how to get a warm, analog sound from a digital recorder?
- Thread starter ceegee
- Start date
PRHunt
I void warranties
Clearly we require more "bits". I recommend a TL074. It has 32 bits and is audibly superior.
Many of us have gone down this long road of digital discovery. In my experience, digital is given an unfair indictment. The quality of audio has less to do with the medium and more to do with the sophistication of the equipment. This is true of an all-analog chain as well as a digital interface.
Digital interfaces that cost less than $1000 have significant limitations simply to bring the price within reach of most of us. And it isn't just the digital section...all of those analog inputs and outputs don't perform anything like a major studio's analog section, either. In fact, I once discovered after switching out analog circuits that the muddiness in my mixes came from the original analog sum amplifier's inability to resolve detail in thick mixes. It wasn't the digital section.
There are four ways to improve these devices, however, and come darn close (in fact, almost indistinguishable) from the big studio rigs (acoustics excepted, of course). Starting with the most significant --
1) Drive your digital boxes with a super accurate word clock ($500-$1500). This may be the biggest improvement that those using an inexpensive all-digital chain can make. Opens up the sound, extends transparency, reveals transients, and eliminates subtle jitter-induced phase cancellations. You are 50% closer to the big guys, and you'll notice it immediately.
2) Modify your analog ins/outs (500-$1000). There are firms that do this (Black Lion Audio is one). Their mods will add transparency to the raw signal as it comes in...as well as send it out unaltered. Especially important when you consider that a single send/return will pass the signal through several analog and digital stages. An alternative is to purchase an interface that is factory designed to perform among the best (Halo, for example)...but this is a large investment. Regardless, either modify or upgrade, and you are now performing at 80% of major studios.
3) Properly sampling analog inputs is more challenging that you might think. Add an external stereo AD (Analog to S/PDIF) box in the $1000-$2000 range for your two most used inputs and your sound will become downright realistic, airy, and reproduce the sheen of the original (Apogee, Benchmark, Drawmer, Black Lion Sparrow, others). We're 90% there.
4) Feed the S/PDIF main output of your digital interface into a studio-grad DAC for subsequent monitoring (Cranesong, Zodiac, Benchmark, Apogee, etc.). You'll spend $1000-$2000, and will make superior decisions during the mix. Bingo, 100% of the performance (if not the versatility) of a major studio's signal chain (well, excepting those $10,000 monitors).
There, your journey is complete. You will experience audio nirvana, and your clients will be euphoric.
Perform all four of these alterations and you will be in the realm of big league sound. You will be more than pleased with the native sound of the digital sound (which will surpass tape with impactful accuracy)...and you will be equally pleased that it reproduces the sound of analog precisely, as you desire, by using well-chosen plug-ins or by bouncing the mix out to reel and back.
Your wallet will be lighter...and your spouse might actually notice the difference and approve. The total investment will still be a fraction compared to all-pro big-house stuff.
Digital interfaces that cost less than $1000 have significant limitations simply to bring the price within reach of most of us. And it isn't just the digital section...all of those analog inputs and outputs don't perform anything like a major studio's analog section, either. In fact, I once discovered after switching out analog circuits that the muddiness in my mixes came from the original analog sum amplifier's inability to resolve detail in thick mixes. It wasn't the digital section.
There are four ways to improve these devices, however, and come darn close (in fact, almost indistinguishable) from the big studio rigs (acoustics excepted, of course). Starting with the most significant --
1) Drive your digital boxes with a super accurate word clock ($500-$1500). This may be the biggest improvement that those using an inexpensive all-digital chain can make. Opens up the sound, extends transparency, reveals transients, and eliminates subtle jitter-induced phase cancellations. You are 50% closer to the big guys, and you'll notice it immediately.
2) Modify your analog ins/outs (500-$1000). There are firms that do this (Black Lion Audio is one). Their mods will add transparency to the raw signal as it comes in...as well as send it out unaltered. Especially important when you consider that a single send/return will pass the signal through several analog and digital stages. An alternative is to purchase an interface that is factory designed to perform among the best (Halo, for example)...but this is a large investment. Regardless, either modify or upgrade, and you are now performing at 80% of major studios.
3) Properly sampling analog inputs is more challenging that you might think. Add an external stereo AD (Analog to S/PDIF) box in the $1000-$2000 range for your two most used inputs and your sound will become downright realistic, airy, and reproduce the sheen of the original (Apogee, Benchmark, Drawmer, Black Lion Sparrow, others). We're 90% there.
4) Feed the S/PDIF main output of your digital interface into a studio-grad DAC for subsequent monitoring (Cranesong, Zodiac, Benchmark, Apogee, etc.). You'll spend $1000-$2000, and will make superior decisions during the mix. Bingo, 100% of the performance (if not the versatility) of a major studio's signal chain (well, excepting those $10,000 monitors).
There, your journey is complete. You will experience audio nirvana, and your clients will be euphoric.
Perform all four of these alterations and you will be in the realm of big league sound. You will be more than pleased with the native sound of the digital sound (which will surpass tape with impactful accuracy)...and you will be equally pleased that it reproduces the sound of analog precisely, as you desire, by using well-chosen plug-ins or by bouncing the mix out to reel and back.
Your wallet will be lighter...and your spouse might actually notice the difference and approve. The total investment will still be a fraction compared to all-pro big-house stuff.
mshilarious
Banned
If you need to sync several boxes you use a word clock, but for any given box the external clock will nearly always *increase*, not decrease jitter, irrespective of the quality of the external clock. Inexpensive converters are likely to do relatively worse on external clock because of the specious quality of their PLL circuit. Note that a converter operating on external clock does not bypass its internal clock--it can't, because the actual serial clock rate is much much faster than the word clock signal--instead, it attempts to pull its internal clock to sync up with the external. If the PLL that must do that isn't robust, you'll get much more rather than less jitter.
This is easily measureable; Sound on Sound did an excellent article on the topic.
miroslav
Cosmic Cowboy
That's the SOP I followed and was recommended by the manufacurer. I use three A/D/A boxes, one is the "primary" and gets internal clock sync, then from that box I take its Word clock and feed it to the other two for their sync.
I've not bothered with an exernal Work clock box.
chuckduffy
Well-known member
This is an interesting thread. I find the idea of a perpetual search for 'warmth' distracting and for the most part dont understand it. I do listen to samples here and in other forums and get the idea. I notice many people looking for warmth are trying for a full 'rock/pop' band sound with virtual instruments for drums, bass, piano, etc. and record guitars and basses through amp sims. The out of the box sounds and presets on all this stuff has a really hyped high end. Combining them all together results in a super bright, too articulate, plastic sounding mix.
One dirt simple thing I do is simply roll off high end on alot of stuff to mimic the frequency responses of older equipment. Recording guitars by with a microphone in front of an amplifier can help alot. Bass through a decent pre-amp and skipping the sims can help. I think alot of people recommend tubes for warmth, but I think alot of the sound people call warmth in older gear comes from transformers not tubes.
I started out on an analog 4track about 25 years ago. I don't miss it all that much, although I do like to listen to the songs once in a while.
One dirt simple thing I do is simply roll off high end on alot of stuff to mimic the frequency responses of older equipment. Recording guitars by with a microphone in front of an amplifier can help alot. Bass through a decent pre-amp and skipping the sims can help. I think alot of people recommend tubes for warmth, but I think alot of the sound people call warmth in older gear comes from transformers not tubes.
I started out on an analog 4track about 25 years ago. I don't miss it all that much, although I do like to listen to the songs once in a while.
Thanks, John. Yes, this article is well known. Good explanation and accurate science. However, it is also a few years old. I suspect inexpensive built-in clocks are not as rock stable as external and apply accumulated error as multiple tracks are mixed and bounce-backs are added. This error is likely greater than the very minor errors associated with the connection of external clocks, in my experience, IF the external clock is very good and cable lengths are minimum.
Also, since that article was written, IC processing speeds have increased and their physical size reduced, making servo PLL response times faster. I'm not disputing the article, just its conclusion that the outcome is always inferior. So far, my clients and I have been gratified at the improvement a very fine clock brings to a $650 digital interface. But it is indeed worth noting that in each of these three cases, there were indeed multiple devices (for example, MOTU 828, MOTU 2408, and Focusrite Octopre).
Also, since that article was written, IC processing speeds have increased and their physical size reduced, making servo PLL response times faster. I'm not disputing the article, just its conclusion that the outcome is always inferior. So far, my clients and I have been gratified at the improvement a very fine clock brings to a $650 digital interface. But it is indeed worth noting that in each of these three cases, there were indeed multiple devices (for example, MOTU 828, MOTU 2408, and Focusrite Octopre).
mshilarious
Banned
Except there really haven't been very many new ADC ICs introduced since the article was written (two years ago). The pace of change in that segment of the industry has slowed substantially, mostly because there are diminishing returns in further improvements: they are starting to hit the limits of what is physically possible with existing technology.
Converters could maybe use a bit more DSP horsepower for better decimation routines, but again the converters on the market are already pretty good and haven't changed much since 2006. Whereas if you go six years back from that there was a bigger difference.
PLLs are of course usually not internal to the IC, but that is a fairly mature technology as well.
Anyway, your post is conjecture. You have an external clock, measure it against internal, post the results. The SoS experiment was very simple and can be performed by anyone. I mean if you need a clock to sync multiple devices, you need a clock, but there is no reason to presume it actually reduces jitter because you like the sound.
volltreffer
Imply depressed
IMHO, the best tip is to get aquainted to your new gear. Most often the added high end of the signals recorded (and maybe the distortions that add further hihg end) together with the missing cheap tape hiss causes the impression of less warmth. It would be ridiculous to add 2000$ of gear to your el-cheapo recorder. Try to roll off via EQ what you don't need, use proper gain and you'll prolly be in the game.
Just my 2c --- altough I enjoyed reading this thread and even learned that word clock is rather slow and only triggers a PLL - a thing I never bothered with until today...
Ciao,
aXel
Just my 2c --- altough I enjoyed reading this thread and even learned that word clock is rather slow and only triggers a PLL - a thing I never bothered with until today...
Ciao,
aXel
volltreffer
Imply depressed
Thank you SOOO much for this one; but the idea of getting 32-Bit in a TL074 is stellar! Now I finally understand why there never was a TL078 series! It was not the housing needing too many pins --- 64 Bits weren't available at that time!
Hey, John,
I passed your points to an digital designer near here who says that all-in-one boxes sometimes have an Achilles' heal...and that has to do with power fluctuations. He says that inexpensive single-rack-space interfaces rarely have the space for the large power supplies and the filtering & isolation they require. He says he has a unit where current fluctuations caused by the analog circuits cause a modulated jitter in the digital section that is audible when downsampled...that one can just make out the beat and stronger sounds. This causes the incoming audio to be sampled incorrectly...and that the error can indeed be multiplied when sending out and back in. So adding external clocks and samplers can contribute to an improvement, but more for their independent power supplies than anything else. An exceptional clock and short cables (he says same-length cables, too) minimize other errors, leaving a net gain in audio quality. I want to hear those downsampled files!
I passed your points to an digital designer near here who says that all-in-one boxes sometimes have an Achilles' heal...and that has to do with power fluctuations. He says that inexpensive single-rack-space interfaces rarely have the space for the large power supplies and the filtering & isolation they require. He says he has a unit where current fluctuations caused by the analog circuits cause a modulated jitter in the digital section that is audible when downsampled...that one can just make out the beat and stronger sounds. This causes the incoming audio to be sampled incorrectly...and that the error can indeed be multiplied when sending out and back in. So adding external clocks and samplers can contribute to an improvement, but more for their independent power supplies than anything else. An exceptional clock and short cables (he says same-length cables, too) minimize other errors, leaving a net gain in audio quality. I want to hear those downsampled files!
mshilarious
Banned
That was an interesting enough theory for me to test. Sounded plausible anyway, although I wondered how the PLL circuit isn't affected by the same noise. Maybe it has better PSRR.
Anyway, I figured what could be worse than a PCI card with onboard converters? PC power is dirty as hell because most digital circuits don't care (just those that need to interface with analogland at high precision I guess). This particular card--a Steiny VSL2020 from 2003--has no significant capacitance, so it should be subject to these woes.
The test methodology was to repeat the SoS test, 1kHz sine loop. Left channel (top) is loop, right channel had no signal so that's crosstalk.
And look! Blue is . . . internal clock! Yellow is clocked to the ADAT coming off an RME ADI 8-DS. Your friend is pretty smart!
The only thing I don't get is why the jitter is 65Hz off rather than 60Hz, but hey, sideband frequencies are jitter so . . .
I will say that the RME unit is a 1u rack device, I dunno why that isn't plenty of space for capacitors . . . and yes, it has a tightly squeezed in switching supply that has a fair amount of ripple. I measured it once but I forget what is it. I suppose its circuits have good enough PSRR that it doesn't matter. I mean not Lavry levels of performance, but it works.
In conclusion, I still have to wonder why you wouldn't just spend the clock money on better converters . . .
Anyway, I figured what could be worse than a PCI card with onboard converters? PC power is dirty as hell because most digital circuits don't care (just those that need to interface with analogland at high precision I guess). This particular card--a Steiny VSL2020 from 2003--has no significant capacitance, so it should be subject to these woes.
The test methodology was to repeat the SoS test, 1kHz sine loop. Left channel (top) is loop, right channel had no signal so that's crosstalk.
And look! Blue is . . . internal clock! Yellow is clocked to the ADAT coming off an RME ADI 8-DS. Your friend is pretty smart!
The only thing I don't get is why the jitter is 65Hz off rather than 60Hz, but hey, sideband frequencies are jitter so . . .
I will say that the RME unit is a 1u rack device, I dunno why that isn't plenty of space for capacitors . . . and yes, it has a tightly squeezed in switching supply that has a fair amount of ripple. I measured it once but I forget what is it. I suppose its circuits have good enough PSRR that it doesn't matter. I mean not Lavry levels of performance, but it works.
In conclusion, I still have to wonder why you wouldn't just spend the clock money on better converters . . .
mshilarious
Banned
Stupid not posting picture BBS software!
mshilarious
Banned
Alright, can anybody else see those? All I get is a blank attached images box 
mshilarious
Banned
Let's try again!
Supercreep
Lizard People
....Meh.
mshilarious
Banned
Tubes work OK as transducers but you have to yell at them pretty loud or maybe bang on them with a drumstick. Capacitor and moving-coil microphones are a bit more practical.
BushmasterM4
New member
Steven Slate VCC plugin does a fantastic job. And you get several vintage consoles to choose from. Its not a slap it on a channel and there you go plugin (magic pill), but used properly it sounds very nice.
Cassette decks don't give warm analog sound. They just remove high frequency detail and add lots of hiss. I would avoid them.
A 1/4" tape deck of decent quality can be a nice way of adding warmth to mixes. But tape is expensive and getting harder to get. Tape machines also require maintenance. I still use mine at times for special projects. Summing mix stems OTB and sending them through nice analog gear such as compressors with tubes or transformers before or after the summing stage can add warmth. IF you are ITB the PSP tape saturator, Mastering compresssor and vintage warmer are handy plugins.
A 1/4" tape deck of decent quality can be a nice way of adding warmth to mixes. But tape is expensive and getting harder to get. Tape machines also require maintenance. I still use mine at times for special projects. Summing mix stems OTB and sending them through nice analog gear such as compressors with tubes or transformers before or after the summing stage can add warmth. IF you are ITB the PSP tape saturator, Mastering compresssor and vintage warmer are handy plugins.