Your Vinyl Transfer Workflow (sharing best needledrop practices)*

Discussion in 'Audio Hardware' started by Vocalpoint, May 11, 2011.

  1. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    I totally missed that when you posted it before. So cutting 5hz and down solves DC-offset. I just tested it and seemed to work.
     
  2. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Yes, DC offset is a "tone" at 0 Hz so removing everything from 5 or 10Hz downward removes it quickly and easily. There are plugins that do it as well but since you're using RX and it's so simple to remove with a quick selection, there's really no need.
     
  3. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    What is it caused by?
     
  4. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    There are several explanations around the net but basically it's almost always a voltage introduced into the signal somewhere in the analog chain before it gets converted to digital. It could be the interface, the phono stage, etc. At the percentages you're getting, it's really of no concern. Heck some of the 80's CDs some folks around here treasure above all others have 0.5% or more DC Offset and that didn't stop them from sounding good to some folks. There's a decent write-up on Dc Offset here that doesn't get too technical: DC offset - Audacity Manual . Again I really wouldn't get preoccupied by it. I know we all want to get things to as good as they can be, but since at the percentage you're experiencing there's no way it's impacting the audible material in your needledrops and because it's so easy to remove with a simple select and delete, don't sweat it!
     
    ghost rider likes this.
  5. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    Very helpful link thanks.:righton:
     
  6. BendBound

    BendBound Forum Resident

    Location:
    Bend, OR
    Now I want to ask about use of de-click in IzotopeRX7. I have not been using it, instead doing manual repair on spot clicks when I encounter them with the instant process de-click tool. ClickRepair is a program I was using, while when I do use it, set it at 6 to 10, reverse processing. But over the last year, I have elected to do manual by ear de-clicking only with the "instant process" approach on identified clicks. That process unfortunately takes me many hours over several days to complete, particularly on older records. I'm looking for other options.

    Today I read what the program says about settings, from the manual and on-line. Then I did a series of experiments on a section of an Aretha Franklin track from a 1972 record in VG+ condition. The record, which looks stellar, had been cleaned in an ultrasonic tank and in a VPI17. But with a very sensitive phono cartridge, I find I am picking up a lot more noise as well as a great signal.

    Looks like some of the better settings – assuming I wanted to scrub an entire recorded lp – are as follows:

    1. multi-band random setting,
    2. sensitivity between 1.0 and 2.0,
    3. click widening at 1.5 to 2.5 ms,
    4. frequency skew, 0.0.

    I did not experiment with frequency skew, and I should have.

    Now, from those of you who are familiar with this tool, have you developed preferences, or should I say, initial default settings?

    Presently, I employ a routine BrilliantBob mentioned here. Once I process for hum, I 'mix' white Gaussian noise at- 85dB. Later I use spectral de-noise at as low a set of setting as I can see. That removes a lot of background gunk, track noise, rumble, etc. I sample a section that is relatively clean and try not to over process that function when I apply it to the entire recorded track. The results are quite good, if not amazing.

    But I have yet to employ entire track processing using de-click and wanted to get some feedback from your collective experiences.

    Finally, when should de-click be run: Before or after the routine I have adopted from BrilliantBob?

    Thank you all in advance.
     
    Last edited: Feb 26, 2021
    gabbleratchet7 likes this.
  7. Grant

    Grant Life is a rock, but the radio rolled me!

    I find that the default "vinyl record" declick settings works great for me. I'm amazed at how well it works. I don't really need to tweak anything except maybe the sensitivity on more difficult records.

    I always run declick as soon as I finish the transfer before anything else.

    I should try out BrilliantBob's denoising tequnique.
     
    Last edited: Feb 26, 2021
  8. marcob1963

    marcob1963 Forum Resident

    If you're applying a High Pass Filter as part of your process, then that will eliminate any DC Offset.
     
    Grant likes this.
  9. Grant

    Grant Life is a rock, but the radio rolled me!

    @BrilliantBob I'm impressed with your Gaussian noise trick. I won't say it takes care of everything, but it not only makes Spectral noise removal easier, but doesn't appear to harm any music. I tried it out on an old 45 and it seems to leave the reverb tails intact.

    whatcha see.wav
     
    ghost rider and BrilliantBob like this.
  10. BrilliantBob

    BrilliantBob Select, process, CTRL+c, CTRL+z, ALT+v

    Location:
    Romania
    I inject (mix) the Gaussian noise with Signal Generator at -100dB and then I remove it with Spectral De-noise with 40dB reduction at max quality settings. I also use the de-clicker to remove some background noise, LOL, with Low-latency algo, Freq skew -2.3, Sens 1.8 Click widening 0.0.

    test.wav

    [​IMG]
     
    BendBound, Grant and ghost rider like this.
  11. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    I missed the discussion on Gaussian noise. I would like to give it a try if my brain can figure out the steps. How far back was the post Bob?
     
  12. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Same here, altough I avoid using the default strength of 5 on transients as it can soften them. I have shortcuts set up for using lower values and try them first and only if there's still audible clicks (or bumps that are left over from declicking at too low a setting), I unzo and then try a stronger setting. If you set up a preset using the Vnyl Record setting but boost the strength to 7, it works really well at removing thumps and just about anything else, but it will really soften transients, so that's only for cases where nothing else works (Interpolate is another option, but sometimes stuff is too wide for it).

    I still like Clickrepair for some basic automatic declicking. Too bad to read that its creator has health issues that look to be terminal (according to another thread on here). I find that when one listens to the click only output, Clickrepair sounds like it's removing only clicks whereas RX sounds like there's a bit of phasing/artifacting going on. In the end, if the music sounds ok, it`s fine, but it's still a bit unnerving to hear some of the swooshy stuff in the clicks only output!
     
    BendBound, Grant and BrilliantBob like this.
  13. BrilliantBob

    BrilliantBob Select, process, CTRL+c, CTRL+z, ALT+v

    Location:
    Romania
    Your Vinyl Transfer Workflow (sharing best needledrop practices)*

    For hard noised needledrops lower the gaussian noise level (eg. from -100dB to -85dB) but keep an eye opened to the audio quality.
     
    ghost rider likes this.
  14. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    So you mix -85db white Gaussian noise then denoise? I'm trying that now.

    How hard can you hit it? I'm at level 15 for my trial run.

    Here is a sample my archived before denoise and denoised using mix Gaussian noise.
    Dropbox - Herbie Hancock sound system.flac - Simplify your life
     
    Last edited: Feb 27, 2021
    BrilliantBob likes this.
  15. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    My first thought is this is the way to go. I just don't know if an artifacts are produced.
     
    Last edited: Feb 27, 2021
    BrilliantBob likes this.
  16. Grant

    Grant Life is a rock, but the radio rolled me!

    The second sample came to life.
     
    BrilliantBob and ghost rider like this.
  17. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    No artifacts? I'm guessing not. This is pretty amazing. Also to be fair the 2nd sample Loudness was used to raise the levels up.
     
  18. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    That sounds good but if you lok closely at the spectral view, there are spots where the ambience is getting chopped off as it fades out.

    Here's an example from your file:
    [​IMG]

    You can reduce or eliminate this by opening the Advanced Settings in the Spectral De-noise module and trying to adjust the release settings.

    Also, I've been experimenting lately with doing more than one pass of noise reduction on problem areas. I just tried doing so on the first half (untreated) section of your file. Here's the result:

    Dropbox - Herbie Hancock sound system_ST.flac - Simplify your life

    Here are the settings I used for the first pass:
    [​IMG]

    and the 2nd:
    [​IMG]

    Note that this was done just Learning from the section of your file between 2.9 and 3.0 sec as it seemed to be pretty clean. Notice as well that I didn`t hit very hard, only 4.0 Reduction, but I also used the Reduction curve (blue line) with smoothing set at 4.3. I adjust this to that the yellow Residual noise line is pretty much level with the higher frequencies of the darker orange line, so it's basically removing the hump at the bottom. I didn't use M/S for this as when I tried I found there was as much low-frequency noise in the Mid channel as in the Side. I also didn't use the Gaussian White Noise method. It can be interesting, but I'd use Pink Noise (or even brown) as they will be much more effective for lower frequency noise. White noise can be good for reducing hiss but a lot of the noise that`s distracting on vinyl comes from the lower frequency whoosh of vinyl and any problems with rumble. I was also doing some testing earlier today using White Noise but instead of just taking an arbitrary number of `85 or -100, I selected a high-frequency patch I an area wanted to remove, checked its Total RMS level in the Waveform stats to find it was at -63. So I took a blank track, added White Noise at -63, Learned from it with Spectral De-noise, then applied that to the problem area. It worked really well in removing the patchy noise. As I write it occurs to me that Triangular White Noise might be a better choise. I recall it's always the preferred choise for flat dither, but I can't recall the reasoning behind it. Time for some research. Nothing like being an armchair audio geek during a pandemic! :)
     
    ghost rider likes this.
  19. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    I don't hear artifacts per se, but there's a sense that the audio drops out abruptly (see my previous post). It can give that "vacuum chamber" effect that a lot of folks complained about when noise reduction became too heavy handed on commercial releases in the 90's and into the new century. In fact, I recall a story that Bob Ludwig told about adding low-level pink noise to a track and playing it back to people who said it sounded better than the completely dry track.
     
    ghost rider likes this.
  20. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    By the way, there's another interesting RX denoise method I sometimes use that brilliantbob described here awhile back. You use the Ambience Match module, on a noise-only section, then render it with "Render ambience only" checked. Then use that to Learn the Spectral De-noise profile and afterwards the section you just changed to rendered ambience. This can give a nice sort of average ambience to Learn for denoising. This can be effective for denoising sections when there's not a lot of variation in the noise.
     
    BrilliantBob and ghost rider like this.
  21. Grant

    Grant Life is a rock, but the radio rolled me!

    The life usually happens when a layer of grunge is removed. That life is like someone lifting the music off and away from the vinyl record. It gets more 3-dimensiality. That's the only way I can describe it.:shrug: But, you have to be careful. If it sounds too 3-D, you're probably removing some bottom.
     
    ghost rider likes this.
  22. Grant

    Grant Life is a rock, but the radio rolled me!


    I wonder if RX plans to continue to sell licenses for the individual RX and Ozone programs. certainly you all got the email for the new suite subscription for $24.99 a month? I swear, if they go the way of Adobe...
     
  23. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Yeah, that would suck.
     
  24. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Mind you, as part of my work, I have a complete Creative Cloud subscription. I hardly ever use Audition (haven't really since 3.0), but it is there if I want it. I like the subscription model for always getting updates (I use Photoshop, Lightroom and Illustrator for work and a photography hobby).
     
  25. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    So you ran denoise twice? with each of those settings or did you run 2 different tests? Also I don't see any changes between 1st and 2nd passes.
    It sounded great so I'm duplicating it on the whole file. The release time was grayed out on my RX 6.0 advanced. I'm still cautious about using denoise I'm pretty much in over my head at this level.
     
    Last edited: Feb 27, 2021

Share This Page

molar-endocrine