OverTone FC-70 Review - Fairchild 670, Really?

[Burillo]

i like how people say that aliasing adds up and damages the sound, while completely disregarding gratuitous DA/AD conversions (and with it - clock skew et al) that inevitably happen when you use analog hardware

Very true! AD/DA conversion over and over again does take it's toll on the sound but as far as I know it doesn't add irrelevant harmonic content, or does it?

I'm also lucky enough to have really good AD/DA conversion so it wouldn't be a big issue for me personally but you do have a good point.

Actually, now that I think about it, isn't the main "damage" done by AD/DA conversion happening at the analogue stage, the filtering stage? That's the most critical component as far as I know and it's the one where cheap sound cards usually do quite a bit of damage. The single chip technology has gotten really good in the past few years though so can it really have such a large impact on the sounds? I don't think I've ever done more than 2 or 3 generations of outboard bounces but you do raise a very interesting question. How many generations of out->in-out->in processing can you do without seriously changing the signal? And how do those changes damage the whole mix? Aliasing damages the whole mix by having a lot of stray harmonics which aren't at all related to the song harmonically.. or that's my theory of the damage caused (and it might be completely wrong!).

Cheers!
bManic

if you look up Ethan Winer's "Audio Myths Workshop", he actually talks about that too (spoiler - sound isn't seriously degraded until 15-20 AD/DA cycles even on a cheap SoundBlaster, let alone good audio interfaces). he also talks about "adding up", about various distortion and how you won't ever hear it because it will be masked by other sounds and lots of other enlightening stuff.

my point though really wasn't about AD/DA distortion per se. my point was about people overhearing horror stories about aliasing in the digital domain and concentrating on them while completely overlooking other sources of potential damage that we invented digital audio to get rid of in the first place.

[bmanic]

Yeah, I know Ethan Winer's view on things and his Audio Myths workshop is indeed quite refreshing but unfortunately he is TOO radical in his assumptions about what you can and can not hear. Perhaps from a layman's perspective, some of his examples would actually make sense but for a person used to critical listening many of his examples are almost ridiculous.

Also, that man has a VERY clear agenda and that is "acoustics are everything! They mask and dwarf all other problems ever!".. which is not completely incorrect, just very extreme view of things. Personally I think he is perhaps not very skilled in critical listening and the finer points of mixing/mastering (listen to some of his Cello recordings and see what you think of them).

He has some very good points but some of them are exaggerated by quite some margin.

From a layman's perspective one of the interesting studies was the way people preferred 128kbps mp3 files to FLAC.. was a few years ago and it exploded all over the net. It was a blind test with a poll running for quite some time. I think it achieved most traffic through the various techblogs like Gizmodo and such.

What it tells us is that very few people from the audience actually care at all about how it sounds. I can very much relate to that as I have no problem watching movies with horrible color correction where some of my professional video/graphics guru friends cringe and claim the movie is unwatchable.

But the more interesting argument is how much does aliasing and other kinds of damage either hurt or enhance the music making process. For me, aliasing at a higher level is a direct irritant, literally! It makes me irritated and makes me want to turn off the signal which is aliased. Of course it's a very minor irritant but it's still there.

AD/DA conversion to me sounds like a "pressure change". The signal's transients and overall impact is changed one way or the other. Perhaps the AD/DA conversion filter causing some phase changes or something? It's not at all irritating to me. The change either fits the signal or it doesn't, either way it's just an aesthetic change for me, nothing else.

Intermodulation distortion can also be "irritating" but only when you drive signals with a lot of high frequency content into the device that causes IM distortion. With bass tones and such it can even be quite interesting sounding, not irritating at all.

I've always been interested in what constitutes ear irritation, or fatigue. Is it purely energy based around a certain frequency area (2-6kHz?) or is it energy + time based around the typical 2-6kHz area? What I mean with time based, can we get away with more distortion if it is dosed at short time intervals, keeping the overall energy distribution the same? Not sure how to explain this and make any sense but that's as good a try as I can come up with.

EDIT: final thoughts about this; It's relatively straight forward to run test tones and analyse things and see what kind of distortion you have but the bad part is that the more you look at these things, the less it makes sense in what ACTUALLY constitutes a "problem" for the ear/brain (or an irritant). Theory and the ear/brain don't seem to always correlate, at least not my ear/brain.

Cheers!
bManic

[Burillo]

well, exaggerated or not, his points are valid in that they make you think whether it's really what you're hearing, or you're talking yourself into hearing what you are expecting to hear. our ears are way too untrustworthy to rely on them 100% of the time.

however, consider the following. say, you have two tracks that are phase-aligned. you then pass each of them through an outboard device - the result is two tracks will have slightly different phase characteristics due to clock skew, jitter etc. of the DAC's.

aliasing is of course a source of irritation, but chances are it will not be evident on a lot of sources (e.g you can clearly hear it in a supersaw doing vibrato, but what about pads?), and not all processes introduce aliasing.

but anyway, we're gravely off topic here.

[gavriloP]

Little more OT about Winer's tests: no wonder you don't hear much difference in his samples because the originals are already pretty crappy SB synth sounds that sound aliased Only thing that this proves is that once the sound is ruined, more ruining isn't so bad... he is a man of agenda, not scientist.

Well, enough of that. I bought FC70 and to my ears it is very nice comp! There is some similar vibe with Bootsie's Density mk.iii but I prefer FC70 way over it. Perhaps the saturation is nicer with Density but FC70 has much better attack (no ticks or clicky sounds with transients).

With this, DCAM freeComp and Tokyo Dawn Feedback comp II, I got my compressors covered for awhile

[bmanic]

well, exaggerated or not, his points are valid in that they make you think whether it's really what you're hearing, or you're talking yourself into hearing what you are expecting to hear. our ears are way too untrustworthy to rely on them 100% of the time.

Absolutely agree! This is why one MUST DO abx tests to confirm ones suspicion. If the test tells you that you can't hear the difference then you just have to trust the test. There's no going against statistics/science, no matter how annoyingly strong placebo can be.

I'm also constantly surprised by how much colors/GUI brightness affects the perception of sound. For instance in Camel Audio's Alchemy synthesizer my brain tricks me that the basic Saw wave sounds brighter with the default GUI colors than with the dark GUI!

Evil brain.. it's evil!

Cheers!
bManic

[meloco_go]

well, exaggerated or not, his points are valid in that they make you think whether it's really what you're hearing, or you're talking yourself into hearing what you are expecting to hear. our ears are way too untrustworthy to rely on them 100% of the time.

Absolutely agree! This is why one MUST DO abx tests to confirm ones suspicion. If the test tells you that you can't hear the difference then you just have to trust the test. There's no going against statistics/science, no matter how annoyingly strong placebo can be.

While this is true, it is easy to construct a bad test (without proper control and on unequcated lustener). And as much as it is easy to convince oneself that there is a difference when there isn't, it works both way; and actually it is easier to not find a difference in a blind-test when you're not trying to find it.

[Burillo]

While this is true, it is easy to construct a bad test (without proper control and on unequcated lustener). And as much as it is easy to convince oneself that there is a difference when there isn't, it works both way; and actually it is easier to not find a difference in a blind-test when you're not trying to find it.

i would argue that this is exactly the point - would you notice if you weren't looking? if not, then why does it matter?

[meloco_go]

While this is true, it is easy to construct a bad test (without proper control and on unequcated lustener). And as much as it is easy to convince oneself that there is a difference when there isn't, it works both way; and actually it is easier to not find a difference in a blind-test when you're not trying to find it.

i would argue that this is exactly the point - would you notice if you weren't looking? if not, then why does it matter?

The idea is to perform tests for yourself, and not trust anybody who claims that he proved with ABX test that there's no difference between A and B. People's abilities and experience varies.

[Tapehead]

While this is true, it is easy to construct a bad test (without proper control and on unequcated lustener). And as much as it is easy to convince oneself that there is a difference when there isn't, it works both way; and actually it is easier to not find a difference in a blind-test when you're not trying to find it.

i would argue that this is exactly the point - would you notice if you weren't looking? if not, then why does it matter?

The idea is to perform tests for yourself, and not trust anybody who claims that he proved with ABX test that there's no difference between A and B. People's abilities and experience varies.

That's called maturity. It takes awhile to learn how to trust yourself and then condition yourself into practicing your OWN findings unless what you want to do is exactly what someone else is doing.
After years of programming the "coolest possible beats" I've realized that chasing "cool" was (is) just too safe. The whole idea of making a beat or creating anything with drums in mind became a totally dead end for me.

Therefore, I also had to rethink my idea about suitable tools for my productions. For example, the majority of saturation tools (including highly praised Sound Toys) were not in fact beneficial to my concept of sonic beauty. For a long, long time I thought I was just not using them right because "this and that awesome engineer" said those tools are god's gift to sonic nirvana.

Let's say when I tried to use bootsie's newest saturation concepts I felt dizzy…it was like I was putting on some 3D cinema glasses but instead of watching something in 3D I saw blurry, falling apart images. Even at the most subtle of settings thing just wouldn't blend.

This is perhaps one of the areas where I really got to like Nebula despite the fact that so many libraries there are just total fluff. BTW this FC-70 combines extremely well with some of the Tim's and Eric's libraries. As I said before in this tread, for me, lots of Nebula stuff works great with very clean, mature digital tools such as Fabien's TDR compressor. Try chaining TDR with Smooth 609 from Cupwise…You might not get Neve but in my experience you might get something much better. Or place some Eric's tapes after FC-70 and drive your signal into it (nice textures for back ground, acapella vocals).

Very good discussion in here!
Thank you

[Kingston]

This whole thread and the review: TLDR

But I will still point out the image of the opened up unit on the review website is not a Fairchild 670. It's someones DIY project (from this decade) based on a design named PM670 (or PM660) that is very different from anything Fairchild.

Fairchild 670 doesn't even have circuit boards, let alone ultra-modern white ones! It was knitted point to point, all of them hand made.

[bmanic]

And the plugin is based on this? Interesting!

Speaking of plugin emulations, I've done some testing of the various IK plugins and something is really starting to bother me with the attack of many of the emulations. Are they using convolution for parts of the sound characteristics? What the hell is going on? I just enabled the demo of the IK SSL compressor model and it has pretty much the exact same "clicky" attack as their Fairchild emulation.

I could have sworn that the real SSL 4000 bus compressor didn't click this way on the attack, especially when set to fastest attack. Instead it actually flattens and "chunks it up" in a nice way. This is also how the SSL Native Bus Comp reacts and Native Instruments SSL plugin. I haven't used any real SSL bus compressor for many years but my memory of it is definitely not what the IK plugin presents.

The actual compression action of the IK plugin is very controlled and rather nice but for me the clickyness of the attack is a real problem. It's the exact sound I do NOT want. Heck it's what I use compressors for, to shape the transient in a non-clicky way. I want them chunky!

Cheers!
bManic

[Cupwise]

so what analog comps do you use that squash transients flat like an L2?

and if such comps existed before L2.. why was it considered revolutionary?

[Davias]

EDIT : Cupwise beat me to it ^^

If you want to flatten a transient, just use a limiter ^^ Every soft comps I tried let's transients go through even with fastest settings.

When I realized that I switch all my limiters for clippers, and all my fast-attacks compressors for limiters ^^ Only keep compressors for ultra-slow attack/fast release

[Cupwise]

im going to go into more detail about this because i'm seeing this being spread around more and more. it's just a growing misconception about how classic analog comp emulations should work, based on peoples' ideas about how compressors in general should work, which are based on their experience with modern software compressors, and based on modern production techniques.

...and something is really starting to bother me with the attack of many of the emulations. Are they using convolution for parts of the sound characteristics? What the hell is going on? I just enabled the demo of the IK SSL compressor model and it has pretty much the exact same "clicky" attack as their Fairchild emulation.

I could have sworn that the real SSL 4000 bus compressor didn't click this way on the attack, especially when set to fastest attack. Instead it actually flattens and "chunks it up" in a nice way. This is also how the SSL Native Bus Comp reacts and Native Instruments SSL plugin. I haven't used any real SSL bus compressor for many years but my memory of it is definitely not what the IK plugin presents.

The actual compression action of the IK plugin is very controlled and rather nice but for me the clickyness of the attack is a real problem. It's the exact sound I do NOT want. Heck it's what I use compressors for, to shape the transient in a non-clicky way. I want them chunky!

Cheers!
bManic

so you are saying that older analog compressors acted like an l2 and grabbed every bit of signal coming in? how did they accomplish that, being in the analog realm and without the benefit of look-ahead? no.

analog comps let some of the signal through. true story. i have recordings of several of them, showing just that. with threshold dropped down and ratio maxed, you will get a big 'clicky' sound right at the start. some can be very fast and maybe the 'click' is only half of a cycle of a 1khz tone, but you will still hear that. in those clips, all attacks were set to fastest speed. those are sine tones, but you will get the same results with drums or anything else with a jump in volume. some are more pronounced than others but it's only because they were made with different amounts of compression- the peak of the 'click' is always at the initial level before compression. so if you had 30db of gain reduction, the click will be 30db louder than what follows.

my opinion is that since L1/L2 and the clones that followed, we've been 'spoiled', and people now just take that type of compression for granted, and are even starting to think that that's how analog comps must have worked, because they are always held in high regard by those who used them. so if a model of one isn't able to catch everything, it's obviously wrong. but the trend of squashing everything super flat is a more recent thing. back in the 70s you probably had 0 engineers complaining about not being able to squash every last iota of a transient out of their drums.

here's what you do- back off on the threshold and/or ratio a bit! yes it's that easy. at least, that's what they did back then. they didn't squash stuff to hell and back, and they didn't want to. most of the time they probably only used a few db of compression on that drum bus. if you only need a few db of reduction that 'click' will only be a few db, and you may actually want to emphasize the transient anyway (yes some people used/use comps to actually make transients LOUDER). this might sound totally insane to some people, but they didn't always use the fastest attack setting possible! drum transients are part of where they get their 'punch', and keeping them allows the drums to be heard better in a busy mix if the drum transients peak above the rest of the mix just a bit. i don't get why everyone wants to suck that out of the sound these days by squashing them flat down to the level of the drum's body, leaving something like a square-wave burst. but fair enough, the ability to grab the whole transient (which kind of requires digital look-ahead, not counting when you overdrive something and get instant compression because that also results in tons of distortion) does open up another area of possibility that didn't exist before. the problem happens when people begin assuming that comps always worked that way. no. they just didn't.

i also have to question your use of the word 'shaping' when you say you like to shape your attacks, even though you seem to be talking about squashing them flat. nowadays lots of productions are made with synths, drum samples or synthetic drums, etc etc etc and there are basically no rules with that. so ok, have at it. on the other hand if you just recorded a band's drummer back in the 70s, the standard thing to do was to do small touches to enhance that performance. that can be done with a few db of eq cut/boost here or there, and a few db of compression. but they wanted it to still sound like a drummer in a band actually hitting drums- unless they didn't. yes, some people were experimenting even back then, but most recordings were mixed so that they still sounded like a performance by a group of musicians. the main concept of mixing was just to get these different elements to work best with each other. that never entailed squashing the life out of every element. drums have peaks in nature. you can squash them out of there and then use ducking and all these other tricks that mastering guys do to get 'louder' end mixes and still have the stuff sound 'good', but those drums are going to sound anything but natural. people might not care about that nowadays but i think most producers/mix engineers who worked with actual bands back in the analog days did.

so now you have the idea that IK may be using convolution, because some nebula comps you saw left a loud click when set to fastest attacks, and obviously that's due to a flaw in nebula (and i'm not saying that nebula is without its flaws). no. it's due to you having false expectations about how 'classic' compressors worked. there's a reason why Waves L1/L2 was/is considered revolutionary. comps couldn't do that before them. L1/L2/L3 etc being digital is WHY they can do what they do. so maybe emulations of classic comps aren't what you are looking for. and thats 100% ok. but people need to stop furthering the misconception that 'classic' analog comps somehow worked like an L2. it's getting out of hand.

if you get a loud click of say 10+db, then maybe it could be said that you are just simply using too much compression. you'd be getting 10+db of gain reduction to get that click. in 'the day' that would probably be considered an extreme amount of compression for most sources, by most engineers. OR maybe instead of saying you are using too much compression, we could say that you are using 'modern techniques'. ok. but that's just the thing. these classic comps that some people are emulating, they are just that. classic. they aren't modern. they weren't designed to do that. the users of them didn't want to. the whole entire culture was different.

[Tapehead]

Well, I'm not bmanic but I think the clicks he is talking about are not necessarily to do with fast attacks. I get clicks with IK's model even on smooth vocal tracks with almost any settings and the clicks are random and often (not always) at the beginning of phrases. Actually, one of the reasons I skipped on investing into this compressor even though I really liked the overall tone.

I own some good compressors ( MC77, Aphex compellor, Action, TL Audio, API 525) and I never heard them click unless I do something totally wrong. Then again in my productions I almost never have to use super fast attacks…