Jazz Version of Blackbird: Vocal Recording Quality Sounds Okay?
I took a couple days to sidestep my main project for a reality check and created a project to demonstrate the results I get when tracking vocals with an SM58 and then applying post-processing in REAPER.
In the mix, the vocal may be a bit too loud, but I did this so that if anyone wants to listen and comment you can pay attention to a few things I've been concerned about:
1. Can you hear any frequencies that are coloring the vocal in an odd way?
2. Can you hear an unacceptable level of hiss or sibilance?
3. Does the vocal sound too "thin" to you for a male voice?
4. Does the vocal sound natural or does it sound under or over-compressed?
5. And ultimately -- If you had this recording quality for a vocal handed to you in a mix would you use it for a final master or redo it?
This is a jazz version of the popular Beatles song I worked out a year ago. There's no limiting at the master. I only turned it up.
I'm not going to reveal what vocal processing I'm using yet. I wanted to post a version of only the vocal, with and without processing, so that someone can tell me if it sounds like the raw recording may indicate the mic may be broken or not the sound of a genuine SM58 (I just learned there are counterfits out there). But I'd like people to first hear this in the context of the mix because that's of course how it would be played.
If anyone is interested, to record the vocal I used an old M-Audio Omni interface along with their Delta 44 PCI soundcard. The Omni doesn't have the best mic preamps (specs are available here: https://www.zzounds.com/item--MDOOMNISTU ), but I like it because it's got a warm sound (probably compliments the SM58). There's no EQ or compression used going in on the mic while tracking.
I "grew up" in digital with an OmniIO and two Delta-66s. I wouldn't call the Omni warm per se, it's just a front end IO and preamp for those cards and generally transparent - it was a nice piece of gear though.
I don't hear huge problems with the vocal. I want to think it could use some more compression; I think that because the vocal sort of jumps in and out and you can't hear as much of the emotional aspects of the singing etc. but could be wrong. Maybe could use a smidge of de-ssing or lose a tad of wispiness (probably gentle high shelf cut) but it isn't horrible and the vocal quality is fine in general other than that and sounds plenty clear.
My ear does want to hear you hold the end of some of the phrases out longer. That will make the vocal less choppy, but that's just a personal observation unrelated to the main ask. That could also be related to my compression statement though.
__________________ Music is what feelings sound like.
1. Can you hear any frequencies that are coloring the vocal in an odd way?
2. Can you hear an unacceptable level of hiss or sibilance?
3. Does the vocal sound too "thin" to you for a male voice?
4. Does the vocal sound natural or does it sound under or over-compressed?
1. There's a touch of peakiness? at ca.1750Hz?
2. Nope
3. It's "light" for the "genre", often a very close, warm (intimate) sound. But that's a matter of taste - it doesn't sound silly. You could dip out (broadly) some at ~5k & warm it up a bit at 200. But I guess that's the 58 for you.
4. The vocal sticks out in a few spots, I dunno if that would necc. call for more compression, but it's definitely worth thinking about.
__________________ it's meant to sound like that...
I want to think it could use some more compression; I think that because the vocal sort of jumps in and out and you can't hear as much of the emotional aspects of the singing etc. but could be wrong.
Quote:
Originally Posted by jrk
1. There's a touch of peakiness? at ca.1750Hz?
4. The vocal sticks out in a few spots, I dunno if that would necc. call for more compression, but it's definitely worth thinking about.
My ear does want to hear you hold the end of some of the phrases out longer. That will make the vocal less choppy, but that's just a personal observation unrelated to the main ask. That could also be related to my compression statement though.
I think Karbo is pointing toward the main thing that I would think about. I know that you are mainly askng about mic choice, EQ, and compression, but these are are secondary considerations to performance. I think that I could appreciate a jazz version of Blackbird, but I am distracted by the the choppy phrasing of the vocal in this version. I am sure that this is highly influenced by my over familiarity with the original version, but a more more drawn out legato phrasing would seem more natural to me and capture some of the feel that makes the song so special. Of course this is a matter of taste, and your opinion is just as valid as mine if you are happy with your stylistic approach. Your voice quality and pitch are quite good, and I am sure that you can achieve whatever you want with the vocal track. It seems to me that those of us who record our own music get so focused on the technology of recording that we lose focus on more important things. I suspect that most people will prefer a recording that has feeling but has mediocre audio quality to a recording with less feel that is sonically perfect. But again, what do I know?
I think Karbo is pointing toward the main thing that I would think about. I know that you are mainly askng about mic choice, EQ, and compression, but these are are secondary considerations to performance. I think that I could appreciate a jazz version of Blackbird, but I am distracted by the the choppy phrasing of the vocal in this version. I am sure that this is highly influenced by my over familiarity with the original version, but a more more drawn out legato phrasing would seem more natural to me and capture some of the feel that makes the song so special. Of course this is a matter of taste, and your opinion is just as valid as mine if you are happy with your stylistic approach. Your voice quality and pitch are quite good, and I am sure that you can achieve whatever you want with the vocal track. It seems to me that those of us who record our own music get so focused on the technology of recording that we lose focus on more important things. I suspect that most people will prefer a recording that has feeling but has mediocre audio quality to a recording with less feel that is sonically perfect. But again, what do I know?
T
I appreciate the input on the style of the vocal, but I'm not posting this for the sake of the vocal performance. If you listen to bossa nova, you'll hear there's a staccato way of singing that suits the music. It accents the rhythm to not sing in a more legato manner. But this can be argued all day based on preferences and taste. I'm trying to do something different unlike the original song. That's jazz.
I'm asking for those who are knowledgeable about recording vocals to judge the recording based on the five questions I had above in my first post.
Thanks for the comments.
Last edited by Peterk312; 02-11-2021 at 10:53 PM.
Reason: More thoughts...
I suspect that most people will prefer a recording that has feeling but has mediocre audio quality to a recording with less feel that is sonically perfect.
T
I sure hope so!
But I also know a few audiophiles out there who would dismiss a great performance because it wasn't recorded well enough.
And I do want to mention that one of reasons I think I was light on the compression is because I'm trying to make the vocal sound as natural as I can, as if I was singing in the room you play the music in. If I was actually there singing live, you would hear the vocal jump in and out at times because nobody can control their voice to the extent that the loudness of the vocal is as even as you would get when using a compressor. I think it fluctuated a little too much after the first two comments agreed, and that's why I created a new version.
Really, I'm more concerned about squashing the vocal than having the level get a little louder on certain phrases, which I actually do sometimes for effect.
I appreciate the input on the style of the vocal, but I'm not posting this for the sake of the vocal performance. If you listen to bossa nova, you'll hear there's a staccato way of singing that suits the music. It accents the rhythm to not sing in a more legato manner. But this can be argued all day based on preferences and taste.
I'm asking for those who are knowledgeable about recording vocals to judge the recording based on the five questions I had above in my first post.
Thanks for the comments.
Fair enough. But I will point out that Astrud Gilberto's delivery in the most famous version of the most recorded bossa nova song of all time (Girl from Ipanema) is not staccato. But you are right, you asked a set of questions that I didn't address. Carry on.
Can you explain what you did there? It sounds like there was midrange removed from the middle and more high frequency added.
Is that what you don't like about the SM58 character? Like 1.5 kHz to 3 kHz?
Also, are you listening through monitor speakers or headphones?
I'm listening through a pair of old (refurb) B&W DM5 speakers w/ a subwoofer
I do like the 58, it's a great "rock" mic. Works really well for pushing the voice over noisy guitars etc. The broad presence peak (+6dB?)from (what?) 3k ->10ksounds far too crispy for this sort of thing IMO. Also, classically, the talent might work the mic quite close - I guess you haven't - and together with the low roll-off - this robs quite a bit of bottom end, thinning the voice a bit.
So - a broad dip at ~ 5k and a boost at ~200. Was the suggestion.
Actually, I used a high shelf -2db @ 360, a broad-ish (q=1) -1.5dB band at ~6k and a high shelf +2dB @ 9k5 cancelling the lower shelf. But that's on a mix, and trying not to screw the other stuff up too much. You'd probably do it differently
But this is all a matter of (your) taste, and picking good reference material. Only you know what you want.
Enough about vocals....
__________________ it's meant to sound like that...
I do like the 58...sounds far too crispy for this sort of thing IMO...the talent might work the mic quite close - I guess you haven't - and together with the low roll-off - this robs quite a bit of bottom end, thinning the voice a bit.
So - a broad dip at ~ 5k and a boost at ~200. Was the suggestion.
No, I did work the mic quite close, about 2-3 inches away. And I like to take advantage of the proximity effect. But my first EQ has a very steep cut in low frequencies, maybe too much. It's always stumped me how much low end to remove from a vocal. I often read people using a high pass from about 100 Hz. I left enough low end to see somewhat of a peak in the 100 - 200 Hz. range, but by 50 Hz it's down by more than -6db according to the ReaEQ spectrum.
What I'm trying to avoid is an unnaturally audible bass tone following the whole vocal, and this of course helps other instruments that need that low frequency tone to sit in the mix (e.g., a standup bass), particularly given the vocal is the loudest element in the mix.
Quote:
Originally Posted by jrk
Actually, I used a high shelf -2db @ 360, a broad-ish (q=1) -1.5dB band at ~6k and a high shelf +2dB @ 9k5 cancelling the lower shelf. But that's on a mix, and trying not to screw the other stuff up too much. You'd probably do it differently
What I heard sounded like a reduction of midrange, something like the typical "smile" shape some people go to with EQs almost by habit cutting out midrange. Are you saying you would have done this only to the vocal and not the whole mix? I'm not trying to get into how the project should be mixed. Just trying to gather opinions on the five questions I asked above and how I might go about fixing these issues if present.
Quote:
Originally Posted by jrk
But this is all a matter of (your) taste, and picking good reference material. Only you know what you want.
I think it's obvious I don't have enough confidence in my recording skills and gear to just go with what I hear in the recording, which is why I'm asking what others think about the recording quality.
So far, I'm getting the impression it's not as bad as I thought it was.
I think it's obvious I don't have enough confidence in my recording skills and gear to just go with what I hear in the recording, which is why I'm asking what others think about the recording quality.
There's a difference between something having problems that make it unacceptable and being a matter of taste. When there are no glaring issues (like in your example), we're mostly down to what you personally like best.
At some point, it really is OK for you to decide what it's going to sound like, it's your creation after all.
__________________ Music is what feelings sound like.
There's a difference between something having problems that make it unacceptable and being a matter of taste. When there are no glaring issues (like in your example), we're mostly down to what you personally like best.
I actually did think there was a glaring issue, mostly in hiss and sibilance, and I thought someone would hear an odd frequency stick out. But so far, most important thing I think I overlooked is the compressor was set too lightly to not have some phrases and words jump out suddenly.
You can probably gather from my other posts (for example: https://forum.cockos.com/showthread....20#post2393820 ) using an SM58 is not helping me feel too confident about the recording quality either, although that thread (along with some videos available demonstrating use of a SM57 or SM58 for vocals) made it clear that you can get good results with an SM58. But even here, I get the impression I'm using the wrong mic for the style of the song, and that means I need to somehow compensate with processing. There are an infinite number of ways ways to do this, some of which would probably destroy the recording quality I managed to get for the vocal, and that's why another set of experienced ears along with knowledge of recording technique are of value to me.
For a few reasons, I'd like to get into some of the processing I used on this vocal track.
The image above is the first EQ used. This is what generally shapes the sound. When bypassed it makes a huge difference.
What I was wondering is does anyone think that cut in the low frequencies is odd or too extreme? I realize I did not yet post the solo vocal (and I might not), but with a genuine SM58, would you expect to need that kind of cut with a male voice? If you can listen to the sample recording again, does it sound like if I used an SM58 this kind of EQ cut would typically be necessary?
What I was wondering is does anyone think that cut in the low frequencies is odd or too extreme?
Yes, both odd and extreme.
Usually when I mix (or "post-process") vocals, there's almost nothing going on in the EQ. Low cut filter sure, but other than that, max. 1dB here or there.
Strangely, the vocals in the mix don't sound tweaked that badly. They're on the thin side and they're sibilant in some spots, but I wouldn't have guessed that you practically attenuated all of the low end!?
EDIT: Even when I (used to) mix FOH, I most likely cut male vocals around 120 to 150, the rest of the tweaking was usually due to avoiding feedback in the room.
__________________ Windows 10x64 | AMD Ryzen 3700X | ATI FirePro 2100 | Marian Seraph AD2, 4.3.8 | Yamaha Steinberg MR816x "If I can hear well, then everything I do is right" (Allen Sides)
Strangely, the vocals in the mix don't sound tweaked that badly. They're on the thin side and they're sibilant in some spots, but I wouldn't have guessed that you practically attenuated all of the low end!?
I'm concerned I may have a counterfeit SM58, a knockoff.
I can't figure out how to embed the video here. Embed code is not working.
@ 2:04 in the video he's speaking through a real SM58. Watch the bass increase when he switches to speaking through the fake SM58.
That could explain why I need such a big cut to get it to sound the way it does. Surprising, isn't it?
Sm 58s are ubiquitous. Find a buddy who has one and have him do an AB comparison. You'll know if something is not right about the mic in 5 minutes. You shouldn't have to record anything as it should be obvious even through a PA system if the mic is as wonky as the eq curve would indicate.
It is good news that Sure will do that. And if it is a faulty Sure mic, there is a good chance that they will fix it at a reasonable cost. They are a good company. It is really a good idea to check on repair/return policies and the like before you buy a mic. My son was given a nonfunctioning small diaphragm Neumann condenser mic (can't remember which one ATM) that would have cost a a couple of thousand dollars if functional (as partial payment for studio time). Turned out that Neumann was willing to replace the microphone cartridge for a couple of hundred dollars including labor and shipping. It is now one of his go-to mics. How can you not love a company like that? Their products might cost more, but in the long run, they are worth it.
Problems with the recording quality of the vocal at this point:
1. Sounds too thin and sibilant at some points.
2. Bass cut is too severe, possibly because I have a damaged or fake SM58, or possibly because I have the wrong idea about how much lower frequencies should be cut to get the vocal to stand out in the mix (and used the EQ above with a severe low frequency cut).
3. Overall, vocal may sound too noticeably like an SM58, but this may be due to post-processing.
What I did today was try a different EQ, a different way of handling control of the lower frequencies, and I used a different compressor not used in the last version above.
I think I managed to get rid of the "thin" sound that appears now and then. The lower frequencies have been added back in, but I overlooked that the thin sound was happening not only because I cut the low frequencies too much. The low frequencies actually jump in and out, most likely because of the proximity effect of an SM58 (or whatever knockoff I may have). So, I added a dynamic EQ (Nova by Tokyo Dawn Records)) to control both the low frequencies as well as sibilance. I also removed ReaComp from the processing chain and switched to VOLA vocal limiter.
If you listen to this with headphones (e.g., Sony MDR-7506) you may not notice much difference. But if you A and B the two with good monitor speakers you'll hear the one with more bass and different compressor that I did today sounds very different.
Let me know did I get closer to solving the 3 issues I noted above.
The excess sibilance is your original recording was minimal to begin with, but the new version is an improvement. Indeed I thought that the quality of your original vocal was good. If the vocal sounded a bit thin in general, I marked that up to a stylistic preference, and I thought that it sat well in the mix. It is common to shape the EQ for one part to fit the mix, so it can be a bit misleading to look at an EQ curve for one part out of context. Indeed, f you listen to isolated tracks from George Martin's/ Jeff Emerik et al mixes for the later Beatles recordings, it can be surprising how thin they sound. The only reason that I would suspect any problem with your mic is the extreme EQ curve you used, but as I said, that might be misleading. Your new version is not overcompressed, and nothing strikes me wrong about it. With respect to recording quality, I would say that the track is ready for mastering. I think that you have done a fine job.
I hypothesized that the reason I used that EQ posted above with the big low frequency cut was possibly because I have a broken or fake SM58. It was purchased many years. Today, I got a new SM58 in the mail and did a comparison.
The two mics sound virtually indistinguishable.
I made a test project in which I recorded both mics at the same time. The mics were right next to each other with my mouth centered between them about 3 inches away. I took this screen shot above to show how I compared two EQ curves. The single green line you see is the old SM58, and the pink colored area is the new SM58.
That snapshot shows there's a slight bit more high frequency above 10 kHz with the new mic, and I noticed at times there was momentarily a slight bit lower frequency response. It suggests the moving parts in the new mic capsule/diaphragm are not as stiff as the older mic, yes?
But the difference you see on the graph is not quite audible. It was a little shocking to see the two EQ curves following each other almost perfectly.
So, even though the old SM58 looks different (it has a sticker under the grill that says "SM58" and "Dynamic" etc. and the new one has silkscreened lettering) it sounds indistinguishable. Before doing the EQ I thought the new one sounded just a hint sweeter, but I think it was psychological because it was a new mic. In a mix it will make no difference for a vocal.
Cannot add anything to the comments already posted by others but that last image does suggest you have two identical mics with differences within manufacturing tolerences and slight differences probably owing to distance to mouth which are inevitable at such a close range (minor uncontrollable changes having a big affect).
I would expect slight differences in the mics design and manufacture over the years (only a guess) and that would explain the differences in the printing.
I wonder if your perceived need for such a large amount of low cut can just be put down to proximity and your voice characteristics but to my old ears the results are pleasing.
Only thing I don't understand though is if the mics are so close in how they sound, how come when I do a null test the sound doesn't cancel out completely?
Two tracks, vocal recorded at the same time. Pan one hard right and the other hard left. Switch the phase on one of them. Mute the track that has SPAN on it (which the two tracks are being sent to). Flip the mono switch at the master, and I can still hear some low frequencies and some higher frequencies. It sounds like all the mid-range is gone.
Only thing I don't understand though is if the mics are so close in how they sound, how come when I do a null test the sound doesn't cancel out completely?
Because they aren't 'that' close... and null isn't going to work on analog sources anyway.
__________________ Music is what feelings sound like.
Jazz Version of Blackbird: Vocal Recording Quality Sounds Okay?
I wouldn't call the Omni warm per se, it's just a front end IO and preamp for those cards and generally transparent - it was a nice piece of gear though.
Linear Mode
