Your Vinyl Transfer Workflow (sharing best needledrop practices)*

I can read a manual, unfortunately they don't write how to adjust the recording level to an average number beforehand. That's what I was wondering all along, how can one set a level referring to a average level before playing a record? With a regular rest tone OK, but music is not regular, it's dynamic. I guess I'll keep doing whatever I'm doing. I like the results.

 

This has all been mind numbing. Personally I'm not recording anything for awhile anyhow but I'm just going to stick to what has been working for years. Listen to multiple tracks on every record and try to keep the peaks below -5

I use Loudness trying to keep things at -18 RMS (dbfs) I stay aware that it is possible to chop the peaks using Loudness. Normalize still may be the simplest way to raise the levels. I was hoping to figure out a setting in Loudness that works well for most recordings. It's not unusual at all for a 2 record set to have significantly different results from the same setting. I still can't wrap my brain around that.

So I set the peak level in Loudness at -0.3 but if I have more than a few peaks hitting that I'll undo and run Loudness again lowering the RMS dbfs to
-18.5 or -19
usually the peaks are below -1

 

Have a look at the actual distortion of the $180 Motu - and good ADCs that are 10x that.



The Motu is the only one with any kind of fault like you describe - the others reach their lowest distortion at the highest signal level. Even with the Motu dialed in so your recording hits peaks at almost 0dB(fs), the bulk of the samples of a peak waveform and the bulk of the waveform peaks will still be below this rise.

(shift the whole Motu M2 graph to the right to compare to others - at 0 trim, it has signal boost compatible with consumer voltage levels)

It takes a special kind of ADC, designed to not work right, to have rising distortion and still cost a lot of money, like the B2 Bomber ADC, $2600.

• -100dB THD+N (+4dBU = -18dBFS)
• -94dB THD+N (+21dBU = -1dBFs)

"BURL AUDIO has solved this problem by designing an analog audio path that is complimentary to your Mic Pre and to the analog to digital process. By using a revolutionary hybrid circuit with a proprietary transformer, the BURL AUDIO BX1, and a discrete class-A, zero feedback, zero capacitor signal path, we have achieved dynamic and tonal balance."​

 

Exactly. If it ain't broke, don't fix it!

 

When I record needledrops through preamp -> MOTU M2 (volume knobs zero) -> Audacity (30/100 recording volume level) I noticed the recorded programme material has total RMS -22 to -24 dBFS about.

According to these results, I increase the volume of the MOTU buttons very little (I keep the full-color, high-resolution LCD (160x120 pixels) full-length meters in the green zone always), until the recorded programme material reaches the -18 dBFS sweet spot.

This is not theory it's practice and the processed needledrop sounds good to me. No distortion at high volume, no clipping.

 

What I think Harby said is that it's a good idea to turn up the MOTU analog volume before converting the signal by its ADC and keep the Audacity digital recording volume to a minimum (eg. 5-10/100). So that the recorded programme material to reach the -18 dBFS total RMS. I will test this variant, too.

 

As long as you don't clip the signal going in don't worry. Forget the MUTU meters. Look at the meters on your PC screen. You should be able to switch to RMS level.

Worse scenario. As long as you don't peak over - 6 dbfs on your converter meters it should be o.k.

 

Agreed.

 

I never said fault sir. The sweet spot is where the converter sounds it best which is - 18 dbfs RMS. In fact with cheap converters this is even more important.

I never said "rising distortion" above - 18 dbfs RMS. The sweet spot is not something you can measure. Usually.

Sweet spot: best stereo imaging, most detail. Best sound. It is not a fault.


Yes, of course you can record higher but your best sound will be at - 18 dbfs RMS. Never said it will sound bad any level higher. You misunderstood.

This is a known fact that every professional mastering /mix engineer knows. It is taught in audio production schools. No matter how good a converter will be it will always have a sweet spot. This is why digital equipment was models after analog. In 24 bit 0 VU = - 18 DBFS RMS.
Regardless of good the design is this will still apply.

Even our $12 000 Burr-Brown converters at the studio have a sweet spot. They are an 8 channel interface. Most ADCs are 20 bit. Our studio's converters operate at 22 bit performance so their sweet spot is wider.
We can hit anywhere between - 12 and - 24 dbfs RMS. But even that range will have a sweet spot.

 

It is possible to design a converter that is easier for amateurs to use. The MUTU may be a converter that was designed to be maxed out. Easier to use but a bad idea for setting headroom.

There is another reason we prefer - 18 dbfs RMS - mixing. There is a certain range where faders are at the optimum. - 10 to +4. At this range every fader move is equal to 2 db or so.
If you have a signal that is too high then you have to turn the fader very low. As in - 20 db. At this level it is next to impossible to mix properly. A move can be 4 db or as much as 6 db.
We can trim the level during the pre-mixong stage so that every track is - 18 dbfs RMS but it takes up time. And in a studio, time is money.

 

Sorry but this is just bad advice! Why ignore the MOTU meters? They were designed by the people who designed the converter. I think they knew a little bit about what they're doing. Instead of using the professionally designed meters on the MOTU (which are probably part of the analog circuitry), you're telling someone to rely on the digital meters on the PC screen but not to read a digital value (for which they are designed) but a value that's supposedly equivalent to an analog value. You're mixing two different technologies. What's more you're sticking with the notion of not having anything peaking above -6 dBFS while sticking with -18 RMS. I've already explained why this is not practical for most needledrop work without adding compression but you keep ignoring that fact. I'll repeat my earlier question: you admitted you're not a vinyl guy, so why are you spending time in a thread dedicated to a hobby you know nothing about and giving misleading advice?

 

Not wanting to give up without trying, I just made a quick recording in Audacity with the meters showing RMS. Going as high as possible into the green, I recorded Anna from the Beatles Please Please Me in mono. When I analyse the contrast, Audacity says -18.87dB. I see peaks around -3db.
Now when I open the same recording in RX and run Waveform statistics, is says Total RMS level -15.90 dB with peaks at -3.59dB.
The recording sounds good, no clipping, but it sure goes 3dB above the -6dB limit.
Now: what do we do with this??

 

If it sounds good to you, leave it.

 

We set the level of a 24 bit converter at - 18 dbfs RMS for 3 reasons.

1. Gain staging. Very important for mixing.
Although this will not apply to most of you

2. It is the sweet spot of most pro/consumer level converters. This isn't a fault.
Take an integrated amplifier. It will not sound the same at all levels. Now imagine a amp that only sounded good at full volume. Silly, I agree.
Regardless of price every integrated amp will have a sweet spot - a level where the amp shines it's brightest. That is not to say that the amp sounds bad at other levels.

Keep in mind many interfaces will have mike preamps as well. As long as you stay within the "line" range you are o.k. However, if you need to turn the input level past line into the mic range then a whole new set of rules will apply.

THD peak is not the same as THD RMS.
Hitting - 3 dbfs peak is not the same as hitting - 3 dbfs RMS. The distortion measurements will read very differently. Hitting the occasional quick - 2 dbfs peak should not matter at all. But there will be a rising distortion if you hit - 2 dbfs RMS.


3. HEADROOM
If you set your level to - 18 dbfs RMS it is 0VU for 24 bit converters. You will hit nice peaks - high but not to high. And you will never run out of headroom. Remember in an emergency you have another 6 db to go. Trying to set a level by peaks is simply well amateur and a lot more trouble. Setting level by LUFS will be easier.

This is not my opinion. I want to make this clear..Keep in my there are exceptions to the rule. Some converters (foolishly) have been designed to sound better as you record higher. Bad idea for setting headroom. The MUTU is possibly one of these converters.

I have posted this video many times. Another time doesn't hurt.





"..... Every spec sheet is different. Some call it headroom, some just list a maximum input level. Sometimes it's listed in straight db relative to the nominal level, sometimes it's listed as dbu, or dbV, and that's not always even the same scale that they used to specify nominal level. Some even just list it as a voltage, so that you have to go find a volts>db calculator.

Does it sound better? That's kind of subjective isn't it? It completely depends on the behavior of the analog chain. Some analog stages might sound kinda good when pushed up into distortion, some just sound like crap, and some folks might like the overdrive from this one or the other one, and some folks might thing it sounds better if it comes in as clean as possible. I'm not talking about "clipping" here. As mentioned above, it's quite possible to distort your analog chain without ever coming close to 0dbfs and incurring "digital clipping", and a lot of analog gear will start to get non-linear well before it even hits its own maximum output level....."

Ashcat_it
Cockus Forum


Some more digital wisdom:
This is a particularly sneaky way to exceed headroom, because it can create problems and you may not even be sure where they’re coming from.

This issue occurs because most digital metering visualizes the actual numerical value, so the numerical value that represents 0dBFS shows 0 on the meter. However, converting digital audio back to analog may result in higher actual values than the samples themselves, which creates the potential for intersample distortion.

This type of distortion can occur if the level of some samples uses up the maximum available headroom, and these high-level samples then pass through the D/A converter’s output smoothing filter to reconstruct the original waveform. This reconstructed waveform might have a higher amplitude than the peak level of the samples, which means the waveform now exceeds a playback system’s maximum available headroom (fig. 2). Unless your channel’s meters can alert you to intersample distortion, you should leave a few dB of headroom to avoid this. Many mastering engineers recommend absolute peak levels not exceeding -1.0dBFS.




Figure 2: With the analog audio waveform sampled in (A), raising the digital audio’s level to the maximum available headroom (B) can exceed the maximum headroom when going through the smoothing filter (C) that reconstructs the analog waveform. (Image courtesy Hal Leonard,Musician’s Guide to Home Recording.)

If you create a digital file (or a CD, for that matter) for playback with the potential for intersample distortion “baked in,” then the listener’s D/A converter will create this distortion. But another problem is that when you’re mixing, if you’re not aware of the possibility of intersample distortion, your monitoring system may be producing distortion, and your mix will be based on a distorted sound.

- Sweetwater

 

You play a the record and look at the meters on your PC screen. Forget the MUTU meters. As long as you have the level at LINE input and are not going over don't worry.

 

Sounds good to me. Just think of - 18 dbfs RMS as 0VU in analog. Perfect level and if you need more headroom it is there. And so much easier then to spend all day searching for peaks.

Most modern converters today will sound pretty good even maxed out - 1 dbfs peak. However they will still have a sweet spot. Older converters used to sound really crappy long before 0 dbfs.

Now you normalize to 0.1 dbfs peak. Good luck.

 

That's MOOOTU and once more you tell people to ignore their meter. Do you think you know better than the engineers at MOTU?

 

Well, thanks for this post. It better explains things, and can be useful for those of us who do these needledrops. It makes sense, too. Everything has a sweet spot, even your car's gears while you're driving! The exchanges have been contentious, but if things are explained, it helps.

Now, whether this will actually apply to the folks on here that want info on best practices will matter to them, it's just nice to be there. Actually, I read the same things years ago, and they back up what you have been saying for a while.

Most of us aren't doing any mixing (I have, and I think it's a bitch!), but some of us definitely do processing. My way around the gain staging is to use 32-bit floating point, but not everyone does, or can.

Now, I don't own or use a MOTU interface, so I cannot comment on it.

 

I'm fascinated by his accent. What is that?

 

No because normalizing to 0.1dBFS risks intersample distortion, which you just posted about.

 

LOL. Lower West Side New York
Lower middle class accent.

He is an old friend of mine.

 

32 floating bit solves some problems. But since I refuse to mix ITB I can't use it. As good as a SSL 4064 board is, it will overload as some point. Unfortunately.

 

If you find RMS annoying you can always use LUFS. Works better.

 

A comment on a post on a post.
Did I really post all these?! Sorry. I am sure there was a way to say this in 30 words but .....

 

This illustrates why I've been so critical about this notion that we should all be recording our needledrops at -18 dBFS RMS and have no peaks over -6 fBFS. People who are doing fine needledrops suddenly think they're doing something wrong! The reality is that albums are cut at different levels, and most of them (other than modern ones cut from brickwalled masters) have a crest factor that's much higher than 12dB.

For example, here are the stats for a needledrop I did this morning of Joni Mitchell's Hejira LP. I actually recorded it at a lower level but did a quick boost to -18dB to show that the peaks are definitely above -6dB. Sure I could apply limiting or compression to reduce the peaks to -6dB but the recording still had peaks reaching 17dB above -18 dBFS. It would have been impossible for me to record this needledrop at -18dBFS RMS AND keep the peaks below -6dB.


Here's another needledrop I did this afternoon (I'm working at home these days so I can record needledrops while I work ). This is an early UK pressing of Elton John's tumbleweed Connection. I didn't boost this at all. My current recording level is set with a test record tone at -18dB but then I've reduced it a further 3dB. Thus, the Elton album at -25.45 / -25.92 is below the average level. However look at the peaks. If I were to boost this one to -18dB, the left channel would have a true peak exceeding 0dBFS by 0.51 and a digital over of 0.31. So much for keeping those peaks at under -6 dBFS and always being able to rely on that headroom being there!

So, as I wrote above, if you did a needledrop that sounds good to you, really that's what this hobby is all about. You're doing fine Stan!