Your Vinyl Transfer Workflow (sharing best needledrop practices)*

I used believe this for sure. I think with analog that is the case. With digital all the noise is from the TT and preamp. When you raise the levels the noise goes up too. I record peaks to-3 or less. So I'm sure you are correct if you record very low.

To me its worth it to have plenty of headroom.

 

Do you guys use a Low Cut switch on your amp if you have one? I was afraid this switch would lose the bass if I used it but I noticed all my needledrops had an unpleasant heaviness to them that is hard to describe. Like there was a constant low bass rumble underneath everything. I went back to my instruction manual (last resort) and it said that this switch can eliminate turntable rumble. I tried it out and now the rumble is gone and needledrops sound much better. Am I losing bass with this though, what's the technical details about what this switch does? Do I just have a crappy turntable and others don't have this rumble?

 

It sounds like you have a subsonic filter. I would use it, but my phono stage doesn't have one.

 

From what Barry Diament has said around here over the years I thought headroom in ADC is more than vital then DAC. I always record with levels under -3db - but actually my ADC (Firewire Solo), I think, anyway, records the straight line level signal from my pre-amp. I have some records with relatively high volume that approach -3 and some (long) that peak at sometimes -15. I normalize those, too, for maximum signal before DAC playback; in terms of noise floor, how different is that than increasing the gain on the analog side?

 

This is one of the things on which (with all the due respect) opinions and technical details aren't necessary. Just play some records with the switch on and off and let your ears to give you the answer.

 

Lots and lots of commercial CDs (even audiophile labels like DCC and Audio Fidelity) have peak levels higher than zero dBTP. Presumably those mastering decisions are made by people who know (or at least believe) that hitting 0 dBFS will not be audible.

Normalizing to -0.3 dBFS is relatively harmless, assuming it's done in at least 24 bit resolution. Making gain moves in a 32 bit domain is better.

 

Peak to -0.3. Having your peak at o.o will introduce clipping.

 

Exactly there is no difference. So there is no longer the need to record as hot as possible without clipping because virtually no added noise is a result of the recording. The danger of clipping far out weighs the need for a hot recording.

 

Before I started recording in RX4, sound forge did not have a way to normalize to 0.03 all I could do is lower it a tad. In reality even when I normalize to 0.0 when something does clip its just a peak here and there that clips and I would never know it by listening.


This is something of interest to me. When I compare my drops to someone else’s usually if theirs sound better at first it’s because it seems louder. There is either some other way they normalize and the peaks are getting compressed or my TT has more peaks and this prevents mine from normalizing hotter.

 

I think you are wondering the same thing I was when I created this thread.

 

It's not you. You can get hotter if you use a peak limiter. Voxengo Elephant works great with RX4. Or if the original track is more compressed (less dynamic), normalizing to -0.3 will result in a louder file.

The trick to using a peak limiter is to not abuse it. One of the things that has been eye opening to me is that peak limiting seems to be used on even "audiophile" CD reissues of rock titles (don't ask, I ain't sayin' here of all places), but sparingly. I guess those reissue companies still want their loudness to be "competitive", as Bob Katz likes to say.

Also, dBFS RMS isn't as good a loudness gauge as LUFS.

 

I record with my low cut off (actually, my tape out bypasses it altogether) but I perform at 20 hz cut in Audition after transferring to remove any unnecessary and largely inaudible frequencies.

 

I’m doing basically the same thing in RX4. Not an EQ function. I highlight everything below 25 Hz and silence it. Has been working fantastic.

 

I was just reading through here and noticed I posted back in 2011. Lots changed for me since then.

VPI Scout II
Ortofon Rondo Bronze
Lehmann Black Cube SE II
Apogee Duet
Izotope RX3 (32/44.1)

I record and edit in Izotope. I split tracks in Audition.

As for editing, I do a high pass filter at 20hz. I sum mono tracks to mono. I declick silent portions between songs, and most quiet portions of music. I train the denoiser between each track and denoise fade outs and silence using default settings. If there's a lot of baked in surface noise, like crappy 60s vinyl, I might denoise the whole thing. Since most of my digital listening is at work with headphones, surface noise can be apparent and needs to go. I normalize to -0.5dbs peak per side. From there I dither, then open the file in audition and split the tracks. Save a lossless file as ALAC and a portable AAC file. Tagging is done around the same time.

I only recently started using the denoiser. Anyone have tricks or settings? I've just been using the Izotope default. Anyone use the rumble filter preset, or any other curve?

 

BTW, do you have a needledrops YouTube channel? I really enjoy listening to good needledrops (and I'm sure that yours are great).

 

Thanks it was an interesting read.
I’m not so caught up about maximum levels any more. Back when I was completely self taught, pre-Steve Hoffman forum. I was compressing everything in Ozone 5.0 with some kind of preset. I had even figured out how to set the thresholds in such a way that it acted as a noise gate chopping everything below -40 db. I had no idea how bad I was butchering the audio. I’m not saying you or anyone else is doing that but my point is. I would get drops sounding good (my definition at the time) at home and play the CD in the truck and it sounded horrible. Even some store bought CD that were compressed loud sounded bad. I determined that as has been stated many times some DtoA converters will clip at those levels and my truck was one of them.


Now the only reason I think about raising the levels is match the levels of samples posted in the needle drop forum. When I’m listening and want it louder I turn up my preamp.

 

Not yet. I may give it a try. Thanks for the encouragement.

 

This is what I do.

 

No, it's not messing with the dynamics as long as you don't 1: exceed 0, and 2: you don't use a limiter.

 

That is why you only normalize up to, say -.1 or -.2 db. Some people normalize right up to 0. As Stefan mentioned earlier, some people disagree on the best margin, but it all works out.

 

You can assume it is peak normalization. RMS normalization happens when you use the Hard Limiter.

 

I have been making so many changes to my process. It has taken me awhile to get it all figured out and post it up.
VPI scout LE
AT-OC9ML/II ( azimuth recently dialed in with video analyzer) a very useful tool!
Mc C2500 onboard phono pre
Focusrite Forte usb soundcard in 32bit float / 96

I am now recording and doing all my processing in RX4 in is my belief that there is some improved audio quality that is noticeable. My guess is there a benefit by recording in 32bit float and not exporting the file to use programs like Click repair.

Screenshot 1 shows how I deal with subsonic noise I highlight and expand in such a way that only 25Hz and below are highlighted then I click cut. This will work with any version of RX. When I was using a HP filter in Ozone 5,0 it was having a strange effect on the entire file and you could see the wave form change and a few peaks would jump up. This way only that portion is affected.
Anyone using a high pass filter I suggest you try this.

img
Screen shot 2 shows how I deal with noise. I no longer denoise the entire file. I only do the gaps in-between tracks. You can see at this setting it would harm the music if I did process the entire file I also highlight a 2nd time just the bottom part and overlap it a bit. This does 2 things it gets that residual noise that was left lowering the floor noise in the low -70s or lower and it tapers off the sudden silence so it sounds natural.

online photo sharing
Unless it is a very clean record I declick the entire file with the algorithm @ single band with level set 0.7. If the record is noisy I may raise that to as high as 1.0. I manually declick the rest.
I also use Ozone 5.0 as a plug-in to role off the highs because the AT-OC9ML/II is bright.

imagur

Last I normalize to 0.03 and dither using the RX4 preset for DVDs and export as a 24/96 flac file. I would like any input on the dither options in RX4.

 

Let's hear a sample then

 

Stefan posted a bunch of extremely helpful info better part of 2 years ago on how to use RX2, I suspect he may have some buried in this thread for RX3 and RX4. Regrettably I have not been following this as of resent so I don’t know where those might be. My earlier post may help you.

 

Is there any reason why you choose single band over multiband for declicking?