Your Vinyl Transfer Workflow (sharing best needledrop practices)*

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

  1. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Just thought I'd give a heads up to anyone looking for a decent EQ plugin. The Tokyo Dawn Labs plugins are all on sale at really good prices until Monday at Plugin Boutique. I've bought a couple of things there without problems.
    You can try the Tokyo Dawn Labs plugins for free without limitations, they just won't save presets (a true analog experience!). I've tried a couple of their EQs now and quite like how they sound and their features (tilt EQ, Equal Sound Contour feature, etc.).
     
    Stan94 and ghost rider like this.
  2. Grant

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

    I meant to write that I have 20 years of data on USB thumb drives. Anyway...

    NVME M.2 drives are great, aren't they? I have slots for two of them on my motherboard which is now standard. I'd like to see manufacturers try to slip in three! I can see someone removing GPU card support on a model to make room for an extra one, as onboard graphics become more common on the CPU itself. I use a CPU with graphics and don't really need a GPU right now, especially when they are almost impossible to obtain, and if you do find a new one, it can cost $700+. Crazy times! Thanks bitcoin miners...not!:realmad: Onboard graphics is common in Intel chips, but not on AMD.

    I use Samsung NVME M.2, but i'm seeing a lot of good stuff about Sabrent drives, too. I haven't heard much about Western Digital Black. It's interesting that I never see anything about internal Seagate SSD. They make them (Ironwolf), I just never see them.

    You're 61? I thought you were younger than me. Oh well, i'm in that same age range, near the end of the post-WWII boomers.
     
    Stefan likes this.
  3. Stan94

    Stan94 Forum Resident

    Location:
    Paris, France
    It's funny how needledropping with good equipment and clean records can make one pick up details that were always there but somehow burried in the music. For instance, I'm now hearing a great deal of lip-smaking on Beatle vocals that I'd never heard before. For good measure I check my rips against the 1987 CDs or good MFSL needledrops, so that I don't take out something that's part of the master. Now I ususally prefer the sound of German vinyl compared to the clinical sound of the old CDs.
     
  4. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Yes, I love the NVME M.2. I was thinking I could put one in my needledrop computer using an adapter since it doesn't have an M.2 slot (it's from 2012), but it turns out the BIOS won't support it without using some hacked BIOS that's kicking around the net. It's not the end of the world as I do have a SATA SSD in it as my boot drive and it's nice and fast. Plus I still intend to build a new system to replace it later this year.
     
  5. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    I think part of it is because we're so focused on the needledrop sound and process so we pick up on details we'd otherwise not notice.
     
    arisinwind and ghost rider like this.
  6. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    How many of us would leave the defects baked into the mastering?

    I'm recording Mountain Climbing and Mississippi Queen has a ground loop. I'm pretty sure the last time I removed it.
     
  7. Grant

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

    My old 2010 Gigabyte board's BIOS didn't support SSD either, but it worked anyway. The BIOS just flashed a little warning on boot-up and continued to load Windows.
     
  8. Grant

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

    I remove them if I can. They didn't have the technology back then, but we do now.
     
    arisinwind, Exotiki and ghost rider like this.
  9. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    When I can remove something like AC hum or a tape glitch without harming the needledrop I always do.
     
    arisinwind and ghost rider like this.
  10. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Oh mine has no problem supporting the SATA SSD, it just won't support an NME M.2 in an adapter installed in a PCIe slot.
     
    Grant likes this.
  11. Exotiki

    Exotiki The Future Ain’t What It Use To Be

    Location:
    Canada
    If the defect was obviously unintended and the fix is significantly less objectionable than the defect. Yes, I will try and fix it.
     
    ghost rider likes this.
  12. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario
    My Uncle Dave had a unique tactic for selling CD players back in the early/mid 1980's. He would throw a CD on the floor, step all over it with his shoe, place it back in the player and PRESTO the CD would play perfectly.
     
  13. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario
    Mmmmm......Sad!
     
  14. john morris

    john morris Everybody's Favorite Quadron

    Location:
    Toronto, Ontario
    I can beat that story! About a year ago I was gently handling a 7th Season Little House And The Prairie DVD when it cracked in half.
     
  15. Grant

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

    All that, and more, is why I just decided it was time to build a new computer.

    Motherboard $159
    Processor $189
    RAM $74
    NvMe 500 GB M.2 $74

    I already had the case, DVD burner, and power supply, so I got away cheap.
     
    Stefan likes this.
  16. harby

    harby Forum Resident

    Location:
    Portland, OR, USA
    Remember, when converted to Mid-Side, the side channel is what creates the stereo width. If you remove audio from only the side channel, you collapse the stereo width by what was removed, blending channels.

    For example, convert to mid-side, then on the side channel, cut everything below 100Hz. You've then not cut out extraneous hum or noise, especially noise that is common to both channels, but you've instead made everything below 100Hz mono.

    This is not a bad thing if it reflects solely what the vinyl mastering engineer also did in summing bass to mono - the only thing that you'll be removing is non-program material.

    Here, though, you are allowing the plugin do more to the side channel. Remember that there isn't magical isolation in mid-side: if you have a completely L signal, then L+R (mid) = L ... but also L-R (side) = L. If you change the side in any way, then when re-combining, then R = (M - S) / 2 will give you a non-zero answer.

    -

    Also it is important to remember that if you apply a filter or plugin to only the side channel, that filter may have a phase delay, or worse, a group delay of several samples. If your plugin delays one channel, they can't recombine.

    For example, see a 17Hz 6th order Chebychev high-pass as just mentioned. The frequency response is flat down to the 17Hz, but the phase response is not - it has effects to 220Hz+ (this is also why you don't evaluate performance on just a null test, as one individual persists). The phase shift will cause incomplete cancellation when summing back to L/R.

    [​IMG]


    Don't know why you'd do such strange settings.
    M = L 50% + R 50%; S = L 50% - R 50%.
    Then to go back,
    L = L 100% + R 100%; R = L 100% - R 100%.

    Example, for signals of 3L and 7R:
    Mid/Side: L 50% + R 50% = 1.5 + 3.5 = 5; L 100% - R 100% = 1.5 - 3.5 = -2
    Reversing: L = 100% + 100% = 5 + (-2) = 3; R = 100% - 100% = 5 - (-2) = 7
     
  17. BrilliantBob

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

    Location:
    Romania
    My NVMe SSD driver is the system disk on C:. Amazing speed of the operating system and DAW. Very goog for gaming, hard processing and net speed. Keep out to the temperature. Mine stays in idle at 46 Celsius degrees in the normal range. I used 11 TB in 6 months out of the guaranteed lifespan of about 300 TB.
    [​IMG]
     
  18. BrilliantBob

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

    Location:
    Romania
    I think the best approach to eliminating rumble is to use a linear phase EQ. Now I play around and testing with the RX6 De-hum module (there is a linear phase EQ here for high pass filtering at 20Hz 2Q in MID and 30Hz 2Q in SIDE) and with the ReaFir subtract mode for infrasound noise reduction. The SplineEQ is interesting too.
     
    harby likes this.
  19. BrilliantBob

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

    Location:
    Romania
    About the recording chain induced noise and why I initially remove noise on my needledrops starting from -75 dBFS...

    With the TT -> preamp -> digital interface -> PC chain, all connected and on, and proper grounded (a single ground point from the PC chassis connected to the preamp chassis and further to the TT chassis) I obtained a signal with no hum or ground loop. But the audio chain noise is still here. -85 dBFS total RMS with peaks up to -75 dBFS. The crosstalk is going up to -85 dBFS too.

    Bearing in mind that the highest vinyl quality in existence can output only a maximum Dynamic Range 70 dB of useful information, the logical conclusion is that below -85 dBFS there is only noise which must be eliminated for a clean and sound needledrop. But maybe is just me.
    [​IMG]
     
  20. Grant

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

    I'm not quite sure how low the dynamic range for vinyl goes, perhaps a bit better than -70db, but I always want that noise gone. The cleaner the background, the better the transfer, and the less work you have to do with the NR. My current Realtek ALC1220-VB codec line-in does 110db. Well, it does DSD too.
     
    Last edited: Apr 8, 2021
    Stefan and BrilliantBob like this.
  21. Grant

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

    Here's the last one I ran a couple of months ago:
    [​IMG]
     
    Stefan and BrilliantBob like this.
  22. Stefan

    Stefan Senior Member

    Location:
    Montreal, Canada
    Which is exactly what we want to do if what's being removed is noise! I've been denoising using M/S now for about a decade over hundreds if not thousands of albums, and I've never seen a case as in the example ghostrider posted where bass that was hard panned to the side beforehand ended up centered afterwards. In fact, when I did my process on the same album section (end of Pink Floyd's The Wall side 2), the bass remained in the left channel after I was done denoising.

    I don't think any of us are cutting everything below 100Hz (Below 20 Hz sure because that's infrasonic and we can't hear it anyway). We're running denoising software on the side channel only because that's where a lot of vinyl noise occurs. Remember that the signal from vinyl originates in M/S anyway with the Mid channel coming from the horizontal movement of the stylus and the Side channel from the vertical movement.

    Not necessarily so. Summing to mono in pre-digital days was done with something called an elliptical filter built into the mastering console. It was of course analog and not brickwall but either 6 or 12 dB per octave, so that means there's still some program material left below the target frequency--100Hz in your example (It also means there's a reduction of side channel level above 100Hz, which results in a perceptually narrower width on vinyl compared to digital. but that's another story).

    Of course! We want to remove the excess noise in the Side channel. When it's recombined, we want a non-zero answer; we want less noise! When I switch to M/S and check RMS levels in gaps between songs before denoising, there's typically 6 to 10 dB greater RMS in the Side channel. If I apply even a 3-6dB denoise to that, when I switch back to L/R, it renders the noise much less audible. Sometimes that's enough. Other times, I do a second pass of light denoising. However, since denoising can never be perfect, there's always the chance of removing program material such as reverb tails, note fades, etc., so the initial denoise in the lower frequencies of the Side channel means that any subsequent denoise I do back in L/R mode is much gentler and therefore it removes much less program material. As I've said several times here, it's always a good idea to listen to what's being removed by previewing with noise output only. I also almost never process the entire file but only areas in the Side channel with less signal, which means more noise will get through.

    As for the merits of Mid/Side, obviously it can't be that destructive or most if not all major digital EQ designers wouldn't offer the option to EQ in M/S.

    Of course! That was exactly my point and I did mention that the presence of significant program material above 17Hz in the null test was most likely due to phase. I'm obviously the "one individual" to whom you referred who persists in using null tests (feel free to use my name in the future ;)), but I'm not the only one who does this (check the mastering forum at Gearspace; you'll see it's used by lots of professionals). By definition, a null test if properly done will show only differences. You seem to be implying that a null test will not show a phase difference, but of course it will and it did when I tested the Chebychev filter!

    I've already explained this several times. It's for convenience of toggling back and forth quickly within RX and without having to use a plugin. I used the same formula in Audition 3 before moving to RX (in fact, it may have been a preset in Audition 3. I don't recall as I no longer have that version installed). In any case, the results are the same and I have one simple preset assigned to a keyboard shortcut so I can quickly and easily toggle back and forth.

    As always, we are all free to do whatever we think is best. I've found ways that work for me and in the true spirit of this thread (when we don't get sidetracked or threadcrapped), I share what works for me with our little community here. When I see something that could cause a problem, I also point it out. It's not to belittle yours or anyone's else's contribution, rather it's too help us move forward and achieve better results in our collective hobby.
     
    ghost rider and BrilliantBob like this.
  23. harby

    harby Forum Resident

    Location:
    Portland, OR, USA
    Vinyl dynamic range is more like 55dB or less when measured unweighted, as we'd measure digital dynamic range. Think that's the figure I saw for DIN rumble on a VMS-70, given in the most "unweighted" of the rumble weighting specs (specs which don't use perceptual "A weighting"). Since 315Hz-centered rumble is measured against a 315Hz 5.42cm/s "0 dB" reference, recorded peaks can be above this, though.

    Dynamic range is the difference between noise floor and the maximum signal (possible).

    However the "maximum" is really the maximum on a particular LP - while velocity signal might go to +10dB or more on a variable-depth lathe's 12" maxi-single, volume can be much lower on a long 60s compilation.

    Also, the noise floor figure can be spoiled by a single loud pop, 0dB signal vs a 0dB click = 0dB dynamic range. We measure noise at it's worst, not its best.


    You can test for yourself unweighted - edit together all the quiet gaps between tracks and find the dB peak, and compare to the dB peak of the whole recording. Results will be worse than you'd hope.


    Don't know what you mean by "cleaner background" here - the electrical noise floor of completely-hum-free Faraday cage battery-operated MM preamp reproduction can't be better than about -81dB, which is orders better than what's on the vinyl. Digitization SNR is hardly a concern if you eliminate its electrical interference.

    Signature-based noise reduction is a lot better than presets - vinyl reproduction definitely has a noise signature. Compile together and profile all the groove silences from an LP, and keep the detection threshold and the dB reduction to the minimum to make that noise acceptable, avoiding the underwater chime-y noise of NR gone too far. Apply to music. Consider: Dolby B is at best a 10dB improvement, and still good.
     
    Grant likes this.
  24. ghost rider

    ghost rider Forum Resident

    Location:
    Bentonville AR
    Thanks guys. Something in my workflow may have caused the conditions for this seemingly bizarre error. I'm going to revisit this in a week or two and see if I have different results.
     
  25. harby

    harby Forum Resident

    Location:
    Portland, OR, USA
    If your gold standard is the null test - it doesn't pass the null test. If your software doesn't have a precisely 1/SQRT(2) volume control setting, what comes out of the two conversions isn't what went in. You wouldn't preview the M/S, so its level shouldn't be important. Channel mixer presets as I describe can also be saved as favorites.
     

Share This Page

molar-endocrine