SOUND ON SOUND

[RichardT]

A couple of years back I mastered a CD and when I played it back straight through there was no problem but when I told the CD player to jump to a particular track the player would miss a couple of notes at the beginning. That was when I learned to put at least 0.25 secs silence at the beginning of a track because it matters on a CD. So far as I'm aware it doesn't much matter for anything else but if your music is ever likely to end up on a red book CD then put that space in. Since you never quite know where your music will end up nowadays I always put it in on everything.

CC

Exactly this! I always add some space at the beginning of my tracks in my DAW before rendering to audio (usually 1 quarter note of space in a 1/4 bar so that the rest of the track lines up to the bars). This prevents glitches at the start of the rendered file which can sometimes happen if the music starts at the very beginning.

[BJG145]

I always leave a small gap, though I can't say it's an exact science. I do find that playback can be cut off otherwise. Not to do so is asking for trouble IMHO.

Definitely. Some people do jump on this a bit too much; don't let these types cut off your final reverb tail either.

[nathanscribe]

This is interesting. I usually leave anything up to a whole bar, it feels much more natural to me to have a gap between the technical start point and the music. Especially if it's got some hiss on it. Not joking! There's something comforting about a bit of low level noise to prep my ears and brain, I find immediate starts and silence to be quite odd.

[Folderol]

I always leave a very tiny gap to allow the playback system to 'settle'. It's not an exact thing but probably about 1/4 sec. As most of my material has quite a bit of reverb I also leave long tails to ensure it has completely faded away.

[James Perrett]

Final question here and thanks everyone : I'm going to just add the gap in Reaper and render the file again, however Reaper's default is 24 bit 44.1, and the file is 16 bit, is it ok to render it as it is or should I change the project settings

There's no need to adjust the project settings - just change the setting in the render dialog box. Make sure that the render settings match the source file setting.

I used to do a bit of DJ'ing in the past so I always like things to start when I press the play button which is probably why I use a short space. I know that other mastering engineers often leave 10 CD frames between the start marker and the start of the music.

[forumuser840717]

In the "old days"* the PQ spec required that the start time reference for each track be offset from the actual start of the programme audio; the first track to be -30 frames** (so a full second before the real audio start time) and subsequent tracks to be -5 frames. (Though different mastering companies/engineers evolved differing policies on the specific numbers.) This was to allow the CD player time to lock into the audio data stream and unmute after seeking and starting playback before the actual music started so as not to clip the start of the track. Different players took different amounts of time to do this but the vast majority were fine with the standard offsets. (**SMPTE frames at 30fps NDF, not CD frames.)

On the original Philips PQ Editor, offsets had to be manually calculated and entered into the system; one got very good at doing timecode mental arithmetic when PQ'ing masters and woe betide anyone who didn't thoroughly check their work as typos or bad maths could lead to record companies chasing compensation when several hundred thousand CDs went out with faulty PQ data! A useful, but expensive, step forwards came with the Harmonia Mundi PQ editors (PQ Senior and PQ Junior) which could be set up automatically to apply various offsets to the manually entered timecode references and, with PQ Senior, could automate the writing of PQ burst(s) to the U-Matic tape masters, and their verification, taking care of all the machine control and timing of the process.

Now that PQ files are generated simply by positioning markers on a DAW timeline (or just ticking the auto marker box with some DAWs taking care of the nuts and bolts automatically or maybe allowing the user to set up some preferences) the actual mechanics of how and why they work the way they do largely has been forgotten, as has the art of setting things other than just track start markers.

PQ'ing used to be a fairly straightforward but painstaking and time consuming job requiring manual finding and entering of timecode references for several different marker types, with some clients having preferred house styles/standards particularly for use of the Track Start, Track End, Pause and Index functions so that their discs behaved in a uniform way. Other clients would have their PQ data manipulated to create specific tricks and/or offer extra options on playback and some didn't care, they just wanted discs that worked.

(* In the early days of CD mastering before DAWs, when the only way to make a master was using a PCM processor like a Sony PCM1630 (or 1610 or 1600 or a few others) recording onto a U-Matic 3/4" tape cassette in one of a small number of compatible U-Matic recorders. The PCM unit encoded the programme audio into a video signal to be recorded by the U-Matic machine (in the same way as the PCM-F1 type domestic systems worked with a suitable domestic video recorder to put 16 bit, 44.1kHz PCM audio onto consumer format video tapes). The U-Matic tape also carried a PQ burst on its analogue audio track 1 at the start of the tape, before the audio, and a SMPTE 30fps NDF reference LTC striped onto analogue audio track 2 which acted as a timing and positional reference.

Later, it was sometimes possible (by special agreement with a glass mastering plant who were prepared to risk it (usually with written waivers aginst the risks involved) and had the kit - most wouldn't and didn't) to use other formats such as PCM-F1 as a CD master but only if the audio data could be 'verified' which required a modified PCM-701 processor that was a pretty rare beastie.

Later still, tape formats other than the PCM/video based formats, such as DASH or Timecoded DAT could be used (perhaps when a special 'over-length' CD was to be mastered) on the same basis as above but it was not normal and was always with extensive warnings about reliability, particularly as the audio data integrity couldn't be verified on these until one or two special (and expensive) dedicated processors became available. By which time DDP masters and CD-R masters were becoming viable.)

[Hazer]

Well I just shifted it a bit in my DAW and exported it at 16 bits and I can hear no difference - there are very slight differences in the waveforms though, can't hear anything or maybe it's Reaper creating them slightly imperfectly?

Now that's done there's literally one snare hit sticking out sooo annoying. Can I do a very swift nick of automation or cross fade to the master to sort that do you guys think

[James Perrett]

Now that's done there's literally one snare hit sticking out sooo annoying. Can I do a very swift nick of automation or cross fade to the master to sort that do you guys think

From a purist point of view you would have to re-dither the result and consider any dithering that the mastering engineer had done - especially if they had used noise shaped dither. Multiple passes of noise shaped dither will result in high levels of high frequency noise which is why you would normally use flat dither during recording and only add noise shaped dither if needed right at the end of the mastering process.

From a practical point of view you could certainly get away with just adjusting that one short part provided you don't go changing any other levels.

[CS70]

Well I just shifted it a bit in my DAW and exported it at 16 bits and I can hear no difference - there are very slight differences in the waveforms though, can't hear anything or maybe it's Reaper creating them slightly imperfectly?

Probably the latter. Mixing engines do introduce small errors, like any calculation made with a computer, but with well-behaved code they're kept very small and well out of where they can affect the audio.. The waveform code is less critical so it may not be as sophisticated.

Now that's done there's literally one snare hit sticking out sooo annoying. Can I do a very swift nick of automation or cross fade to the master to sort that do you guys think

As James says. If you load a 16 bit sample stream into 24 bits samples, the last bits of each 24 bit word will be zero, so when they're truncated back to 16 bits (as you don't dither), there is no effect.

But the moment you change something you're creating some 24 bit samples in the result which do have data in the last 8 bits... and without dithering you risk to introduce distortion. It's a risk however, doesn't have to happen (by luck, the processing may not generate anything in the 8 bits you end up truncating away, or the truncations do not generate any issue at D/A time). Listen carefully with very detailed headphones to hear if there's anything untoward.

Mastering engineers normally have no problem with a slightly different revision if the mix is not changed and you could ask him/her to fix the stray snare as well.

[Hazer]

Thanks again all I went back to the engineer to fix the gap and snare and he's fine to proceed which is brilliant. I tried a little automation/ crossfade thing but sounded terrible and can't with my skills / software get anywhere near the desired result.

[Dave B]

It varies and it doesn't have to be silence. The CD will read the index for the disk and go to the correct location and play what it finds. Some systems are better than others at being able to play back instantly. On older systems, they might lose the first few milliseconds of audio, so when creating a master image for a CD, the mastering engineer will attempt to set the position for the beginning of the track ever so slightly early. Given that albums tend to have silence at the end of the tracks, this meant that you usually landed on, or very just before, your optimum point to play back which was in the 'silence' of the preceding track.

Which is a bugger if you have an album that has continuous tracks as it means that the skip point has to be ever so slightly out to compensate for older / lesser gear. Which was ok when we all had CD players, but is tedious if listening from audio files.

Ho hum

[Arpangel]

Different for an album, where the job includes the right amount of silence between consecutive tracks.. but who listens to albums nowadays (present forum excluded)

I still listen to "double albums" man.
The right amount of silence? On my albums it’s anything from 3/5 seconds between tracks, you have to cleanse your mind man, before the next track starts.

:D :D :D

[James Perrett]

The right amount of silence? On my albums it’s anything from 3/5 seconds between tracks, you have to cleanse your mind man, before the next track starts.

You're lucky if you get a second on many of the punk albums I master - but that's the way the label like it.

On the other hand, I have one client who likes10 second gaps.

[Arpangel]

The right amount of silence? On my albums it’s anything from 3/5 seconds between tracks, you have to cleanse your mind man, before the next track starts.

You're lucky if you get a second on many of the punk albums I master - but that's the way the label like it.

On the other hand, I have one client who likes10 second gaps.

Blimey, do people still write Punk music? I thought Grime had replaced it for the angry young man of today.

:)

[James Perrett]

Blimey, do people still write Punk music? I thought Grime had replaced it for the angry young man of today.

These are mostly recordings dating back to the late 70's (I was going to say reissues until I realised that much of it never saw the light of day in the first place) although there are still current bands playing punk.