Keeping an effect the same when converting it from a send effect to an insert effect

[Thousand Ways]

Would much appreciate advice on this:

I have a delay effect connected as a send effect on one channel of the master section. The delay is used on only one channel of my song, so instead I want to move the delay so that it's an insert on just that one channel, thereby freeing-up one send channel on the master section. However, when I move the delay from being a send effect to an insert effect, it's much louder, hitting around -12dB instead of the former -16dB. The effect now also sounds too crisp and sharp. Turning the delay effect dials further toward "dry" instead of "wet" does not make the signal any quieter. Is there some way to shift an effect from being a send effect to an insert effect while ensuring that the final sound remains exactly as it was?

This might be a very basic problem, and I might be missing something very obvious. I can't find anything on YouTube relating to it, but it might be a case of not knowing the right terms to search for.

[deigm]

I would have thought the wet/dry knob was the answer, but if that's not working for you, you could try putting your delay device in a new mix channel, split the output of the track you want to add delay to and send the new split signal to your new delay mix channel. You should then be able to ride that delay channels fader to whatever level you like without effecting the original.

[jam-s]

Yes, imho the easiest way would be to first create a parallel channel and put the (former) send FX there as an insert FX. Then you can easily level it to your liking.

[integerpoet]

Yes, imho the easiest way would be to first create a parallel channel and put the (former) send FX there as an insert FX. Then you can easily level it to your liking.

This.

A send is simply fundamentally different from an insert.

Routing a mixer channel to a bus with only one input might be the closest thing to a send that was only being used on one channel.

But I have to ask… Have you really exhausted all your sends?

[Thousand Ways]

Many thanks for the replies, all.

you could try putting your delay device in a new mix channel, split the output of the track you want to add delay to and send the new split signal to your new delay mix channel. You should then be able to ride that delay channels fader to whatever level you like without effecting the original.

That's clever; thanks.

A send is simply fundamentally different from an insert.

I understand this, but I don't understand in what way it's different. When the effect is used as a send, it seems to smother or blur the original signal more (which in this case is what I want), whereas when the effect is used as an insert, it seems to be behind the original dry signal.

But I have to ask… Have you really exhausted all your sends?

Okay, this is the moment when I find out that it's possible to have more than the eight sends available via the main mixer. But how? I've never needed more than eight.

[robussc]

Okay, this is the moment when I find out that it's possible to have more than the eight sends available via the main mixer. But how? I've never needed more than eight.

8 is the limit on the SSL. As others have noted, you get a decent approximation of the current setup by simply adding a parallel channel and moving the send effect to that and blend to suit. The send is not doing any special blending beyond just routing audio, but there are two level controls that you don't explicitly have with an insert: the send level and the return level. If you have the parallel channel set to have the insert effects before the fader you might be able to emulate the "send level" with the input trim control (not sure if that's bypassed with parallel channels) and then the "return level" with the fader.

(If it is bypassed then you could have a simple 6:2 line mixer before the effect - in the inserts area - and use that to set the send level.)

[Thousand Ways]

This is really useful, thanks.

[integerpoet]

A send is simply fundamentally different from an insert.

I understand this, but I don't understand in what way it's different. When the effect is used as a send, it seems to smother or blur the original signal more (which in this case is what I want), whereas when the effect is used as an insert, it seems to be behind the original dry signal.

An insert is the simple case. An effect takes in the signal and processes it to the extent specified by any dry/wet knob and spits out the result.

A send is far more complicated, even in the simplest case (for a send). A portion of the input is sent to the effect, which processes that portion to the extent specified by any dry/wet knob and spits out the result. And then a portion of the result is added back in to the original signal. So for the simplest send you end up with all the original signal plus an additional signal which is merely based on the original. (Clearly I love to bold things. Please forgive me.)

One way to think about this is to imagine an effect which doesn't do anything: It just outputs its input. If you were to send all of a signal to such an "effect" and accept all of the effect's output in the return part of the send flow, you'd essentially (assuming zero latency) be doubling the amplitude ("volume") of your original signal. (Circuitously to a comedic degree, I might add. )

An insert is fairly natural and intuitive to think about (especially for an effect which lacks a dry/wet knob): You're mutating a sound in some interesting way. A send is a complicated synthetic unnatural notion which occurs only in a mixing desk (or a simulation thereof, such as within Reason). When I apply a send, I don't merely think I'm mutating a sound. I think something along the lines of: "I'm doing that weird thing recording engineers have done for the last few decades to make things sound the way consumers of studio recordings understand and expect."

You can of course use sends creatively for any reason that occurs to you, but the reason that Reason only offers eight of them is … well, maybe it's because an SSL desk only has eight of them. But the reason there are only eight of them is that as a practical matter a single recording doesn't often "need" more instances of this particular complicated flow.

I don't know the original thinking behind the way sends work today — or even why they came to exist in the first place. Somebody had to believe they might be a good idea before going to the trouble of building the first (analog) circuit. Maybe someone like Selig will come along and school us on the history.

But I have to ask… Have you really exhausted all your sends?

Okay, this is the moment when I find out that it's possible to have more than the eight sends available via the main mixer. But how? I've never needed more than eight.

What I really meant was something like: "Wow: you have a lot of sends. Are you sure you're using them as intended?"

[Thousand Ways]

A send is far more complicated, even in the simplest case (for a send).

Thanks for the explanation. I had vaguely guessed at some aspects of this, but certainly not the whole picture.

My own rationale for choosing to send or insert has been down to whether the effect is to be applied to several instruments (send) or just one (insert). I've generally thought of send effects, being fewer and being applied to several instruments, as forming the acoustic of the track as a whole, and insert effects as anomalies that are somehow exempt from that overall sound.

What I really meant was something like: "Wow: you have a lot of sends. Are you sure you're using them as intended?"

I managed to read between those lines. I'm using six; the remaining two are taken up by Selig's ducker.

[robussc]

I don't know the original thinking behind the way sends work today — or even why they came to exist in the first place. Somebody had to believe they might be a good idea before going to the trouble of building the first (analog) circuit. Maybe someone like Selig will come along and school us on the history.

Not Selig, but back in the mists of time adding some reverb back into the mix (because as studio recording advanced the incoming signal was getting very dry) it required sending the signal to another room where the audio was played through a speaker and the resulting sound captured with a mic and then returned to the mixer. Insert effects weren't really a thing yet. Those were the original room reverbs. Soon plate reverbs were added which required less space but still weren't cheap. But that desire to send the mix to an effect processor and return it has been a standard for a very long time and is one of the best ways to bring a mix together.

[DaveyG]

Unless you are being weird or experimental, 8 sends should be more than enough.
I've usually got two for reverbs and one for delay. Beyond that it's inserts all the way.

[integerpoet]

I don't know the original thinking behind the way sends work today — or even why they came to exist in the first place. Somebody had to believe they might be a good idea before going to the trouble of building the first (analog) circuit. Maybe someone like Selig will come along and school us on the history.

Not Selig, but back in the mists of time adding some reverb back into the mix (because as studio recording advanced the incoming signal was getting very dry) it required sending the signal to another room where the audio was played through a speaker and the resulting sound captured with a mic and then returned to the mixer. Insert effects weren't really a thing yet. Those were the original room reverbs. Soon plate reverbs were added which required less space but still weren't cheap. But that desire to send the mix to an effect processor and return it has been a standard for a very long time and is one of the best ways to bring a mix together.

That makes a terrifying kind of sense. This is a bit like that story about why rocket size can be traced back through history to the width of a horse butt. I doubt that story holds up, but I love it. I asked GPT-4 for a version of it, and it said:

This seemingly absurd connection actually has a basis in historical infrastructure development. The story goes like this:

1. In ancient Rome, the width of a chariot was determined by the width of a horse's rear, as chariots were pulled by horses. This width was approximately 4 feet, 8.5 inches.
2. When the Romans built their roads, they made them just wide enough to accommodate the chariots. These roads were used for centuries and their width became a standard.
3. Fast forward to the 1800s, when the first railways were being built in England. The engineers used the same standard measurement for the width of the train tracks, which is known as the "gauge". This was partly because the same jigs and tools used for building wagons were used for building railways.
4. When railways were built in the United States, they also adopted the same gauge, which is now known as the "standard gauge".
5. Now, when the U.S. started building rockets (like the Space Shuttle), they had to be transported by train from the place they were built to the launch site. The rockets had to fit through railway tunnels, which were just wide enough to accommodate a train car of standard gauge.

Therefore, the size of the rockets was indirectly determined by the width of a horse's rear, as it influenced the width of the railway gauge. This is a simplified version of the story and there are many other factors that have influenced the size of rockets, but it's a fun way to illustrate how historical decisions can have long-lasting effects.

[Thousand Ways]

Not Selig, but back in the mists of time adding some reverb back into the mix (because as studio recording advanced the incoming signal was getting very dry) it required sending the signal to another room where the audio was played through a speaker and the resulting sound captured with a mic and then returned to the mixer. Insert effects weren't really a thing yet. Those were the original room reverbs.

Brilliant. And a means of remembering what a send is as opposed to an insert.

Unless you are being weird or experimental, 8 sends should be more than enough.
I've usually got two for reverbs and one for delay. Beyond that it's inserts all the way.

Isn't having a different insert on every instrument more weird and more "experimental"? At the very least, it presumably makes the file much larger. If you have a 48-track song, you might end up with 48 different insert units. If you were sticking within the limits of the main mixer you'd only have 8.

Therefore, the size of the rockets was indirectly determined by the width of a horse's rear, as it influenced the width of the railway gauge. This is a simplified version of the story and there are many other factors that have influenced the size of rockets, but it's a fun way to illustrate how historical decisions can have long-lasting effects.

To use McLuhan's adage, "we walk backwards through the rearview mirror", and the old malaise is grafted on to the new technology. Hence early television sets looked like the radios that preceded them, and so on. But the horse's rear story seems to overlook the influence of human scale: the width of a carriage was surely influenced by the horses being harnessed two abreast, because the human coach riders wanted to sit side-by-side in pairs.

[Ottostrom]

Isn't having a different insert on every instrument more weird and more "experimental"? At the very least, it presumably makes the file much larger. If you have a 48-track song, you might end up with 48 different insert units. If you were sticking within the limits of the main mixer you'd only have 8.

This comment makes it sound like you're not using the sends in a common practice way.
In most cases the sends are used for time effects (reverb/delay) and not for something like EQ, compression or saturation. If you have 48 tracks then indeed you'll want 48 compressors as inserts (probably group together some to a bus but you get my point) and not try to send them into one placed as a send effect.

[Thousand Ways]

This comment makes it sound like you're not using the sends in a common practice way.
In most cases the sends are used for time effects (reverb/delay) and not for something like EQ, compression or saturation.

Yes, this is what I do. I've never used the send inputs for compression. But my tracks often incorporate speech samples from many different sources, and I've found that it helps to have several different reverbs and several different delays as sends for these.

[robussc]

Yes, this is what I do. I've never used the send inputs for compression. But my tracks often incorporate speech samples from many different sources, and I've found that it helps to have several different reverbs and several different delays as sends for these.

In that case it might help to route a collection of samples to a mixer subgroup and then add a parallel channel to that,, to mix in the reverb for that set of voices? Assuming some voices are talking together in the same space of course

[selig]

Quick answer: if you want a PERFECT translation do this.

In the insert, use a Line Mixer and put your delay on the send/return of the line mixer and patch the line mixer into the insert points via channel one. On channel one, adjust the send level according to the included handy chart below.

For example:
If your original send was set to -6 dB, check the handy chart to see where that falls on the "0-127" scale – it happens to be "79", so set the send to "79" and you be matched perfectly (aka: passes the "null test"). That's it, no additional parallel channels, no SSL sends, no loss of dry signal when adjusting dry/wet.

Here's the chart, keep this in a handy place, it can be valuable in certain situations!

[Thousand Ways]

it might help to route a collection of samples to a mixer subgroup and then add a parallel channel to that,, to mix in the reverb for that set of voices? Assuming some voices are talking together in the same space of course

Thanks. I think that this would work in at least some of my cases.

Selig, many thanks for the cheat sheet. I had wondered about dB-to-0–127 conversion. I wasn't sure that the relationship between these values was the same in the case of each 0-127-measured instrument/mix channel/whatever.

In the insert, use a Line Mixer and put your delay on the send/return of the line mixer and patch the line mixer into the insert points via channel one. On channel one, adjust the send level according to the included handy chart below.

By "line mixer", you mean by adding, say, a 14:2-style mixer within the rack area?

[integerpoet]

…my tracks often incorporate speech samples from many different sources, and I've found that it helps to have several different reverbs and several different delays as sends for these.

One tricky bit about sends (which I deliberately avoided covering in my earlier post for the sake of pedagogical simplicity) is that when you send multiple — say three — mixer channels to a single effect, it's not as if that implicitly invisibly clones the effect twice. It actually sends the three channels to a send bus and then sends the output of that bus to the single effect. The output of the effect then comes back into your mix once. This is why for each numbered send there are three send amount knobs, one for each mixer channel, and only one return amount knob.

To apply the room analogy from earlier, let's assume the effect is reverb. The three signals are sent to three speakers in another room, where there's also a single microphone whose output returns to your mix. If you're using multiple reverb effects as sends and it sounds good to you, that's fine, but it's a bit like having multiple rooms at your studio dedicated to reverb, which is … extravagant? Luxurious? But often what one wants — independent of luxury — is to send multiple channels to a single effect so one can vary the amount sent by each channel to suggest the three speakers being situated at their own positions relative to the microphone, which in turn suggests three musicians in the room.

(This isn't like 3D object mixing at all, which is why I say "suggest"; audio engineers targeting stereo in a traditional way are more likely to describe their goal being along the lines of "sense of place" than virtual reality.)

There are of course many famous stories of different approaches. The one that comes to mind is the oft-told story about John Bonham at Headley Grange. This would be a bit like recording him with close mics in an anechoic chamber and then running him through a single reverb insert which modeled the stairwell and the mic position and then trying to integrate that into a mix of channels recorded in other rooms. So not even a send much less many channels into a single send. It worked, but are you Andy Johns? Last time I checked, I'm not, but I suppose that doesn't rule out the possibility that you are. Plus he apparently compressed the drum track, which is even more boggling but suggests he was after not psychoacoustic glue but bigness, which he oh-so-evidently got.

[selig]

it might help to route a collection of samples to a mixer subgroup and then add a parallel channel to that,, to mix in the reverb for that set of voices? Assuming some voices are talking together in the same space of course

Thanks. I think that this would work in at least some of my cases.

Selig, many thanks for the cheat sheet. I had wondered about dB-to-0–127 conversion. I wasn't sure that the relationship between these values was the same in the case of each 0-127-measured instrument/mix channel/whatever.

In the insert, use a Line Mixer and put your delay on the send/return of the line mixer and patch the line mixer into the insert points via channel one. On channel one, adjust the send level according to the included handy chart below.

By "line mixer", you mean by adding, say, a 14:2-style mixer within the rack area?

There are two smaller mixers in Reason, the 14:2 and the Line Mixer (6 channel). The Line Mixer is my first choice for simple things since it’s so compact. You can do the same setup with either mixer, the Line Mixer is all you need for this setup (and saves space!).

[Thousand Ways]

when you send multiple — say three — mixer channels to a single effect, it's not as if that implicitly invisibly clones the effect twice. It actually sends the three channels to a send bus and then sends the output of that bus to the single effect. The output of the effect then comes back into your mix once.
[…]
If you're using multiple reverb effects as sends and it sounds good to you, that's fine, but it's a bit like having multiple rooms at your studio dedicated to reverb, which is … extravagant? Luxurious? But often what one wants — independent of luxury — is to send multiple channels to a single effect so one can vary the amount sent by each channel to suggest the three speakers being situated at their own positions relative to the microphone, which in turn suggests three musicians in the room.

As I understood it, using one send effect for multiple channels is much more like sending them all to "one room", and using insert effects (each one potentially different) on each channel is much more like putting each of those channels in discreet rooms with their own acoustics and effects. No?

Bonham's levee-breaking snare exists mainly for the benefit of this.

There are two smaller mixers in Reason, the 14:2 and the Line Mixer (6 channel).

I confess that I've never really noticed them; I've only used the 6-channel mixer as part of a Combinator. Have never thought about having additional mixers within the rack, and then feeding those into a main mixer.

[integerpoet]

As I understood it, using one send effect for multiple channels is much more like sending them all to "one room", and using insert effects (each one potentially different) on each channel is much more like putting each of those channels in discreet rooms with their own acoustics and effects. No?

Yes, but "putting" — as opposed to "sending" — is doing a heavy lift in that sentence.

With an insert, the effect typically yields a portion of the dry signal and the inverse proportion of the wet signal. In other words, dry and wet portions exist at each other's expense because that's how dry/wet knobs typically roll.

With a send, people typically set the effect's dry/wet knob fully wet. And then the typical send signal chain yields all the dry signal plus a portion of the fully wet signal.

I'm very pointedly using the word "typically" here because it's possible to do all kinds of crazy things and therefore some people do them. And that's fine because neither signal chain makes sense if the goal were an accurate simulation of reality. Like all photographs, all audio recordings are lies.

Bonham's levee-breaking snare exists mainly for the benefit of this.

I am just the right age to appreciate both.