTunes and then brought the files into his audio work station. He lined up his original source files (from which the CD was made) against the iTunes rips from the CD, then inverted the phase on the rips. In virtually all cases, he got a perfect null, which meant that the rips were a clone of the original files that he had sent off for CD production. Any time he didn’t get a perfect null, he told the CD duplicator to start over and make a new CD for his approval.

    1. That’s fascinating. I would be interested to know if the subsequent CDs passed muster, or if he was just a nightmare client (from the duplicators POV). If there was an issue, it could just as easily have been with the rip as the CD.

      1. I assume the subsequent CDs would have passed muster. Unfortunately I didn’t get into the details of the production process with Alan; that would have been interesting. Alan was trying to convince Steve that iTunes rips really are bit-perfect clones of the original files he mastered, that nothing is lost or changed in the manufacturing process or in the ripping process. Audiophiles who listen for subtle differences between ripping applications are wasting their time and fooling themselves.

        However, I have certainly seen CDs go bad or get damaged. Sometimes even brand new CDs are hard to rip because they contained so many errors and require slow reading and re-reading. Apple’s iTunes (like other good ripping applications) will slow down to a crawl on a bad CD. Because I’m a video editor, I have a small collection of optical drives and several different applications I can use. On those one-in-a-thousand CDs that iTunes has trouble with, I’ll try different software and hardware. Sometimes, though not too often, I can get a bad CD ripped acceptably when iTunes gives up on it. But that’s a separate issue from whether an iTunes rip sounds less good than any other rip. It doesn’t.

        1. I’m an editor as well, and started getting less errors on CD rips when I switched from the internal drive on my Mac to external BD burner/readers. I have one Mini whose drive I had to quit using, as it cannot spin a disc without excessive vibration. Laptop drives can be fairly lame as well. When you look at the mechanism (I’ve replaced them with HDs in laptops) it’s no surprise – they’re built to be as light and thin as possible.

  11. Remember that when comparing rips there is a lot of System between the files and one’s ears. It may be that limitations in one’s equipment make it impossible to hear a difference above a certain file size or type.
    Using file sizes or types above that resolution level then becomes an academic exercise, until one’s ears or one’s system improves in resolution! Of course, if time and/or expense are not obstacles, there’s no reason not to use the highest file size/type available as both a way to future-proof the music, and to insure that if a golden-eared person stops by you won’t be embarrassed to demonstrate your system to them 😉

  12. Chris J is on the right “track”. Absolute digital accuracy is a myth. We live in an analog world. Digital circuits are clocked, highly distorted analog circuits. The lower the energy per bit, the more likely that the noise of the Universe will falsify it – and we are at nanoWatt levels.

    The density of storage media is always pushed until they make frequent errors reading, but with redundancy mechanisms to re-construct accurate data. These fall into two classes: enough error correction bits to re-create a perfect copy from the bits read, or just enough bits to detect errors and trigger re-reading the data.

    CDs are played back at the speed of the motor rotation, there is never time for a re-send, and by the Redbook standard redundancy for error correction is not sufficient to fix the bits every time. When you burn an audio CD, this standard produces playback errors unless everything about both record and playback mechanisms is perfect to millionths of an inch and there is no shock or vibration, EMI/RFI, radiation like Radon seepage or cosmic rays. Computer drives are less than perfect mechanisms because re-reading computer data is assumed, and it would be a waste of money to make them better mechanically, electrically and optically.

    The ethernet and internet also assume re-sending packets to maintain data integrity so they are not reliable for 100% accurate streaming. Note that disc reads and wire connections are subject to BURST ERRORS where a section of data is wiped out, and error correction codes like CD data will not handle bursts.

    This is why you need specialized software that re-sends flawed packets, which means CD playback is not real time.

  13. I work professionally as a Linux DevOps Systems Administrator. When computing devices copy data from one to the other, they typically use an error checking protocol. That’s what the “TCP” does in “TCP/IP”. Every hour exabyte levels of data get transferred between computing systems scattered around the planet without error. And y’all are worried about whether or not you can accurately copy data off of a music CD?! Really?

    I also use dbPowerAmp (running under WINE on my Linux media player) and expect that it would use error correction. I do find CDs that don’t copy off and chalk it up to it being a bad CD. That does happen. But for the most part, bits is bits. And if you don’t believe that, remember that your whole personal financial estate is defined by bits on disk drives scattered around the planet. But I’m sure you don’t worry about that…

    1. Well, this is not entirely true for Reed Book CD and that’s part of the problem. As my friend Max writes me:

      “audio CDs (“Red Book”) used to run at 1x speed, and there is no error
      correction capability in Red Book, only a very simple form of error
      detection. What do you do when your drive spins at 1x speed, and you
      know the sector the drive just read has an error? There is no time to
      attempt to read the faulty sector again at 1x speed. The firmware
      then usually “guesses” what could have been there.

      So if there is a read error on a Red Book, the drive will never report
      the error to your computer; you will get bogus data.

      That’s a major difference to Yellow Book (data CD): there are more
      error detection/correction bits (thus a data CD has less storage space
      than an audio CD), and read errors are reported to the operating
      system. Bogus data (even a single bit flip) would be fatal for most
      data files, so the approach here must be different.

      With data CDs, you either get bit-perfect data or an error; but with
      audio CDs, you don’t know what you’re getting.

      This is why some rip programs such as cdparanoia do many overlapping
      reads, and compare the sectors which were read multiple times. If
      there are difference, cdparanoia will read them again and again, until
      there is a consistent picture (or until it decides to give up).”

      1. It’s a pretty crappy coder who expects everything to always work. That’s why there’s error correction. Every time I hear stories about tech like this, I just shake my head.

        I’d expect a ripping tool to do multiple reads and that there be a handshake between the sending and receiving devices.

        Why is it that digital audio designers aren’t willing to be as serious about data accuracy as everyone else? We depend on absolute 100% accuracy in the handling of financial data in banking, as well as many other kinds of data. You’d think it wouldn’t be that hard to do in audio.

        1. Replying to: “Why is it that digital audio designers aren’t willing to be as serious about data accuracy as everyone else?”
          They sure are willing, but you can’t have everything, and the requirements for real-time audio playback are completely different than for financial transactions. You have to choose the right balance and tradeoffs in your technical design decisions; financial transactions must be correct, and audio playback must be real-time.
          Today, nobody needs to trade correctness for real-time, because contemporary storage devices are 1000 times faster than necessary for raw audio playback, and you can afford a large buffer which allows the storage device to retry over and over and over in the presence of a read error.
          When the Red Book standard was published in 1980, CD drives had 1x speed – i.e. the drive would read at exactly the rate needed for playback. If you encounter a read error (which happens very often because the CD surface, unlike your hard disk, is exposed to a hostile environment) you just can’t afford to read it again. You just have to play “something” instead of the real data you don’t have. If you try again, you fall behind and have no chance to ever catch up. Impossible at 1x speed.
          In the 90ies, when CD drives in computers became mainstream, “double speed” drives emerged, which allowed you to retry, though most CD drives were pretty bad at ripping Red Book. I remember buying only Plextor SCSI CD drives in the late 90ies, because only Plextor was reliable enough for ripping Red Book at an acceptable quality.

      2. Paul, my understanding of the C1 and C2 codes on audio CD’s was that they did provide error correction, and that the data was interleaved so that the correction would work with a drop out of over 2mm along a track. This is why CD’s are more sensitive to scratches/marks running circumferentially (is there such a word?) along a data track than a radial scratch which briefly damages multiple tracks.

      3. Hi Paul,

        What your friend Max says sounds right to me. So, I would expect that if you rip 100 CDs with iTunes, on average, you’d get the same results as you would with dBpoweramp (with secure settings enabled) and EAC about 98% of the time. The difference is that the roughly 2 CDs out of 100 that have errors would likely go undetected with iTunes, possibly resulting clicks or other audible artifacts in some of the tracks (but not all errors are clearly audible). In contrast, EAC and dBpoweramp would have tried very hard to recover the bad frames by re-reading the tracks, and if that failed, they would have reported the error or at least let you know that the tracks were ripped as insecure because they did not match the checksum in the AcourateRip database for that pressing. At that point, you could try cleaning the discs, ripping with a different optical drive or purchasing another copy of the disc from Amazon or discogs to remedy the issue.

        In the case of iTunes, you’re left to discover any imperfect tracks in your library by listening…which would absolutely drive me crazy! I’d prefer to know upfront about the errors so that I could address them during the ripping process.

        For the other ~98% of the discs that ripped perfectly, they are verifiably identical to discs ripped using more sophisticated methods. I say verifiably because you can actually purchase a license for PerfectTUNES from illustrate (same folks who make dBpoweramp) and have it scan your entire library. It will let you know which tracks were ripped with errors so that you can go back and fix them if you want. Pretty cool for folks (like you) who may have bad rips lurking in their libraries.

        Back on Crazy Street, I do know some people who claim that they can reliably hear differences among tracks that were ripped to a spinning hard disk drive vs. an SSD vs. a thumb drive….even though the tracks were verified using EAC or dBpoweramp. I don’t question that they hear what they hear, but the brain is a complicated thing. Not all differences that our brains detect are due to differences in the actual sound heard. The power of suggestion is difficult to overcome.

      4. The ripping machine I use does multiple scans to avoid errors and online data checks. It takes about 20 minutes per disc. As a result, it took 800 discs before I got an error, which was because the disc was too thick! Harmonia Mundi decided to make it CD on one side, DVD on the other. The first system I had, a Windows Media thing, had a 1/3 failure rate and dbpoweramp had quite a high failure rate.
        The difference was that my ripper has bespoke software written primarily for Naim to get error-free bit perfect rips. It is not needed for playing in my system anymore, but I still use it for ripping and drag the data files over to my QNAP server.

  14. It is a very simple program to compare two files. Just find a sync point and let it run. Anytime the data doesn’t match, report it.
    There probably already is at least one freeware CD/RIP file compare program on line someplace.

    Now want to talk about error correction and concealment – digital video tape. The error correction systems in broadcast machines is extensive because tape has a lot of errors. Computer backup tape works differently. It’s not realtime so bad sectors are retried then abandoned if bad and the address mapped. Even at that computer tape is not 100% reliable but tape still continues to have the highest density to date. This is why today while almost feature film is shot digitally and archived on LTO tape, a YCM seperation negative is still made and stored in the salt mine. You just can’t trust that tape to last 100 years like the film will.

  15. I agree with chrisj1948 and paul_riordan.

    I also use DB Poweramp because if it’s ability to rip from my drive in “secure” mode, re-ripping data frames until it gets a consistent read, and then comparing against an “Accurate Rip” database to confirm my rip matched exactly with others. If I don’t get accurate results on some tracks where I would expect to, I will then clean and in some cases resurface the disc until I do. My thought is I only want to rip my collection once and I’m willing to spend the time to get it as perfect as I can so I don’t have to mess with it again later. The other benefit to DB Poweramp is it has a version of it’s application that works with automated CD Ripping machines that hold 50 – 200 CDs. I have over 10k CDs and the software allowed me to automate the whole ripping process while I was sleeping or at work.

    I’ve used both Exact Audio Copy (which I also like) and iTunes in the past. My experience with iTunes is if it runs into an error reading the disc, the results can sometimes be audible in the form of clicks or drops. Not often, but enough to be annoying, particularly with used discs. However, if the rip doesn’t have read error problems and has a confirmed Accurate Rip, then the music sounds identical to me (you can compare previous iTunes rips against the Accurate Rip database using an application within DB Poweramp’s Perfect Tunes).

    p.s… I recently purchased your DirectStream DAC and Wow!! Love it!!

  16. How do you know the copies are identical? A scientific study makes no assumptions, jumps to no conclusions. Even conclusions reached through evidence and logic are only tentative subject to being disproved at some future point. If you took two of the same factory made recordings would they be identical? Only a hunch but yes and they would sound that way. Therefore the first thing to look at is if the copies you ripped are also identical. I’ve noticed that recording at speeds greater than 1X with older software created audible glitches and this was directly from one disc drive to another in the same computer not going through the computer’s hard drive.

  17. If you have to AB test CD rips? And, must keep repeating it over and over again to see what might be better? Stop right there if the differences are not obvious. Save time.

  18. I use the rip and burn software that is included with JRiver. When I digitize vinyl, I do it using Vinyl Studio with the ADC outputting 24/96 saving as Flac files.

    In the last few years we have seen advances in digital playback where people who dissed digital and would only listen to vinyl are now starting to accept their latest comparisons with some DACs as giving them playback that is acceptable. These believed advances as best I can understand are a combination of the way the bits are processed, such as the Directstreams converting to 10, now 20 times DSD, to the Lampzinator [?] DACs that use tube based output stages along with carefully chosen DAC chips, and more recordings being created using higher resolutions. I had bought a Benchmark Dac2 as my former DAC did not have a USB input or decode DSD. This is the DAC that Waldrep uses, and for reasons I will never understand received an A+ in Stereophile’s Recommended Components. When I got the Lindemann 825, the differences were subtle but obvious. To the point that a Directstream would not be my first upgrade if finances changed. My first upgrade would be to move up the Coda line of power amps. I have been and am happy with my amp, but doubling the power I believe would give me some benefits at higher volumes.
    Anyway seeing as we are still seeing advances in the digital playback, I would never choose to use MP3 or buy a 16/44 download that was recorded at or converted from tape at a higher resolution. I would choose a minimum of 24/96. The folks who believe MP3 or CD quality is good enough, may find as their system gets better, particularly their DACs that what you aren’t hearing now may become very obvious. I believe that, if someone would have told me 5 years ago, that I would be listening to music streamed from the internet for anything other than evaluating for purchase, I would not have believed it. I now use Tidal regularly and am able to just listen without be distracted by quality issues.
    Another thing if you go back to Windows XP, the included software to rip CDs did not produce identical copies. I am not sure why, but the few times I used it, the results were not good. I made copies of two jazz CDs for Bob Crump, the man who voiced the JC1s. After he received them, he let me know that they weren’t good. So, feeling bad about it, I ordered both from Amazon and had them shipped to him. At the time my system was lacking in detail and resolution, Bob was using, if I recall correctly, the CTC Blowtorch preamp and JC1s, so when he got the store bought versions and reported very real differences, I trusted him. I only mention Bob by name as he is no longer with us. A real loss to the audiophile community. Bob was responsible for getting skeptics to try power cords, offering a recipe for a simple but effective cord. His reasoning for sharing it was that it was an upgrade, but that he sold a more refined cord, so sharing an earlier version helped the community, and would promote interest in even better results.

    Other than a small segment of audiophiles who use and believe vintage DACs sound better, I think most of us know there have been real advances in all things digital, and I don’t believe we are done yet.

  19. A lot of CD players/drives have their servo motor amplifiers adjacent to their crystals. Variations in CD labeling can cause different amounts of servo activity which result in variations in crystal temperatures that will cause overall speed variations. This was why green pens, etc caused changes in sound. Comparing a CD player to a rip could sound subtly different because the clock is always good old analog while the data is digital.

  20. Paul,
    I enjoy your blog and especially the YouTube videos, rarely do I not find something of value in them. I think the most important answer to your question is if you can’t tell the difference, why should you give a shit? Sometimes I think having my ears is a curse that creates chaos with my wife and even internally! I just received Mark Waldrep’s new book and Blu-ray disc and something significant occurred while listening to the BD – my wife bolted into the room wanting to know what it was that I was playing. More significantly, once in the living room she actually noticed the quality and depth of the audio. I thought I’d died and gone where few people think I’ll wind up. I was auditioning the 5.1 stage mix of classical instrumental music. All of the mixes are spectacular, even in the genres that I’d normally avoid were played repeatedly. I’d love to see more 5.1 recordings of better than CD quality and some honesty in terms like high resolution audio, digitally remastered, etc. I haven’t gotten very far in the book yet, but I’m sure it’ll create more than a few arguments! Thanks for all you do Paul!

  21. As often happens, the comments on this blog get me looking around the ‘net. I just came across one about mixing engineers, musicians and their systems, which seems indirectly relevant to today’s post as well:

    “I do not claim to be an audiophile. But when I’m mixing a thing, I listen like a mixing engineer. That means critical listening: I turn off my mind to outside sources, I listen deeply to the music I’m working on, I listen for very subtile [sic] differences between what I’ve got and what I want… +/- a dB or two on the bass here, a little more or less reverb on a vocal there. I have an idea in my mind, and that song becomes my whole world for an hour or two.

    “And my belief is that at least some audiophiles listen this way as well… not necessarily with the intent to fix this or that, but rather to drink in the subtilities…

    “But I don’t usually listen that way. Most of the time, I listen like a musician, not a studio guy. That’s kind of the opposite. When I’m listening in my car, at the gym, pretty much anywhere but in front of my big stereo playing from a FLAC/WAV or a Blu-ray or an SACD or, of course, mixing at my computer, I’m playing that music in my head as well as listening. There’s a saying, “musicians have the worst stereos”. Some of that’s the cliche of musicians being broke all the time, but it’s also because, if you listen that way, your brain is filling in a very large part of the song. Not what you actually hear, but the sum total of your listening experience.”

  22. I’m a bit late to the party here but I’ve read all the previous posts. If I have a simple rip from Cd to my thumb drive it’s simple mp3 at 320 fixed bitrate since it’s most compatible with my Infinity system in my Genesis Coupe. But when I have something very dear to me I do something no one else mentioned here (I don’t think) is I rip a wave file pcm 44.1 etc. I see no reason to expect more from any higher resolution. But then I make a critical listen playing the file back through Steinberg Wavelab. Here if I audibly detect simple errors like pops or clicks I manually blow up and expand the file and many times I can fix it by realigning or deleting the mismatched graphical stream image. And how many times have we come across track change errors, particularly when one song continues into the next? I fix these too. Then I save the file still as a wave and use eac to convert for my players an additional file in mp3. I now have both and use the wav master for any additional copies to any other format I might deem needed. But here’s my question. I understood that mp3 throws away data that it determines redundant but also data which reproduces ambiance. This has been my main complaint with mp3 in general. This has been verified over years of listening tests and a rather peculiar test of my own. When I m playing a file back in 2 channel and I think I’m hearing a loss of ambiance I then switch to a 5.1 surround mode with artificial surround. It seems any ambiant signal in the initial recording excites the surround outputs which can be momentarily increased in volume for examination. Lower bitrate mp3s are dead as a doornail. But higher bitrate flacs sound amazing in this surround mode with all the beautiful ambiance in the sound field. And some higher bitrate mp3s actually have intermittent dropouts in the ambiant “tracks”. This may well be revealing something of the nature of these perceived, hard to describe differences many of us experience . But when it just doesn’t matter for cheap ear buds or other poor quality play back I’ll just use Audio grabber with external lame codec at 320 variable bitrate, sometimes less. And finally what about those early cds that just sound soooo harsh? I have a Hall & Oates cd released in the late 80,s, the name escapes me, that, at moderate to high volumes can curdle fresh milk. But an online download at the higher resolution is so much better. I am assuming it is a better job done from the original masters but have no proof of it. That’s just my 2 cents on the matter and I am thankful for all the incites from all you far more educated folks.

  23. Dr. Charles Zeiliq and Jay Clawson’s four part article in The Absolute Sound, Dec 2011 thru Mar 2012 issues, titled, Computer Music Audio Quality, compared the sound quality of music on CD to music ripped by programs (CD to WAV) using listening tests.

    A score difference of 10 was deemed significant and 20-40 was a large significance. Music on CD was given a base score of 100 and the CD ripping programs placed above or below that score.

    Test results:
    140 – dBpoweramp (highest score with test comment ,”best rhythm and pace”)
    135 – Foobar2000
    130 – JRMC
    120 – Easy CD-DA Extractor
    110 – Exact Audio Copy
    100 – music on CD (base line)
    95 – iTunes (with test comment, “less detailed”)
    85 – Media Monkey
    75 – Winamp
    60 – Nero

    So, music ripped in iTunes fell a bit below CD in audio quality. Your friend Jason Serinus may be onto something.

    It seems to me that this article had little influence. At least I never saw it referred to in my stereo magazines or on the internet or heard anyone mention it at audio shows and stereo club meetings I attended. To this day, I have never found out who Dr. Charles Zeiliq and Jay Clawson are.

    BTW, your company was mentioned in the article. Dr. Zeiliq stated in regard to sound quality improvements related to the computer equipment that, “The largest improvement occurred when we replaced the standard computer power cord with a top quality PS Audio Statement cable.”

  24. When I read the subject line of this post I got the impression you were doing some heavy partying in that cool listening room at the PS Audio factory. Then I read it and it seems to be about some sort of digital audio or some such. Disappointed!

  25. I’m no digital audio maven but Beethoven could hear the difference between original CDs and WAV or FLAC rips using JRiver MC23 with the rips winning hands down.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram