Skip navigation
Currently Being Moderated

Why Audition CC don't show sample rate correctly?

Sep 23, 2013 1:58 PM

Hi

Using Audition CC on Mac OS X with Mackie 1220, 12-channel premium analog mixer with integrated 24-bit/96kHz Onyx FireWire I/O (Quality is amazing!)

Need to record on mono a horrible old tape recording with voice only to "hero"'s repairing audio and configure the best sample rate for recording.

My preferences Audio Hardware on Audition are:

Device Class: CoreAudio

Default Input: Onyx Firewire (0025)

Default Output:  Onyx Firewire (0025)

Clock Source: Device

I/O Buffer size 512 Samples

Sample Rate: 96000 Hz

 

 

Question 1: Expert said do not use more than 48000 Hz for recording voice only and use mono. Humans can't listen more than 20000 Hz... well... My recording on 96000 Hz is better than 48000 Hz and I'm not Superman... and can hear better difference with 96000 Hz recording. Don't know why.

 

Question 2: Why after selecting Audio Hardware preferences to 96000 Hz after recording the File Panel said my track is 48000 Hz Mono 32 (Float)? My hardware can record at 96000 Hz and preferences is set correcty at that Bit Depth then WHY don't appears 96000Hz on File Panel after recording and instead appears 48000Hz?

 

Question 3: Don't understand the differences of my 24-bit/96kHz Onyx FireWire I/O and 32 bit (Float) of Audition. If I record using 24 Bits (Maximum bit depth of my hardware) why appears 32 bit (Float)? My recording are 24 Bits or 32 Bits?

 

Thanks :-)

Tom

 
Replies
  • Currently Being Moderated
    Sep 23, 2013 4:05 PM   in reply to Tom Adobe

    Some answers to questions 1 and 3 may be found in this thread:

     

    http://forums.adobe.com/thread/1291155?tstart=60

     

    If Audition is reporting your audio file as 48K then it probably is. Were you recording it in Waveform or Multitrack view? If in Waveform make sure that you opened a New file as 96K to record into. If Multitrack then make sure that your Audition session is started as a 96K one. Not that you should need to use 96K for recording an old tape recording whose frequency response probably doesn't go any higher than 10kHz even with a well set up recorder.

     
    |
    Mark as:
  • SteveG(AudioMasters)
    5,602 posts
    Oct 26, 2006
    Currently Being Moderated
    Sep 24, 2013 1:48 AM   in reply to Tom Adobe

    Tom Adobe wrote:

     

    But is not clear to me which is the best way to record old tape of voice only to repair. For example, if records is not higher like 10kHz, it's the same or not to record on 48-96kHz to get better result when repair audio? Similar like images when you have better resolution to repair or fix something get better results? Want to understand the concept on audio.

    You don't get a 'better' result using 96k unless there's something seriously wrong with your system at 48k. If it's an old tape and you are going to end up with a result on a CD, you should be using 44.1k anyway.

     

    Using 96k doesn't give you any more resolution at all - it just enables you to store a load of high frequency noise that no human can hear. You don't always get better results with higher resolution images for editing either - once you've reached the pixel limit of what you're editing, you don't gain a thing. Same with audio - everything that adult humans are capable of hearing (and a bit more) can be captured accurately and fully at 44.1k. There is no more 'detail' to be had, and even if there was it wouldn't be a useful part of the sound.

     

    Recording at higher sample rates used sometimes to make a difference when old converters with brick-wall anti-alias filters were used, but these fell out of use ages ago to be replaced with internal oversampling systems that don't have artefacts at the top of the spectrum at all. So as I said, if you have the means to detect a difference between the sample rates, then something's wrong with the lower rate recordings - but it isn't the sample rate itself that's causing this.

     
    |
    Mark as:
  • SteveG(AudioMasters)
    5,602 posts
    Oct 26, 2006
    Currently Being Moderated
    Sep 25, 2013 3:11 AM   in reply to Tom Adobe

    Tom Adobe wrote:

     

    OK, I'm not audio technnical guy and can't discuss your concepts but my ear shows me quality differences. For example if I hear 48kHz recording is better that audio CD quality of 44.1kHz and why if human can hear only from 20Hz to 20kHz and less for older humans?

     

    I told you why in my last paragraph above. If you don't understand it, then I'm afraid you'll have to take my word for it. But I should point out that the research into what you can and can't hear at different sample rates is academically sound, tested and verified, so I simply don't believe you - based on excellent and credible evidence. So what top-end monitors are you using when you claim to detect these changes?

     

    Lots of people claim (or claimed) to be able to tell the difference, but in controlled tests they simply can't - as has been thoroughly proven.

     

    Let's look at your image comparison. If you have a 72 dpi image, then scanning and editing detail at 600 dpi is going to add artefacts to the 72 dpi original. The audio equivalent would be the use of a harmonic exciter. If used in small doses on bandwidth-limited material it can, by a piece of auditory trickery, make things sound a little clearer, but it's not actually adding anything that was in the original - just distortion. In a picture, a similar process (adding edge detail, etc) simply makes it look artificial - not an improvement. The difference though (and it is significant) is that your eyes can easily see this - the same thing in audio doesn't work because your ears can't percieve the changes you make above their resolution limit. If you make the comparison in visual terms, it amounts to making changes to the minute granular structure of whatever your image is printed on - not going to alter the image, just the background to it - which is pointless.

     

    My original statement about this stands, and whether you agree or not is irrelevant, I'm afraid - these are demonstrable facts.

    Tom Adobe wrote:

     

    Can apply this image's concepts on audio to better "cosmetic surgery" repairing old and horrible original audio of old tapes?

    That would not be a repair - by definition, as you are not reinstating what was originally there. If you applied a harmonic exciter in small doses, then you could do this without needing to alter the sample rate in the slightest, and you might perceive a slight improvement. It is very easy to overdo this though, and it gets rather hard on the ears in a very short time.

     
    |
    Mark as:
  • SteveG(AudioMasters)
    5,602 posts
    Oct 26, 2006
    Currently Being Moderated
    Sep 25, 2013 3:25 AM   in reply to SteveG(AudioMasters)

    If you want to look at a powerpoint slideshow about some of the audio research, it's here. There's also additional research, and the results of the ABX tests, but the link for that is currently broken.

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 19, 2014 8:22 PM   in reply to SteveG(AudioMasters)

    CS6 has a serious issue with saving files correctly. The program is asuming that 48kHz is the maximum you will be using and in my case it saved a 96kHz recording with a 48kHz internal header. The file size is consistent with all my previous 24/96 recordings and it sounds just fine interpreted correctly - but played an octave low in frequency and tempo it really sucks unless you are a Blue Whale..

     

    I can play it "interpreted" as a 192kHz file just fine, and it now sounds 100% right, but I cannot save it correctly. I have yet to find out how to recover because when I use "convert Sample Type" it saves it with the same mistake - the wrong sample rate off by the same ratio again.

     

    It is a program flaw - so at this point you cannot record with sample rates higher than 48kHz in CS6 and depend on your file being OK.

     

    Tom the reason you can tell the difference between sample rates is that your hearing has two dimensions, frequency and timing, I sure hope you can't hear the bats at night but I expect that you can tell a good drummer from a bad drummer. In addition there is also the issue (dimension) of bit-depth - instrument decay and acoustic space occupies the time space between notes and if it is not sampled at the right time and place or rounded off to the nearest digital digit you have a problem.

     

    You all know that some humans have perfect pitch and others dont, this gives you some indication how much we each differ. Some people have even learned to use echolocation; the best know cases being blind people because they are not supposed to be able to find their way and know where they are. You can learn underlying concepts of this little discussed aspect of human hearing here and hit the university libraries for the rest. http://en.wikipedia.org/wiki/Human_echolocationhttp://

     

    SteveG's comment are only true in one of the three dimensions - frequency range. We all hear about the same for starters anyway.

     

    The second dimension, timing, is what provides spatial information. Coarse sampling effects this as well. My most recent experience was when transcribing some old casette tapes - when I experimented with FLAC and WMA lossless I found that they din NOT downsample well. Spatial information was significantly deminished be that aucoustic or studio work. Needless to say this surprised me because I was auming that I could drop the samplerate to 48kHz as soon I was done editing and save a lot of drive space. For now I made some excellent 48kHz/24 bit mp3s (320) because they altered the sound the least.

     

    Now you know some additional reasons why older guys like me who have lost the high end and a lot of decibel as well can still tell the difference:we have good sense of timing.

     

    Anyway - I need to learn how to edit my files to reset the sample frequency header - real fast. I just recorder a fabulous Madrigal group for 2 hours and my files are lost in the vortex.

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 20, 2014 4:34 AM   in reply to alextp

    To anyone hearing a difference by using 96 or 192 kHz, are your listening tests done blind, i.e. with somebody controlling what version you're hearing and not letting you know what's been suggested.  If not, the "audio placebo" effect is a powerful thing.

     

    As for the "coarse sampling" argument, watch this:  http://www.youtube.com/watch?v=cIQ9IXSUzuM

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 20, 2014 8:38 AM   in reply to Tom Adobe

    Tom Adobe wrote:

     

    For example if I hear 48kHz recording is better that audio CD quality of 44.1kHz and why if human can hear only from 20Hz to 20kHz and less for older humans?


    Hi Tom,

    If you hear recordings at 48kHz and 96kHz as better sound to your ears, it is because when recording is done at 44.1kHz frequencies 1khz - 20 kHz have been recorded close to 0dB level, your recording at 96kHz has that part of spectrum (1kHz - 20 kHz) recorded at PINK NOISE level (difference +6dB at very top of spectrum). In simple words, your spectrum what you listen is still let's say 20Hz-20kHz, but high frequencies have been recorded louder. If you know a little about mastering, you can get same sound with EXCITER or just simple EQ. I do agree with you that MP3 done at 320/48kHz sounds best on your mobile device, and recording at 48khz was done in recording studios for ages, somehow RED BOOK believers will always tell you 44.1 is the only way to go. Can't help you with rest of your question since i still use much older version of Audition (and it does, what i need without any problems......  

     
    |
    Mark as:
  • SteveG(AudioMasters)
    5,602 posts
    Oct 26, 2006
    Currently Being Moderated
    Jan 20, 2014 9:47 AM   in reply to nahravka

    nahravka wrote:

     

    If you hear recordings at 48kHz and 96kHz as better sound to your ears, it is because when recording is done at 44.1kHz frequencies 1khz - 20 kHz have been recorded close to 0dB level, your recording at 96kHz has that part of spectrum (1kHz - 20 kHz) recorded at PINK NOISE level (difference +6dB at very top of spectrum). In simple words, your spectrum what you listen is still let's say 20Hz-20kHz, but high frequencies have been recorded louder.

    No truth in that at all. Not a shred...

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 20, 2014 11:43 AM   in reply to SteveG(AudioMasters)

    Problem Solved: With a work-around!

     

    I recorded 4 tracks at  24/96 of a Madrigal group in a small church. So far so god everything worked correctly and playback was very consistent.

     

    The files were saved as a multitrack session but the resulting file did not play correctly. Interpreting it as a 192 kHz rate worked for playback because it was a 96 kHz file that Adobe thought should be played at 48 kHz (the multitrack max in CS 6).

     

    Converting the file and re-saving it did not work because Adobe included the wrong multiplier in the new file.

     

    The solution:

     

    • Open a new file with the correct settings in the editor.
    • Open the damaged file and choose the right Interpretation.
    • Copy and paste the damaged file into the new file and save it.

     

    You now have the correct header and the same content.

     

    I won't be staying for the argument about why some people can hear what others can't. One interesting fact is how the HDCD standard squeezes extra bit depth out of an ordinary 44.1 CD and improvers the sound as least some - you probably have some lying around. It was an improvement but it never caught on because the music industry was greedy and wanted SACD so we all had to buy new copies of our disks.

     

    I don't destibute what I record at 24/96 but it gives me the headroom I need to manipulate the files without a percieved loss.

    I like the 24bit/48kHz mp3 at 320 because the loss is minimal and playback is not dependent on expensive CD players to sound good

     
    |
    Mark as:
  • SteveG(AudioMasters)
    5,602 posts
    Oct 26, 2006
    Currently Being Moderated
    Jan 20, 2014 1:23 PM   in reply to alextp

    alextp wrote:

     

    I won't be staying for the argument about why some people can hear what others can't. One interesting fact is how the HDCD standard squeezes extra bit depth out of an ordinary 44.1 CD and improvers the sound as least some - you probably have some lying around. It was an improvement but it never caught on because the music industry was greedy and wanted SACD so we all had to buy new copies of our disks.

     

    If you dithered a CD correctly you could get virtually the same bit depth out of a normal CD as you could out of the clearly distorted HDCD with its weird amplitude and gain coding. And there won't be an argument, as there's nothing to argue about - whatever you might mistakenly think.

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 20, 2014 1:58 PM   in reply to SteveG(AudioMasters)

    In apologize for having bruised your ego with information that is outside your permanent and perfect knowledge.

     
    |
    Mark as:
  • SteveG(AudioMasters)
    5,602 posts
    Oct 26, 2006
    Currently Being Moderated
    Jan 20, 2014 2:00 PM   in reply to alextp

    alextp wrote:

     

    In apologize for having bruised your ego with information that is outside your permanent and perfect knowledge.

    Well by your own definition, that's clearly not possible!

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 20, 2014 6:39 PM   in reply to alextp

    alextp wrote:

     

     

    I won't be staying for the argument about why some people can hear what others can't. One interesting fact is how the HDCD standard squeezes extra bit depth out of an ordinary 44.1 CD and improvers the sound as least some - you probably have some lying around. It was an improvement but it never caught on because the music industry was greedy and wanted SACD so we all had to buy new copies of our disks.

     

    I don't destibute what I record at 24/96 but it gives me the headroom I need to manipulate the files without a percieved loss.

    I like the 24bit/48kHz mp3 at 320 because the loss is minimal and playback is not dependent on expensive CD players to sound good

     

    HDCD doesn't "squeeze" extra bit depth.  It uses dithering, filtering and a form of companding to "fake" extra bit depth.  The trade off for all this processing is a slight increase in noise.

     

    Yeah, many (even most) people can hear a difference between ordinary CDs and HDCDs but "different" is not the same as "more accurate".  The audio world is full of contraptions that people find subjectively more pleasing to the ear but this is not the same as being a better reproduction.  For example, many home hifi speakers are examples of this.  They're designed to make a variety of programme material sound good but this makes them useless as studio monitors since the sound is artificial.

     

    The 44.1/48/96/192 argument is different though.  Some claim to hear a difference but I have yet to find anyone who can hear that difference in a true double blind trial.  The human brain can and does let you hear what you think you should.

     

    To then go on and say you prefer MP3 quality, even at 320kbps, to uncompressed CDs proves to me you're more concerned with your subjective view of the sound you like rather than what is the most accurate reproduction of the original.

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 20, 2014 6:41 PM   in reply to SteveG(AudioMasters)

    I stand corrected:

     

    Saving as described above, the files comes out half size.  In other words the Audition Fixed Session Spec of 48 kHz is carried forward making it necessary to interpret the sample rate as double. The problem may be caused by a structural error one level below the header or a proprietary RAM format..

     

    I tried the Rail Jon Raguts Header Investigator ( http://www.railjonrogut.com/HeaderInvestigator.htm) and the files still need to be set to 192 kHz to play right in other players like VLS, Win Media Player and for that sake Audition.

     

    The one working strategy that let me work on the complete dataset ( The file is 6144 bitrate = 24/96) is to Interpret thefile as 192 kHz and output it as mp3 - that strangely enough makes a file with the right bitrate, time and size.

     

     

    I thing I'll just feed the mics straight to my little Roland next time.

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 20, 2014 6:59 PM   in reply to Bob Howes

    Bob you may very well be right about HDCD - I only bought stuff I like suck as Mark Knopfler who always had god sound. It was an attempt at moving forward that as you say has been supeceeded by better filtering, dithering and more than anything much better DACs.

     

    About mp3s they are not my first choice but when I give people a 320 mp3 that sounds good they won't rip it at 128 like they do from CD.

     

    Hard Drives are getting cheap so lossless at a 24/48 will be quite nice - I don't have that much stuff I would play at 24/96 - I just like editing and storing at that rate. I just recorded an acapella Madrigal group and the Sopranos are quite capable of overloading anything including the room and most peoples ears. It is important to leave at least 10 dB headroom when the room starts ringing so high sampling leaves me a bit more of a margin.

     
    |
    Mark as:
  • Currently Being Moderated
    Jan 21, 2014 5:50 AM   in reply to alextp

    alextp wrote:

     

    I just recorded an acapella Madrigal group and the Sopranos are quite capable of overloading anything including the room and most peoples ears. It is important to leave at least 10 dB headroom when the room starts ringing so high sampling leaves me a bit more of a margin.

    But that is all to do with bit depth (24 bit is good here) rather than sample frequency (48kHz is quite enough).

     
    |
    Mark as:

More Like This

  • Retrieving data ...

Bookmarked By (0)

Answers + Points = Status

  • 10 points awarded for Correct Answers
  • 5 points awarded for Helpful Answers
  • 10,000+ points
  • 1,001-10,000 points
  • 501-1,000 points
  • 5-500 points