Announcement

Collapse
No announcement yet.

How do I clean up Edison Blue Amberol Cylinders?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • How do I clean up Edison Blue Amberol Cylinders?

    I have followed the directions for using the continuous filter but no matter where I move the blue squares the preview sound stays the same. The same with the easy-clean filter. Did I forget something in software setup?

  • #2
    First, you must de-click the cylinder with aggressive settings of all three filter controls of the EZ Impulse Filter (up between 70 to 95 for each, depending on the condition of the record).

    Once that is done, set the CNF mode to normal. Set the attack and release to around 100. Open a file. Find an area in the file (five or so seconds) that is just noise. Highlight it with your mouse. Next, click the sample noise button on the CNF. Now, preview the file and adjust the attenuation until you are satisfied with the level of noise reduction. When you are pleased with the sound, run the filter. Done.

    Let us know if that did the trick.

    Thanks,

    Craig
    Last edited by Craig Maier; 11-16-2017, 10:31 PM.
    "Who put orange juice in my orange juice?" - - - William Claude Dukenfield

    Comment


    • #3
      The effect from the Continuous Noise filter was noticeable but didn't remove the noise. I have tried to upload small clips of Before and After in WAV and MP3 but I keep getting an error that says 'invalid file'. How can I upload these clips to you?

      Comment


      • #4
        First, to increase the aggressiveness of the filter, you can raise the overall threshold (shift threshold) via it's up and/or down arrow. If you want to send a snippet to me, I can clean it up for you. Send a short snippet of the file to my personal account at dctools@aol.com. Keep he file under 10 MBytes so my emaill system can handle it. Make sure that the snippet includes a section having only noise. You can probably just send it as a wave file (44.1 kHz, 16 bit, stereo). I will look it over and then give you settings that work.

        Craig
        "Who put orange juice in my orange juice?" - - - William Claude Dukenfield

        Comment


        • #5
          Hi,

          I see why you were having difficulty with that cylinder recording. It took six different filters to produce a result. I sent the result to you via email. If you like it, let me know and I shall publish the filters, parameters and steps that I used.

          Craig
          "Who put orange juice in my orange juice?" - - - William Claude Dukenfield

          Comment


          • #6
            Here is the procedure that I used to clean up your Edison Blue Amberol Cylinder: (Note: you can hear the result of each step by Previewing before running if you want to modify my settings). Also note that you need to perform the steps in order. Using a different order will produce a different (less desirable) result.

            1. Applied the File Conversion Filter (Filter Menu) from Stereo to Mono L+R. Very Important Note - your particular transfer was made using a stereo microphone and not a phono cartridge. Had a stereo phono cartridge been used, then I would have used Stereo to Mono L-R because cylinders are vertically modulated.

            2. Applied the EZ Impulse Filter with the following settings: Scratch: 91, Crackle: 91, Narrow Crackle: 94, Big Clicks: Checked Preview to hear the results and then run the filter.

            3. IIR Bandpass Filter Applied: Set for Butterworth, 24 dB / Octave, Low Frequency: 140 Hz, High Frequency: 4200 Hz. Preview to hear the results and then run the filter.

            4. Continuous Noise Filter (CNF) Applied: Set fft size for 4096, Set Mode for Normal CNF, Attack: 40 mSec, Release: 80 mSec, Smoothing turned on via checkbox, Now, goto the lead-in noise portion of the file and highlight about 5 seconds of the noise with the mouse. Then, click Sample Noise on the CNF. Next, ratchet up the threshold line of the CNF using the up arrow feature in the CNF (I used 10 clicks - more clicks has greater effect and vice versa). Set the Attenuation for 8. Preview and see if you like it. If you want more noise rejection, increase the Attenuation control up to 10 to 12 (or whatever you like). Note: too much attenuation will produce artifacts. When you are satisfied, run the filter.

            5. Dynamic Noise Filter: Set the Noise threshold to 5000, Filter Frequency to 2200, Attack: 100 mSec, Release: 300 mSecHigh Frequency Gain : -8.0 dB, Enhancer Mode disabled (not checkboxed). Preview and tweek the parameter to your taste and then Run the filter.

            6. I did not perform the following steps, but you can consider doing these. Bring up the 20 Band Graphic EQ and adjust the controls that will have an effect on your results which will be the controls between 177 and 5.6 KHz. Set the other sliders that exist from 22 to 125 Hz all the way down. Do the same thing with the sliders from 8 kHz and 16 kHz. Just work with the ones between 177 annd 5.6 kHz.

            7. I also did not perform the following step, but it is recommended at the end of the process. Goto the CD Prep menu and apply the Gain Normalize function.

            8. Done.

            Craig

            ps - if you become satisfied with this filter lineup, it can be made into one single filter via the Multifilter function in the Diamond Cut Software.
            Last edited by Craig Maier; 11-16-2017, 10:32 PM.
            "Who put orange juice in my orange juice?" - - - William Claude Dukenfield

            Comment


            • #7
              It is worth noting that cylinder recordings (ie. Edison 2 minute wax or 4 minute Amberols) are the toughest restoration projects. I noted that the overall signal to noise ratio on the sample that you sent over was around 6 dB in the raw form. Ultimately, I was able to improve it up to around 20 dB. That is because of the basic technology (acoustic recording - no amplification) coupled with the high degree of degradation of most cylinders (they were often played over and over again with a relatively heavy stylus). If they were of the earlier 2 minute wax type, often they were stored when then became out of vogue in damp basements. As such, the wax surfaces often grew mold on the grooves, making the noise problem even tougher.

              The Blue Amberol in question here on this thread was from 1919 and was 'overplayed' by it's owners and not treated very nicely, making the restoration tricky, but possible. It just took more steps than one would generally expect for a normal old acoustic recording.

              Craig
              Last edited by Craig Maier; 11-20-2017, 03:07 PM.
              "Who put orange juice in my orange juice?" - - - William Claude Dukenfield

              Comment

              Working...
              X